ie. PLES AMERICAN NATURALIST A MONTHLY JOURNAL DEVOTED TO THE ADVANCEMENT OF THE BIOLOGICAL SCIENCES WITH SPECIAL REFERENCE TO THE FACTORS OF EVOLUTION VOLUME LI NEW YORK THE SCIENCE PRESS 1917 THE AMERICAN NATURALIST Vou. LI. January, 1917 No. 601 THE PERSONALITY, HEREDITY AND WORK OF CHARLES OTIS WHITMAN, 1843-1910 DR. CHARLES B. DAVENPORT COLD SPRING HARBOR Ar Pinhook, in the town of Woodstock, Maine, was born December 12, 1843,! a male child named by his par- ents, Joseph and Marcia Whitman, Charles Otis. Some time thereafter the parents and their children removed to Waterford (where they were in the ’50s) and in 1861 to Bryant’s Pond (Fig. 1). All of these places are in the southern part of Oxford County, 50 miles or so from the coast at Portland, in about the latitude of the Presidential Range of the White Mountains—a country of high hills and small mountains, covered with. pine and hardwood © forests, abounding in lakes with some fertile plains be- tween; a country affording a stimulating environment to a boy with an inherent love of nature. And a love of na- ture was widespread in the boys of this country and in their fathers. Take, for example, Jacob Whitman. Born in Easton, Mass., twenty miles south of Boston and ten miles north of Taunton it would appear that his lines had been pleasantly cast. But? he had an adventurous spirit and was perhaps ‘‘ very wilful.’’ He was with the ‘‘min- 1The date is given as December 14, 1842, by Lillie, yoy and Morse, 1912; also National Cyclopedia of American Biography, XI, p. 73. The date given above is that of the Whitman Genealogy. It Es agrees with that pasgain by Mrs. H. D. Smith (Mary Whitman) because her sister as born September 10, 1843, and Charles O. was born the December pes i 2Cole and Whitman, 1915, p. 42. i Ee 6 THE AMERICAN NATURALIST [ Vou. LI ute men”’ in 1775, probably fought at Bunker Hill, was at Harlem Heights, stormed Stony Point, helped capture Saratoga, and was at Trenton. He had married, 1777, and, after the war, 1781, came to the frontier settlement of Buckfield, Maine, which had been settled some four years before. Fellow townsmen, by the name of Record, had previously settled there and had told Jacob Whitman about the town and their tales of conditions at Buckfield evidently appealed to him, and induced him to migrate thither. He was a stout and musecularly built man ... and was inured to hardship. . . . He was an industrious citizen and paid strict attention to the clearing up of his lot, rearing a family and attaining a compe- tence. He was often selected for road surveyor, school agent, and ‘school committee and his name was always kept in the jury box until he got too old to attend to such service. He was a man of fierce spirit when aroused and positive in his opinions and fond of argument. No. 601] CHARLES OTIS WHITMAN 7 He declined to join the first church started at Buckfield as he did not believe its creed. Of other founders and early settlers of Buckfield it is known that they were hunters and were attracted to the locality by observations made on hunting expeditions. Such were Nathaniel Buck ‘fa man of great physical strength and endurance, and noted for being an expert hunter and skilled in wood- craft’; Thomas Allen, a deserter from the English army to the side of the colonists, a man of adventurous disposi- tion, fiery temper and obstinate in his opinions; and Ben- jamin Spaulding who had retired from Chelmsford, near Lowell, Mass., to this wilderness, partly to avoid certain financial obligations and partly to trap and hunt.* Ox- ford County was, indeed, as much a frontier of civiliza- tion at the end of the eighteenth century as the Rocky Mountains were fifty years later and attracted much the same sort of adventurers, lovers of untouched nature, the forests and wild animals. It was the sort of blood from which naturalists might be expected to arise, and from Oxford County and from the adjacent counties of Cum- berland and York have arisen such men as A. E. Verrill, of Greenwood, H. C. Bumpus, of Buckfield, and C. O. Whitman, of Woodstock (Oxford Co.); E. S. Morse, of Portland; A. S. Packard, of Brunswick (Cumberland Co.) ; Geo. B. Emerson, of Wells, and Geo. L. Goodale, of Saco (York Co.); and Elliot Coues from just over the boundary at Portsmouth, N. H. Charles Otis Whitman was born December 12, 1843, in Woodstock, Me. As a boy he attended the town schools at Woodstock and Waterford, and worked on the home place. During the summers of 1857 and 1858, while at Waterford, he helped his father’s brother, Elhanan, on the farm, as Elhanan’s daughter, Mrs. H. D. Smith, of Norway, Maine, recalls very well. He was a good-na- tured boy and good company to his cousins. His cousin can not recall that he was engrossed in birds at that time; but by 1860 he had made a considerable collection which 3 Cole and Whitman, 1915, pp. 24-5. IG. 3 F FIG oe r 9 Fre FIG. 4 No. 601] CHARLES OTIS WHITMAN 9 Fie. 2. Charles O. Whitman, at the age of eighteen years. This picture shows a great resemblance to his son, Carroll Whitman, at the time he entered Cornell University. From a tintype contributed by Mrs. E. B. Reynolds (Myra Whitman). FIG Charles O. Whitman at the age of about twenty-one years. Re- produced from gee PSA where it is hemp “1857, enlargement type. Kindness of Professor E. B. Mor This wo teen years. he cousin ai BD. iain: assures me that the correct age is as stated. G. 4. Charles O. Whitman at the age of twenty-five years. This is one of a group d is probably of his regi = bene sree from Bowdoin College. (Note the “class pin.”) Some of the of the gr are full- bearded young men. From a silver er patar by his sister, Adrianna Whitman Evans of Lynn. Fie. 5. Charles O. Whitman at about thirty-nine years. No. 601] CHARLES OTIS WHITMAN 11 Fic. 6. Solomon Leonard, the father of C. O. Whitman’s mother, from a daguerreotype in the possession of Mrs. Myra Whitman Reynolds. Fic. 7. Esther French, wife of Solomon Leonard and mother of C. O. Whit- man’s mother. From a daguerreotype in the possession of Mrs. Reynolds. Fic. 8. arcia Leonard, wife of Joseph Whitman and mother of C. O. Whitman. Note the thick lower ns and sitehtty raised tip of nose. From a photographic print; kindness of Mrs. Reynolds. Fig. 9. Charles O. Whitman, at about twenty years, for comparison with his mother. Note the restricted broad lip and the slightly raised tip of nose. From a tintype; kindness of Mrs. Smith. No. 601] CHARLES OTIS WHITMAN Joseph Whitman (2d), father of via O. Whitman, shia the wide nd w gra nostrils, long tip of nose an y hair, turning man, probably on his return from Germany, about The hair of the head is now quite gray Fig. 10. Fic. 11. Charles O. Whitm 1878, age thirty-five years. Fic. 12. Charles O. alap ern professor at hei cg of Chicago, probably pring of 1900, Age Fig. 13. “The Father,” about 1887. THE AMERICAN NATURALIST [Vou. LI Fie. 14. Handwriting of C. O. Whitman. Above, at Westbury Academy at about the age of twenty-six years. Below, at fifty-six years. he had mounted in two great glass cases, nearly covering one side of the room. Lapham (1882, p. 166) says: So artistically prepared were they, and so naturally mounted, that they attracted much attention among ornithological students. It is said that he was dissuaded from volunteering for the Civil War by his father’s opposition; was drafted but re- No. 601] CHARLES OTIS WHITMAN 15 jected after examination. In the autumns of 1861, ’62 and ’63 and in the spring of 1864 Charles attended the Norway ‘‘Liberal Institute,’’ an academy of high school grade, where his uncle, George F. Leonard, taught. It is said that he taught winters to obtain the means for pay- ing his school expenses. He entered Bowdoin College as a sophomore in Sep- tember, 1865, and was given the degree of B.A. in July, 1868. Lillie attributes Whitman’s strong views in favor of the requirement of Latin for a college education to the emphasis laid in his day on the classics at Bowdoin. He participated in the student literary and debating inter- ests. He ranked about ninth in a class of 23, his rank having been decreased through the necessity he was under of teaching to earn funds for his college expenses. The following autumn he was appointed principal of the Acad- emy at Westford, Massachusetts, and continued in this capacity until the spring of 1872. Here he apparently taught a variety of subjects. In September, 1872, he was appointed sub-master at the English High School in Bos- ton where his uncle, George F. Leonard, had been for some years master. Here he taught general high school sub- jects until the summer of 1875. Morse states that he dis- tinguished himself there by his original methods in teach- ing certain branches of elementary science. ‘While in Boston,’’ says Lillie, ‘‘Whitman came un- der the influence of Louis Agassiz and was one of the 50 stndents who, in July and August, 1873, attended the Anderson School of Natural History founded by Agassiz on the island of Penikese.’’ Here he met Professor E. S. Morse, one of the instructors, who became interested in him because of his drawings, superior even to Morse’s own. Whitman returned to Penikese in the summer of 1874, after which the laboratory was abandoned. The short-lived Penikese laboratory greatly impressed those who attended it and several of these became founders of other marine laboratories—thus, Hyatt of a laboratory at Annisauam, Whitman of one at Woods Hole, Franklin W. Hoone~ of one at Cold Spring Harbor, David Starr Jor- 16 THE AMERICAN NATURALIST [ Vou. LI dan of one on the Pacific coast at Pacific Grove, Califor- nia. It was at Penikese, says Morse, that Whitman seri- ously began his life work in science, for shortly after- wards he went to Dr. Dohrn’s laboratory at Naples and then to. Leipzig where, under the great Leuckart, he learned methods of microtome section-eutting, staining of tissues, and processes of preparation far ahead of those used at Penikese; in short the modern methods of the em- bryologist and morphologist.’’ He was in Germany for three years, being on leave of absence all this time from the Boston high school. In 1878, at the mature age of thirty-five, he received the degree of doctor of philosophy from the University of Leipzig and in the same year ap- peared his doctor’s thesis, ‘‘The Embryology of Clep- sine,’’ in the Quarterly Journal of Microscopical Science. This paper of 100 pages introduced important new prin- ciples, as well as facts, into embryological science and was beautifully illustrated by his own drawings. On his return to America Whitman took up again his work at the Boston high school, teaching English; but at the end of the year he had decided to become a zoologist, applied for and received a fellowship in biology at Johns Hopkins University and resigned from the English high school. But during the summer of 1879 he received an invitation from Professor Morse, who had been teaching zoology at the Imperial University in Japan, to fill his place. Accordingly Whitman set sail for Yokohama, August 21, 1879, and taught at Tokyo until the summer of 1881. Here he trained four investigators who all þe- came professors of zoology in the university: conse- quently he may properly be called the father of zoology in Japan. He and the administration of the university be- came estranged, as he would not adapt himself to what. seemed to him a desire for official control of intellectual property; also he felt that his work was insufficiently sup- ported. As he left he published a somewhat harsh bro- chure entitled ‘‘Zoology in the University of Tokyo.’’ Leaving Japan in August, 1881, Whitman worked at the Zoological Station of Naples as a guest of Professor No. 601] CHARLES OTIS WHITMAN 17 Dohrn, November, 1881, to May, 1882. He was interested in, and recorded, the methods of microscopical research at that station, and published an important paper on the em- bryology, life-history and classification of Dicyemids, which had some time earlier but, as Whitman proved, without warrant been elevated to the position of a sub- kingdom of animals, the Mesozoa. Returning to America in the autumn of 1882 he was ap- pointed assistant in zoology at the museum of compara- tive zoology of Harvard University. While here he worked in cooperation with Alexander Agassiz on the de- velopment of pelagic fish eggs. Two papers were pub- lished on this matter and plates prepared, but no text, for another. At this time he published his book on ‘‘ Methods of Research in Microscopical Anatomy and Embryology.’’ In 1886 Whitman was invited to take charge of a pri- vate laboratory for biological and related research that Edward Phelps Allis, Jr., had planned to found on the Lake at Milwaukee, Wisconsin. With Allis’s cooperation Whitman then started the Journal of Morphology, the first of its kind in America and based on such high ideals of quality and of beauty of plates that the eighteen vol- umes published entailed considerable financial loss, so that finally it had to be discontinued—although it was re- vived, after a lapse of five years, under other auspices. The three years, ’86-’89, at the Lake Laboratory saw the beginning of some researches by Whitman, which were mostly never completed, and by his associates, Howard Ayers, William Patten, A. C. Eyeleshymer, as well by the founder of the laboratory. Some of the published work is of the finest quality. In 1888 the Marine Biological Laboratory at Woods - Hole was organized by certain residents of Boston and Professor Whitman accepted their invitation to become its director. It was in the development of this laboratory that Whitman’s greatest work for science was done. For he introduced and upheld ideals of cooperation and scien- tific democracy which led to its loyal and devoted support by a large body of the working biologists of the country. 18 THE AMERICAN NATURALIST [Vou. LI In the year 1897 a disagreement occurred between Whit- man and a minor, but influential, portion of the trustees. The latter regarded him as extravagant, while he thought they hampered the proper development and organization of the laboratory. Some of the Boston supporters of the ‘laboratory thereupon withdrew from the board; however, the development of the laboratory suffered no serious interruption. At the end of twenty-one years as director Whitman was led to resign, owing to the growing demands upon his time of his researches. It was in the early years of the Marine Biological Labo- ratory that Whitman and some of his co-workers saw the need of a technical zoological society and a call for such a society was made by a committee of which he was chairman. He served as president for the first four years. At first known as the American Morphological Society, it has been known since 1902 as ‘*The American Society of Zoologists,’’ and is still the national society for this science. In 1889, Clark University having been organized exclu- sively for graduate and research work, Whitman was called from Milwaukee to become its professor of zoology. He now took up again his work as a teacher which con- tinued to his death. There gathered about him at Wor- cester, Massachusetts, a small body of devoted zoological investigators. What with the Journal of Morphology, the Marine Laboratory and the organization of the new | department his time was pretty well filled. He published three short papers on the leech, Clepsine. But affairs at Clark were not altogether to his liking and when a call to the University of Chicago came in 1892 Whitman and the heads of the departments of physics, chemistry, anatomy, neurology and paleontology seceded from Clark Univer- sity and formed the major portion of the scientific faculty of the new university. Here at Chicago Whitman had more graduate students than ever before, but as his re- search grew more engrossing he kept more and more to his home where his experiments were conducted. For a period of fifteen years he bred pigeons to get at an under- No. 601] CHARLES OTIS WHITMAN 19 standing of the evolution of their color markings. He usually had about 500 individuals, representing about 40 wild species, in dovecotes surrounding his house. He hybridized nearly 40 species, most of which had never been crossed before. His work included the phylogeny of pigeons, instinct and animal behavior, voice, fertility and the nature of sex. Professor Whitman suffered considerably from indi- gestion for some years before his death. He contracted a heavy cold while caring for his pigeons and died sud- denly, December 6, 1910, at the age of 67, of pneumonia, the same disease that his mother died of at the age of 57. We have now to consider some of the personal charac- teristics of Charles O. Whitman, and their distribution in his family. First, white hair. Charles Whitman’s hair turned white unusually early. Morse quotes Judge Clarence Hale of Portland as saying that his hair was perfectly white when in college but his class portrait (Fig. 4) shows his hair still dark. At the age of 39 (Fig. 5) the hair is quite white. His father’s hair turned gray early ; Charles’s brother at 49 has very gray hair. Early graying is ap- parently a ‘‘ Whitman trait.’’ 2.. Wave in hair is shown by his father (Fig. 10) and as a dominant trait is doubtless inherited from this side. 3. Nostril. His father had broad nostrils (Fig. 10), but the ridge is carried apically beyond the level of the nos- tril. In Charles, as in his mother (Fig. 8), the apical knob was lacking, hence the nostrils appeared unusually full | 4. The lower lip is thick, especially near the median plane. His mother’s portrait (Fig. 8) shows an excep- tionally broad lower lip and this is exactly reproduced in her daughter, Adrianna. The mother’s father (Fig. 6) also has a broad lower lip. Charles gets the breadth of lower lip from his mother’s side and its restriction to the middle of this lip (Fig. 9) is doubtless due to another factor. 5. The forehead is broad and full (Fig. 12) ; doubtless a 20 THE AMERICAN NATURALIST [ Von. LI Whitman trait, as it is strikingly reproduced in his cousin, Mrs. Smith. Moreover his grandfather Leonard had an exceptionally fine head (Fig. 7). 6. The general slenderness of form is a Leonard trait. His father and some of his sibs were stout, as is one of his sisters, while his other sibs are slender, like himself. Taste for Natural History.—This was, doubtless, in- nate and it was very strong, developed very early, as we have seen above, and led to his keeping animals and mounting birds. Thus he kept, as a small boy, mud turtles, rabbits, guinea pigs, woodchucks, and raccoons. Later, his sister recalls, he kept two gray squirrels which were so tame that they would go to sleep in his pockets. He kept doves, which he would tend carefully and of which he would watch the young after they were hatched. These doves would cover him as he fed them. He was fond of hunting. His stuffed eagle was especially famous. On one occasion, at the age of twenty-four years, he was gone two nights and a day after an owl. Mr. H. T. Libbey, . of Bryant’s Pond, told me that on one occasion he went out with Charles after a large bird and was gone all day. It is interesting that the special objects of interest of his early years should have been those of his late years. If we seek for the origin of this strong taste in such an out- of-the-way part of the world we find it, as stated above, in the blood of the pioneer, adventurer, hunter and lover of nature which sought the wilderness of Maine at the end of the eighteenth century. Actually, one finds evidence of love of hunting and of nature in close relatives, especially on the father’s side. First, Charles’s sister, Adrianna, was as fond of natural history as himself. She had a tame bluejay which stayed about the house like a crow. Charles’s uncle, Chauncey C. Whitman, was a farmer who devoted himself, in his later years, particularly to the raising of horses and cattle. Itis told of him‘ that, when he was twenty-four, he joined his fellow townsmen on a bear hunt and finally the bear was found in his den under the roots of a large tree. Chauncey ‘‘had the temerity to 4 Lapham, 1882. No. 601] CHARLES OTIS WHITMAN 21 thrust in his arm and catch hold of her hair, when she quickly turned and came out before them all.” She had two cubs. One was taken out alive and kept for several months by Chauncey Whitman until, having become mis- chievous, it was killed. Here we see the same love of wild animals that Charles showed. This Chauncey had a son, Albert, who is a hunter living in Oxford County. He keeps squirrels, woodchucks, and other pets. Al- bert’s brother, Thomas, had a poultry farm and his son still carries on the business. Charles’s father’s father, Joseph Whitman, father of Chauncey, was also a hunter and told his grandchildren hunting stories, including ad- ventures with bears and other animals, which are keenly remembered to this day. It appears that the mother’s father, Solomon Leonard, had a love of nature and took pleasure in his trout brook. So Charles got his love of nature from both sides of the house. Scholarship.—But Charles had more than a love of na- ture; he desired to study nature. This scholarship is a family trait and comes to him from both sides of the house. It is noteworthy that in Lapham’s ‘‘History of Woodstock, Me.,’’ 1882, p. 56, it is stated: Three Woodstock men have graduated from college, George F. Leonard, Harrison S. Whitman and Charles O. Whitman. The first is an uncle and the second a cousin of the third. According to Cole and Whitman (1915, p. 707): Judge Mitchell’s “History of Bridgewater” says that more persons of this (Whitman) name in that town in early times received a college edueation than any other. The next most numerous were Packards. And Charles’s father’s father’s mother was Abigail Pack- ard. It is worth while to consider in detail the direction taken by the scholarship of the close relatives of Charles. First, the mother’s father, Solomon Leonard, was a close student of ancient history. But it was to the influence of his mother’s brother, George F. Leonard, that the direc- tion of Charles’s early life was chiefly due. George Leon- ard graduated from Dartmouth with the class of 1859. He adopted teaching as his profession and instructed in the academies. in Norway and Paris, in Maine, and in Northbridge, Mass. He was 22 THE AMERICAN NATURALIST [ Von. LI also a teacher in the English high school in Boston, following this occu- pation for over twenty years. ... He is a profound scholar and suc- ceeded well in his teaching. It will be noted that Leonard was teaching at the English high school, 1862-’82, at the time Charles O. Whitman was there and had preceded him there probably ten years. Leonard was especially interested in mathematics and during the later years of his life devoted much time to ‘squaring the circle,’’ i. e., to determining if the relation 7 can be exactly expressed by a fraction. He is said to have worked out the decimal to over 1,250 places. Relatives say that George Leonard helped secure a college educa- tion for Charles and his influence must have been great in getting his parents to let Charles go instead of, as eldest son, helping his father with his business. Leonard graduated, 1859, and taught, probably 1859-’62, at the academy in Norway and that at Paris only a mile or two away; Charles studied at the Norway Academy, 1861—’64. George Leonard, as we have seen, was already a veteran at the English high when Charles came there to teach, re- taining his connection with it for seven years, and during this entire period his uncle taught there. Another instance of scholasticism in this family is that of Charles’s first cousin, Rev. Harrison Spofford Whit- man, who was born 1844, a year or two later than Charles. Harrison is the son of Harrison Whitman, who was a farmer, at one time captain of an infantry company of Woodstock that saw some service at the time when war was threatened over the boundary of Maine; he was also some time coroner of Oxford County, and died at the age of thirty-one years, leaving a widow and three children. Harrison, Jr., early showed remarkable aptitude in com- position (as did his brother and sister); he wrote both prose and poetry; he was fond of study, graduated at Bowdoin in 1869, teaching school meanwhile to earn his tuition and expenses, was for two years principal of the Thomaston Academy, ’69-’71; then taught mathematics and later classics at the Dean Academy, Franklin, Massa- chusetts, ’71-’74; then he studied theology at Tufts Col- No. 601] CHARLES OTIS WHITMAN 23 lege, graduating 1877. He has ever since taken a leading position in the Universalist ministry of Maine, having been pastor at Augusta and being now at Portland, Maine. He is an able preacher and popular among his parish- ioners. This brief account of close relatives who have been dis- tinguished scholars makes it clear how Charles O. Whit- man found it natural to follow a scholarly career. Thoughtfulness and Classicism.—Not all scholars are of the same type. In one type (the hyperkinetic or ro- mantic) there is a rich flow of brilliant ideas and a rapid passage from one subject of interest to another. In the other type (the hypokinetic or classic) there is a pro- fundity of consideration of a subject and a persistence of interest in it. Charles O. Whitman belonged to the latter type. This type ordinarily shows a recessive in- heritance. His mother apparently showed this type. She was gentle and pleasant and never lost her even tem- per; while in many of the Whitmans the temper was of the periodically explosive sort; but there is evidence of the classic temperament on the paternal side, e. g., in Har- rison S. Whitman. That Charles Whitman was of the classic type and was thoughtful will be generally conceded; he was, indeed, one of the best examples of this type that one could find. He felt little pressure to express himself. His principal biographer (Lillie, 1911) records only 67 titles of which 7 are his annual reports,—reports that, toward the end of the series, were secured only, with much difficulty and after long delay, and for the last twelve years not at all. Of the remaining 60 there are hardly 20 that are to be classed as typical professional papers, giving the final results of finished observation. Some of the papers are brief notices of technical methods (Methods in the Zoolog- ical Station in Naples, 1882, with a French translation ; treatment of pelagic fish eggs, 1883; means of differen- tiating embryonic tissues, 1885; osmice acid and Merkel’s fluid, 1886). Other papers are polemical (new facts 24 THE AMERICAN NATURALIST [ Vou. LI about the Hirudinea critique of Apathy, 1888; Apathy’s grief and consolation, 1899). A group of papers (9) of a semi-popular sort relate to the work and aims of the Biological Laboratory. Most of the remaining non-technical papers are delightful essays, chiefly upon philosophical biological matters. Such are: “The Seat of Formative and Regenerative Energy,” 1887; ‘‘The Naturalist’s Occupation,’’ 1891; ‘‘The Inade- quacy of the Cell Theory of Development,’’ 1893; ‘‘Gen- eral Physiology and Its Relation to Morphology,’’ 1893; ‘Evolution and Epigenesis,’’ 1895; ‘‘Bonnet’s Theory of Evolution; a System of Negations’’; also ‘‘The Palin- genesia and the Germ Doctrine of Bonnet,’’ 1895; ‘‘ Ani- mal Behavior,’’ 1899; ‘‘Myths in Animal Psychology,’’ 1899. The more strictly investigational papers fall into three periods: (1) The invertebrate period—devoted chiefly to the leech, Clepsine, which was the subject of his doctor’s thesis and upon which he wrote more or less from 1878 to 1899—a period of 21 years. Here also belongs his Naples work on Dicyemids. (2) The period of vertebrate em- bryology, beginning with work done with Alexander Agassiz on pelagic fish eggs, 1883-1889, on amphibian eggs, 1888, and the ganoid fish, Amia, 1896. (3) The period of genetics, foreshadowed in his note ‘‘Artificial Production of Variation in Types,’’ 1892, and continued with the pigeons to the end, 1910, in all eighteen years. In his work with worms, amphibians and pigeons he was led to reflections upon animal psychology and to the pub- lication of his classic paper on animal behavior, 1898, and his brief paper on myths in animal psychology, 1899. Many of these papers are highly finished and give the results of prolonged contemplation. Professor Whit- man. especially in his later years, repeatedly spoke dis- dainfully of rushing into print and making an annual ‘‘dump’’ of scientific gleanings—and these were the nat- ural expressions of his own nature; perhaps he insuffi- ciently recognized that all persons were not constituted like himself and could not react in the same way. Whit- No. 601] CHARLES OTIS WHITMAN 25 man, indeed, planned to publish far more than he did; he had accumulated materials that were nearly ready for the printer. But the absence of the hyperkinetic drive com- bined with the manifold duties of the moment led to pro- crastination for the more convenient period of prolonged quiet which an overactive world never afforded him. On the other hand, had it not been for the exceptional pres- sure brought to bear on him for an address or a contribu- tion we might have had less than we have from his pen. Of his writings Lillie truly says (p. lxxiii): His published papers, mostly short, are models of condensed thought, written in a fine, polished, characteristic style. No less care was de- voted to the form than to the substance, and some of his papers will certainly endure as classics of the biology of his time. . . . He rarely had occasion to correct any published statement, and even less rarely, perhaps, to change in any radical way a point of view to which he had once committed himself. Conservatism, so frequently associated with hypoki- nesis, was marked in Whitman. He was not very cordial to developmental mechanics, and was critical of the en- thusiastic rush to the mutation theory and Mendelism. He could not easily abandon old ideas for new, and ally himself with the latest biological fad. A strong philosophic and argumentative tendency is found on both sides of the house. His mother’s father was much given to theological discussion, and was an ar- dent adventist. Charles’s father, too, was argumenta- tive. The philosophic tendency in Whitman was marked in his writings. In one of his manuscripts of Westford days, at the age of 26, he discusses the topic ‘‘Progress Has No Goal” and again ‘‘Womanhood Suffrage,” of which the tenor is shown by the concluding sentence: Female suffrage . . . may meet with opposition, as indeed, every re- form does, but all this opposition is but the alarm of the great ¿lock of human progress, which is soon to strike the hour when all enlightened nations shall recognize not only manhood but also womanhood suffrage. This is quoted as an illustration of youthful style and not a statement of his views in later life. Whitman’s hypokinesis shows itself even in his hand- writing. It altogether lacks the running dash of the 26 THE AMERICAN NATURALIST [ Von. LI hyperkinetic. Each word is worked out with some effort. It is interesting to compare his chirography of Westford Academy days (when he was about twenty-six) and that when he was fifty-six (Fig. 14). There is the same loop from the end of the word to cross the ‘‘t’’ in ‘‘the’’; the same scanting of the terminal ‘‘y’’; the same form of the capital ‘‘I.’? In thirty and indeed in the course of forty years his handwriting showed no important change. Deliberateness.— Professor Whitman’s movements and speech were characterized by deliberateness—another characteristic of the classical type. It is, of course, a mere caricature to say, as an unkind critic once did, that his lectures consisted of pauses punctuated by sentences. I mention this because it brings vividly to mind a way he had in lecturing or addressing his seminar students to pause frequently for some seconds looking pleasantly over the room before beginning the next sentence. Even in conversation he would turn a calm, thoughtful face to- ward you and express himself clearly and deliberately. So marked a deliberateness is not shown by his sibs, but his brother shows something of it, and I am told that his father was slow of speech. Literary Ability.—While Whitman did not have a strong internal impulse to write, what he published is mostly characterized by high literary finish. Speaking of cooperation between the organic and the inorganic sci- ences he says (1895, p. iv): Comparison of standpoints must benefit both sides. Cross fertiliza- tion works rejuvenation in theories as in organisms. The biologist may pause to see how the individual vanishes in the abyss of the universal, and how self determination dissolves in the pressure of the physicist’s fundamental postulate of inertia. The physicist may find it agreeable from time to time to turn from the Nirvana where self and not self, rocked ‘in blissful reciprocity of vibration, annul each other, to the world where self asserts itself in organie determinations, issuing in pur- poseful adaptations and eonscious, intelligent action. Again, speaking of Bonnet he says: With a zeal never daunted, and an ingenuity seldom baffled, never defeated, he piled mountain upon mountain of negation, rolling Ossa upon Olympus and Pelia upon Ossa, until the whole organie world No. 601] CHARLES OTIS WHITMAN 27 seemed to be completely buried under a stupendous mass of negations, blinding in one infinite negation—No Change. Perhaps something of Whitman’s interesting and vivid style may be referred to the influence of his teaching: of English in the high school; but much of it appears in papers written before the Boston days. In them, too, he frequently uses the rhetorical question so often found in his later writings (cf. Fig. 14, upper). I think we must conclude that this literary ability has a constitu- tional basis. We have seen above that his cousins, chil- dren of Harrison Whitman, ‘‘early showed a remark- able aptitude for composition both in prose and poetry.’’ A granddaughter of Chauncey Whitman, brother of Charles’s father, was an authoress of poetry, for which she found a market. Pertinacity.—In Whitman’s combination of traits was found an element of pertinacity that was at times very formidable. Had it been less he could hardly have suc- cessfully overcome the handicap of comparative poverty, despite which he went through college. It showed itself again when he insisted in Japan that his student’s papers should be published under their own names: and, when, since he was overruled, he resigned his professorship. It showed itself again in his struggle with a minority of the trustees of the Biological Laboratory, in which his views prevailed. It showed itself still again in his relations with Clark University, which led to his acceptance of the offer to go to Chicago. In minor departmental and labo- ratory affairs, as his colleagues well recall, this gentle- mannered man would show at times uncommon resistance to suggestion and persuasion. One, therefore, learns with interest that his father also was set in his opinions and could not readily be made to change his views. Indeed, Charles is seen to be an interesting mixture of gentleness, as shown also by his mother, and tenacity of purpose, as shown also by his father, his mother’s father and other members of the family. Closely akin to his insistence on his ideals was his uncompromising disposition. As Lillie well remarks, _ It is questionable whether his life would have been so valuable, had : his disposition been more pliable. 28 THE AMERICAN NATURALIST [Vou. LI One other trait that sometimes showed itself, especially in his writings, was his capacity for trenchant and satir- ical criticism. Morse (1912, p. 283) has given some ex- amples. It is characteristic of many hypokinetics to feel deeply and to resent warmly. In weighing any criticism we must always consider the personality of the critic. Perhaps in this reaction of Whitman’s we see trace of a Whitman ‘‘sternness’’ shown also by his father. Artistic Taste.— Professor Whitman had a keen artis- tic sense. This is proved by the success, in his boyhood, of his mounts of birds. I think we may go back of this and find evidence of an artistic sense in the appeal made to him by natural beauty—the beauty of forests, flowers, birds and beasts. To a person without the sense of beauty natural forms have little attraction. Later this sensitiveness to and love of form shows itself in the beau- tiful drawings he made at Penikese, the exquisite plates of leeches and the delicate pencil drawings of the Dicy- emids. It is clear that the art of Japan appealed strongly to him and he had in his house at Chicago many examples of that art. It was love of form that made him a mor- phologist and led him in the Journal of Morphology to in- troduce a beauty of execution of plates exceeding any- thing then current in America. It was this sense of beauty that led him to select excellent Japanese artists to draw and paint his pigeons. While I have not been able to make an exhaustive study of the distribution of artistic sense in his relatives, it appears that his father, who was a carriage manufacturer, was an artistic one. The wood- work of the finish was done very carefully, the wheels made by hand and so artistically was the whole executed that he once received a silver cup as prize for one of his carriages at a competitive exhibition. On the other side we find his mother’s father, Solomon Leonard, was also an artisan. He established an iron foundry at Pinhook and his kitchen ware was so satisfactory that ‘‘the name of Solomon Leonard was known in every household.’’ When later he retired from business he maintained ‘‘a small furnace at Bryant’s Pond, where he made small castings to pass away the time.’’ So it seems probable No. 601] CHARLES OTIS WHITMAN 29 that a taste for the beautiful in form and color had a con- stitutional basis in Charles. : Musical ability is frequent in the Whitman side of the house. Nearly all of the Whitmans could sing; espe- cially good were his father’s brothers. But Marcia Leon- ard, the mother, could not carry a tune; and the children were apparently not good singers. Charles tried to learn to play the melodeon when he was a boy, but had no apti- tude for it; his elder sister, however, plays well on the piano and her son is a performer on the violin, of great ability, and a brother of Charles plays the violin well. Of mechanical ability Dr. Whitman had more than the average man. His success in mounting birds (in his father’s carriage shop) required manual dexterity. His sister recalls that he made a martin house. This ability shows itself again in his interest in methods of micro- scopical research which culminated in a volume—the only book he ever wrote—entitled: ‘‘Methods of Research in Microscopical Anatomy and Embryology,’’ 255 pp., Bos- ton, 1885. Of this book 55 pages are devoted to instru- ments and methods of imbedding, including 28 figures of machinery and apparatus, and pages of descriptions of the principles and details of this apparatus. It is im- probable that Whitman would have collected such ma- terials and written such a book had he not been interested in and had an insight into mechanical devices. This taste he maintained to the end, and he tarried long over a stu- dent who had invented some new instrument.” This me- chanical insight enabled him to do things in ship-shape fashion. His charts and general arrangements of his laboratory and breeding pens showed this savoir faire. With less mechanical skill the feat of moving his pigeons to Woods Hole and back each year would hardly have been feasible. Whitman came from a race of artisans on both sides. His father was a carriage maker; his father’s sister Loney had a son, Edwin, who is a machinist and an- other, Edson, who is overseer in a fabric mill. One of his uncle Cyprian Whitman’s sons is an engineer. Of Elhanan Whitman’s sons, Austin was a farmer with more 5 Dr. Osear Riddle. , 30 THE AMERICAN NATURALIST [Vou. LI than average mechanical ability and Edgar was a carpen- ' ter and always ingenious. Even Elhanan’s daughter, Mary, says she is handy with tools and so are all of her sons. On the mother’s side we have doubtless mechan- ical interests in at least the grandfather, Solomon Leon- ard, the iron founder. The social instinct was highly developed in Whitman. While he rarely appeared at social functions and seemed to shrink from exercising the presidential office to which learned societies repeatedly elected him, yet he liked the society of his younger colleagues, frequently had infor- mal lunches or dinners and, after the excellent meal, would encourage the guests to sit and talk on the matters that were uppermost in the minds of such young and ardent students of biology. Given a young man of good health, fairly ambitious, dogged, scholarly, gentle, friendly, philosophical, conserv- ative, with an ardent love of natural objects, literary and artistic ability of a high order, and with a savoir faire and place him in an environment of a scholarly uncle who is in a position to assist his nephew; of the stimulus of magnetically charged biological atmospheres at Penikese and at Leipzig; of an epoch when new machinery and new methods are being rapidly introduced into a new science; of a period in a young science that demanded organiza- tion of laboratories, university departments, societies and journals; and the reaction is that of the propositus; even though handicapped (?) with a hypokinetic tempera- ment, an uncompromising attitude and, in his earlier years, an occasional tendency toward caustic criticism. At any rate such was the man, such the environment that provided the stimulus, and such the life history of Charles Otis Whitman. LITERATURE CITED Cole, Alfred and C. F. Whitman, 1915. A History of Buckfield, Oxford County, Maine. Buckfield, Me. 758 pp. Farnam, C. H., 1889. Genealogy of Whitman Family. Lapham, W. B., 1882. History of Woodstock, Maine. Lillie, F. R., 1911. Charles Otis Whitman. Journal of Morphològy, XXII, pp. š Č; Morse, E. S., 1912. Biographical Memoir of Charles Otis Whitman, 1812- 1910. Nat. Acad. of Sci. Biographical Memoirs, Vol. VII, pp. 269-288. MENDELIAN FACTOR DIFFERENCES VERSUS REACTION SYSTEM CONTRASTS IN HEREDITY! T. H. GOODSPEED AND R. E. CLAUSEN Durine recent years there has been a remarkable ad- vance in our knowledge of Mendelian principles of hered- ity. This advance has for the most part had its source in the important and fundamental work of Morgan and his associates (1915) in which they have been concerned with the mutations of the common fruit fly, Drosophila ampe- lophila. The results of other work, in so far as agreement has permitted, have been brought into harmony with the principles arrived at through these investigations. As a result there has been developed a fairly clear and compre- hensive conception of the constitution of the hereditary. material and the nature of the mechanism by which it is distributed in gametogenesis, a conception which fur- nishes a consistent explanation of the products of Men- delian studies. Morgan has stated that the fundamental principle of Mendelism may be reduced to this, that the units con- tributed by two parents separate in the germ cells of the offspring without having had any effect on each other. This conception of the absence of any factorial variability _ save that concerned in the discontinuous changes in fac- tors involved in mutations has furnished the working hy- pothesis for Morgan’s brilliant analysis of the germ plasm of Drosophila. Although the results of the Droso- phila investigations have been ably presented elsewhere (Morgan et al., l. c.) it seems well briefly to review them here especially as they are vital to the argument presented in this paper. 1The investigations herein reported have been assisted by that portion of the Adam’s fund granted to the department of botany by the depart- ment of agriculture of the University of California. | a by 4 THE AMERICAN NATURALIST [ Vou. LI The large amount of data which Morgan and his asso- ciates have collected within the past six years has clearly demonstrated that the chromosome mechanism furnishes a basis upon which the behavior of Mendelian units may be logically and consistently explained. In an investiga- tion of Drosophila over a hundred factor mutations have been discovered and studied, and these have been found to fall into four groups with respect to the linkage rela- tions they display with one another. These four groups correspond to the four pairs of chromosomes. By deter- mining the linkage values which are displayed within groups it has been possible to demonstrate that there is a consistent, invariable, linear arrangement of factors within the chromosomes at some time in their history. From this data Morgan and his associates have been able to prepare a map of the relative positions of the factors in the chromosomes. The complete conception, therefore, pictures the chromosome at some stage in its history as a linear series of loci. Whena change occurs in some locus, a corresponding change of some sort may occur in soma- togenesis, so that the individual which develops from such a set of factors with the changed locus differs in some particular way from an individual which develops from the normal unchanged series of loci. The change in the characters of the individual will depend not only upon the particular locus which has been changed, but also upon the particular way in which that locus has been changed. A changed locus, however, maintains the same position. with reference to the other loci as did the unchanged locus, and this fact is the basis of Mendelian behavior, for knowing the behavior of the chromosomes in reduction, it enables us to gain a clear conception of the nature of Mendelian segregation. When now we consider the particular factors them- selves, the changed loci of the system, we see clearly that important physiological relations exist among the various loci. It is an appreciation of this fact that has led many investigators, among them Conklin (1908), Jennings (1914), Morgan (1915 b), Pearl (1915), and Wilson (1914), No. 601] § MENDELIAN FACTOR DIFFERENCES 33 to insist that the factors can not be regarded as deter- miners in themselves; but rather that they are differenti- ators, that working together with other factors in the system a difference is produced in somatogenesis which has its origin in some difference, some change in a locus in the system. For when in Drosophila a change in the locus W is produced, or in Y, such that the individuals developing from systems with these changed loci are white-eyed in the one case and yellow-bodied in the other, it seems evident that the change is more profound than the color of eyes or of body; that beyond these changes there is an underlying, elusive, physiological change re- sulting in individuals that are less vigorous and less fer- tile than those which develop from the normal unchanged system. The fact that a factor may have a primary, simple, easily recognizable effect and secondary far reach- ing effects, the latter to be attributed to the modified physiological relations resulting from a change in one of the members of a system, is one which has often been ob- served and which is of fundamental significance in our conception of the interrelations of the genetic factors. There are, however, other instances which may be cited of a somewhat similar nature, locus changes which pro- duce certain characteristic effects under particular en- vironmental conditions, but fail to disturb the normal behavior of the factor system when these conditions are not met. It would, of course, be possible to recount almost in- definitely the specific effects of particular locus changes, whereas evidence concerning these far reaching effects of single factors is too scanty to warrant further discussion of this point. However, Morgan (1915 a) has been able to establish the relations displayed by the factor for ab- normal abdomen and to demonstrate that only under very particular conditions is the presence of this genetic factor manifested by its characteristic expression, and that when these conditions are not present the product of somato- genesis may not differ in appearance from the normal fly, although differing from it both in genetic constitution and 34 THE AMERICAN NATURALIST [Vou. LI hereditary behavior. We have thus a factor here which evidently has such a relation to the other members of the factor system that only under peculiar environmental con- ditions does it disturb the normal course of somato- genesis. Miles (1915) has likewise investigated a type of chlorophyll reduction in maize in which the recessive forms are yellow seedlings which usually show a dis- tinct greenish tinge at the tips of the leaves. This type of chlorophyll reduction displays normal Mendelian be- havior in inheritance giving in the progeny of hetero- zygous plants a ratio of approximately three seedlings which are of the normal green coloration to one which is of the yellowish type. The heterozygous plants possess the normal depth of coloration and can not be distin- guished from those which are homozygous for the produc- tion of normal chlorophyll coloration. The yellow seed- lings on the other hand form a distinct and easily recog- nizable class with no tendency toward intergradation, an observation which we have ourselves been able to con- firm in independent mutations involving this locus. Usu- ally these seedlings die as soon as the food material in the endosperm is exhausted, for under ordinary conditions the change in the locus is incompatible with a normal de- velopment of the individual, it is too profound an altera- tion to give a normally functioning factorial system. Miles found, however, that when the yellow seedlings were grown under particularly favorable conditions, they developed a normal chlorophyll coloration and produced plants which were able to go on through the cycle of changes included in the normal development of the maize plant. Now this behavior can not be referred to any change of the reduction locus back to the original condi- tion, for the progeny of such plants consisted entirely of yellow seedlings; indeed, such a reversion would be incon- ceivable. Obviously the explanation of the situation will be found only through a consideration of the system with the recessive reduction locus. Normally this unchanged locus performs a definite function in determining the pro- duction of chlorophyll in the plant, but this function is No.601] § MENDELIAN FACTOR DIFFERENCES 35 performed in conjunction with a number, perhaps a very great number or even all, of the other loci within the system. With a change in this particular locus, how- ever, comes a change in the normal course of events in chlorophyll production, in that the rate at which the sys- tem is able to produce chlorophyll has been altered. Nevertheless, this change does not completely prevent the system under favorable conditions from going on and ultimately developing the same reaction end product which would have developed in the normal unchanged condition, but more rapidly. Among such factors as have a profound influence upon the interrelations within the systems of which they area part are those which Morgan (1914) has called lethals. Morgan’s work with lethals is particularly suggestive because he has been able to demonstrate that they, like other normal Mendelian factors, occupy a definite locus in the chromatin system and display the same perfectly defi- nite and consistent behavior with reference to the other loci of the system as do all other changed loci which do not interfere with the normal development of the indi- vidual. It is entirely possible that some of the lethals, like the chlorophyll reduction locus which we have dis- cussed above, may yield systems which occasionally per- mit of the normal development of the individual, at least certain peculiar sex ratios which have been obtained might indicate that fact (Morgan, l. c.) ; but the important result of these investigations with lethal factors lies in that fact that certain kinds of changes in some loci are incompatible with normal functioning of the chromatin system. It might in addition be noted that there seems to be no particular reason why we should not include in the same category with lethals, the type of chlorophyll reduc- tion mentioned above and those other types in maize which result from such profound factor changes that no development is possible after the food supply of the endo- sperm is exhausted. Now giving the above results their broader and more general interpretation, it would appear that the factors 36 THE AMERICAN NATURALIST [ Von, LI make up a reaction system the elements of which bear a more or less specific relationship to one another. It is this specific interrelation of the factors of the reaction system which determines that wheat produces wheat, and corn, corn, and so on through the whole realm of living matter. With this in mind it is at once apparent that normal Mendelian behavior can not be considered as a contrast of different reaction systems, but that in such cases the two organisms contrasted must possess funda- mentally the same reaction systems, only a relatively few elements within the reaction system differing, and these not in a fundamental fashion. Jn fact it seems entirely logical in the light of modern Mendelian developments to consider each particular locus as made up of a definite nucleus, some complex organic compound perhaps, with a number of end chains which may be altered in various ways without changing the structure of the nucleus of the locus. According to this conception the fundamental rela- tion of the locus to the other elements of the reaction sys- tem would remain unchanged, while the end product would be altered in some particular manner. There is probably no more striking confirmation of this conception than the suggestive hypothesis of multiple allelomorphs which Morgan and his associates (1915, J. c.) have developed. Their results and those of others in this connection seem to show clearly that the explanation of Mendelian differ- ences on the basis of such a profound change as the drop- ping out of an element from a delicately balanced reaction system is practically out of the question. In multiple alle- lomorphs we have not one, but several, changes within the same locus. The similar effect which these changes have on certain organs of the body, for example, that relation shown in the locus W in Drosophila as a consequence of which the normal red eye color may be changed to white, eosin, or cherry depending upon a particular change in the locus, are such as to indicate that these are probably changes around the fringe of the molecule and not such as fundamentally affect the structure of the entire locus. Moreover, the relations thus exhibited again indicate that No. 601] MENDELIAN FACTOR DIFFERENCES 37 the locus has a particular place and function in the reac- tion system, that it bears a specific relation to the other elements of the system. It is true that investigation in plant breeding has not yet progressed far enough to furnish a definite confirma- tion all along the line of the work with Drosophila. The reasons for this are very obvious and they lie in the tech- nical difficulties involved in such work rather than in fun- damental disagreements in principle. It has not as yet been possible to study as many factors in any plant spe- cies as have been investigated in Drosophila, nor to carry the work through as many generations nor to employ as large populations. Moreover, most plant material is more difficult to handle from the standpoint of a chromo- some analysis on account of the longer period of time necessary to secure data and the greater amount of atten- tion which must be given the cultures and the larger num- ber of chromosomes which are usually involved in such species. It must also be borne in mind that practically all the Drosophila differences have arisen under observation as simple factor mutations, and it has therefore been rela- tively easy to determine their relations to the other fac- tors in the system. In plants, on the other hand, the ma- terial has presented itself as a confusing array of varieties containing for the most part a large number of recessive factors, and usually the original form from which they were derived, corresponding to the normal fly of Morgan’s work, has not been obtainable and would not have been varticularly useful, had it been available. Nevertheless, there appears to be no real difficulty in the way of accept- ing the conception derived from the Drosophila studies as a definite, consistent working hypothesis; for it is diffieult to believe that the behavior of plant material should be fundamentally different, and indeed points of correspond- ence are not lacking to warrant us in viewing somewhat sceptically any undue emphasis placed upon the differ- ences which may seem to obtain. We may now return to the conception of the reaction system as a unit in itself in the sense that it is made up of 38 THE AMERICAN NATURALIST | Vou. Li a large number of elements which bear a more or less spe- cific relation to one another. This is the important physi- ological conception which has grown out of the vast amount of work which has been done in recent years in the analysis of the hereditary material. This is no new contention; it has been advanced and ably advocated by many investigators, but we feel that certain consequences of this conception have not been given the consideration their importance deserves. For if this conception be valid then it should not be possible, in certain cases at least, to shift and recombine the elements from which systems have been built up in the haphazard way that some advocates of Mendelism have attempted to do. If, for example, it is possible to obtain hybrids involving not a contrast between factors within a single system, but a contrast of systems all along the line, then it is obvious that we must consider the phenomenon on a higher plane, we must lift our point of consideration as it were from the units of the system to the systems as units in themselves. Our attention has been called to this extension of the Mendelian conception by the behavior of species hybrids of Nicotiana which have been studied at the University of California during the past six years. This study has . been concerned particularly with hybrids between N. syl- vestris and varieties of N. Tabacum. These species, the former represented in the collections of the University of California Botanical Garden by a single type and the latter by a considerable variety of distinct forms, belong to entirely distinct sections of the genus Nicotiana and differ in important particulars which have been described elsewhere (Setchell, 1912). Goodspeed (1913) has studied a large number of different reciprocal hybrids between sylvestris and various of the distinct varieties of Taba- cum. These hybrids are all partially sterile. It is pos- sible to obtain a few viable seeds from open pollinated flowers and from those pollinated with Tabacum and syl- vestris, but it has never been found possible to obtain any selfed seed. The phenomena displayed by these hybrids in development and inheritance admit of a consistent ex- No. 601] | MENDELIAN FACTOR DIFFERENCES 39 planation, if we regard them as the outcome of a contrast of two distinct Mendelian reaction systems the elements of which can not be freely interchanged without pro- foundly affecting the general functions of the reaction systems thereby resulting. We shall take up in very gen- eral fashion the points which have inclined us to this view, reserving for a later treatment the discussion of the hy- pothesis in detail and also the presentation of the ex- tended data. When hybrids are obtained between sylvestris and the various varieties of Tabacum they agree throughout in F, in presenting the entire set of characters of the Taba- cum parent to the exclusion of those of sylvestris. This behavior may be definitely accounted for as a dominance of the Tabacum reaction system as such over the sylves- tris reaction system. For point by point and character by character throughout, the correspondence between the Tabacum variety and its F, sylvestris hybrid may be demonstrated in a remarkable fashion and this irrespec- tive of whether the factors concerned in these character expressions in the Tabacum varieties are dominant or recessive in varietal crosses. This correspondence is not only apparent in general appearance, but it extends to minute details of form and structure, and it is displayed even in the more intangible characteristics generally in- cluded under the term habit—i. e., such characteristic varietal peculiarities of expression as the method of branching, insertion and inclination of the leaves, the type of inflorescence, and so on through a whole series of de- tails. For example, when N. Tabacum var macrophylla (Setchell, Z. c., p. 8) is the Tabacum parent, the F, hybrids display the particular appearance and also the particular characteristics of macrophylla, except that throughout they are expressed on a very much enlarged scale (cf. East and Hayes, 1912). The broad clasping leaf of mac- rophylla with its distinctly pointed tip is faithfully repro- duced in the hybrids. The flowers show no effect of the very much elongated corolla tube and the lobing of the limb peculiar to sylvestris, but display the macrophylla 40 THE AMERICAN NATURALIST [Von. LI proportions throughout. The stout tube, swollen infun- dibulum, and pentagonal limb are clearly derived from macrophylla, and the color is rose red of approximately the same depth and tone as that of macrophylla and in striking contrast to the pure white of sylvestris. In habit the hybrids resemble macrophylla. In early growth they are not characterized by the long maintained rosette which is so characteristic of sylvestris, and in leaf distri- bution, branching, and type of inflorescence they again correspond to their macrophylla parent. When an entirely different set of characters is con- cerned in the Tabacum parent the F, hybrid with syl- vestris is still an exact replica of the particular Tabacum used. For instance N. angustifolia (Setchell, l. c., p. 9) and sylvestris give a hybrid which is entirely different in general appearance and all details from that obtained between macrophylla and sylvestris, and which displays throughout the angustifolia characters. The leaves of the hybrid are obliquely ovate-lanceolate and taper gradually to a long, curved point. They are also distinctly petioled like those of angustifolia. These characters are in strik- ing contrast to those characteristic of the leaves of syl- vestris, which are broad throughout, broadly pointed, and have a broad clasping base. When flower characters are - examined, angustifolia is again faithfully reproduced in the F, hybrid for its flowers have the slender, straight corolla tube with practically no suggestion of an infun- dibulum and the deeply divided limb with narrow lobes that taper into long slender tips, all so characteristic of angustifolia. Like those of angustifolia the flowers are pink. In habit these hybrids again resemble angustifolia. This resemblance is displayed in a particularly striking fashion in the graceful, drooping manner in which the leaves are borne in marked contrast to the stiff, erect manner in which the leaves are borne by sylvestris. Throughout, macrophylla and angustifolia present a strikingly contrasted set of characters, yet in each case they are reproduced in their entirety in F, of the hybrid with sylvestris, No. 601] MENDELIAN FACTOR DIFFERENCES 41 When particular tagged Mendelian factors are con- sidered the same behavior is displayed. Perhaps there is no more striking instance of this than that shown by the expression of the calycine flower type in these sylvestris hybrids. When N. Tabacum var. calycina (Setchell, l. c., p. 6) with its peculiar split, hose-in-hose flowers is crossed with Tabacum varieties of the normal flower type, the F, hybrids display the normal flower form and segregation occurs in F, into normal and ealycine in accordance with simple Mendelian expectations. But when the -Tabacwm reaction system carries the recessive calycine flower factor into these species hybrids with sylvestris then every flower on the F, plants displays a more or less caly- cine structure. Similarly when the parthenocarpic char- acteristics of N. Tabacum var. ‘‘ Cuba’’ (Goodspeed, 1915) are carried in by the Tabacum parent then the F, hybrid, instead of shedding its capsules soon after anthe- sis as is the case in all the other T'abacum-sylvestris hy- brids, retains them indefinitely, in spite of the fact that no good pollen is produced, and thus non-fertilization, the stimulus for fruit abscission in Nicotiana, here also is the rule. So far as present evidence indicates this character- istic is rather strictly confined in Nicotiana to the variety ‘c Cuba.’’ This behavior of recessive factors of Tabacum varieties in hybrids with sylvestris is a striking confirma- tion of the conception that in such cases there is a con- trast between distinct reaction systems rather than be- tween certain factors as opposed to each other. In gen- eral when Tabacum varieties of the type mentioned above are crossed with each other the hybrids, especially with respect to flower color, leaf shape, etc., are intermediate. The contrast in this case is not one between two distinct Mendelian reaction systems, but it is merely a contrast of certain differences within a common system, and the segregation in subsequent generations, although complex, indicates a general accordance with normal Mendelian expectation. But in the case of species hybrids between Tabacum and sylvestris the contrast is between distinct 42 THE AMERICAN NATURALIST [ Vox. LI Mendelian reaction systems and the consistent reproduc- tion of all Tabacum characters, whether qualitative or _ quantitative, indicates at one and the same time that these are fundamentally of the same nature, depending essen- tially for their expression on a complex set of Mendelian factors, and, moreover, that the Tabacum system asa unit — dominates the course of somatogenesis and determines the reaction end products of the two systems. This domination of the somatogenic processes by the _ Tabacum reaction system is followed by important experi- mental possibilities. If the species hybrids always dis- play the Tabacum characteristics as completely as all our present evidence indicates that they do, then they will furnish a powerful method of attack on the problem of Mendelian behavior in the Tabacum section of the genus Nicotiana. For by crossing hybrids between Tabacum varieties with sylvestris, it should be possible to secure in the partially sterile hybrids resulting a phenotypic repro- duction of the gametic series of the Tabacum parent. This series would not be complicated by intergrading of heterozygous forms, because the plants thus obtained would exhibit the phenotypic characters of homozygotes, and recessive factors as well as dominant ones would be reproduced in their proper place in the Zabacum system. Apparently it should be possible, therefore, to demon- strate the fundamentally similar nature of linkage in Nicotiana and Drosophila. Such an analysis will still be very difficult in Nicotiana on account of its high chromo- some number, but by the method of procedure outlined above some distinct advance at least seems perfectly feasible. These, however, are matters on which we have as yet very little data. Since, therefore, the Tabacum reaction system domi- nates the somatogenic processes in the hybrid to nearly or quite the exclusion of the sylvestris system, the ele- ments of the two systems must be largely mutually in- compatible. Free interchanges between the two systems would not, therefore, necessarily result in the formation No. 601] § MENDELIAN FACTOR DIFFERENCES 43 of functional Mendelian reaction systems. This high de- gree of mutual incompatibility of the two reaction sys- tems exhibits itself in the high degree of sterility of the F, hybrids. As Goodspeed (1912, J. c.) has shown, how- ever, this sterility is only partial and a few good ovules are formed which produce viable seed in the case of open pollination or when crossed back with the parents. In evidence which is presented elsewhere (Goodspeed and Ayres, 1916) it has been shown that, while it is experi- mentally possible to modify the behavior of the F, plants in such a manner that the fruits without pollination are retained for a considerable period rather than falling soon after anthesis, the percentage of good ovules pro- duced can not be appreciably modified. This is an im- portant point, for it indicates that the number of good ovules produced is a function of the chromatin behavior and not to be influenced by environmental factors, and that they should, therefore, exhibit a consistent behavior and lend themselves to a logical interpretation. In fact, evidence at hand indicates that the small percentage of functional ovules represents the Tabacum and sylvestris extremes of a recombination series, and that, therefore, the middle members of the series, which are made up of relatively high proportions of both Tabacum and sylves- tris elements, fail to function because they produce in- compatible reaction systems. This is shown clearly by back crosses which have been made with the parents, although here portions of our evidence are not so well controlled as we would prefer. When back crosses are made with sylvestris as the pollen. parent there is pro- duced a variety of forms many of which are highly ab- normal, but among them there is a considerable propor- tion of plants which are pure sylvestris in all characters. These plants are fertile and have bred true to the sylves- tris type for three generations. A number of the remain- ing plants resemble sylvestris, but show contamination . With other elements presumably derived from the Taba- 44 THE AMERICAN NATURALIST [Vou. LI cum reaction system. These contaminations affect the whole plant, not alone any particular character complex. All plants except those which were of the pure sylvestris type were sterile. Until this year we have not been able to secure seed from back crosses with the Tabacum par- ent, but from open pollinated seed a variety of forms is produced, practically all of which are of the Tabacum type in general appearance. This result is evidently due to pollination of the F, flowers with pollen from the wide series of Tabacum forms which, in predominating num- bers, have always been grown in the cultures. Some of these plants resulting from uncontrolled pollination were likewise fertile. They have been grown for several gen- erations and although displaying segregation, this segre- gation has never involved the production of sylvestris characters, but has been of a type normally found within varietal hybrids of Tabacum. The sterile forms in this series largely resembled the F, hybrids of Tabacum and sylvestris, and the occurrence of a few aberrant and syl- vestris forms which were obtained from the sowing of the open pollinated seed are what would be expected, if pol- lination was sometimes effected with sylvestris pollen. It appears, therefore, that for these species hybrids the conception of the factors as making up for each species a reaction system in which the elements have a specific relation to one another harmonizes the results obtained with the more recent Mendelian developments. The ob- jection which might be made that interchanges of factors which behave normally in one system should not logically be followed by such profound disturbances as to com- pletely prevent the formation of a functional reaction system is met by several counter considerations. In the discussion of lethal factors it has been pointed out that ` Morgan (1914, l. c.) has demonstrated that changes in many loci of the Drosophila system have been followed by failure of the resulting individual to develop. It is en- tirely conceivable that, if a certain factor A in one system - No. 601] MENDELIAN FACTOR DIFFERENCES 45 be considered, the corresponding factor A’ in the other system, if there be such a factor, might be just as differ- ent from A as a lethal factor is different from its normal allelomorph. Further, from a modern Mendelian view- point there is no basis for assuming that recombinations could be obtained involving only exchanges in isolated loci in the systems. For if the behavior in segregation in the F, hybrids corresponds to that in Drosophila then such combinations as could be obtained would depend on the shifting of entire choromosomes or in case of crossing- over of relatively large portions of chromosomes. The recombinations obtained in the hybrids between T'abacum varieties and sylvestris, provided chromosome distribu- tion takes place after the normal fashion in these hybrids, would involve, therefore, for the most part, the formation of systems containing either whole chromosomes or large sections of chromosomes of opposing systems. Such attempted reconstruction of systems might well fail in cases of any marked specificity in the relations of the factors of the opposing reaction systems, since any large proportions of both systems in a gamete might conceiv- ably destroy the continuity or the balanced relations neces- sary for the continuance of system reactions. When the proportions of one or the other system are relatively small, the system reactions might be merely disturbed, resulting in the production of the abnormal forms which have been secured in our cultures. It is to such relations between the two systems involved that we ascribe the selective elimination of the greater portion of the possible gametic combinations in the crosses between Tabacum varieties and sylvestris, and which, therefore, results in a high degree of sterility in these hybrids. It is of course obvious that there are many categories of sterility, in the hereditary sense as well as in the physi- ological sense. The particular type herein considered is that which results from relatively wide crossing such as is involved in species hybrids. That certain types of steril- 46 THE AMERICAN NATURALIST [Von. LI ity are due to specific factors which display consistent Mendelian behavior has been demonstrated by Bateson and Punnett (1908) and others and has been suggested by Correns (1913) for cases which are concerned with the specific category of normal self-sterility in Cardamine pratensis. It appears, nevertheless, as Hast (1915) in principle has suggested, that many cases in which sterility has followed rather wide crossing would seem to be sus- ceptible of a more logical treatment from the standpoint of non-specific disturbances in the reaction systems involved. (To be continued) COMPARATIVE RESISTANCE OF PRUNUS TO CROWN GALL Proressor CLAYTON O. SMITH UNIVERSITY OF CALIFORNIA In making a study of the plant disease popularly known as crown gall, plant tumor, or plant cancer, it seemed desirable to ascertain the relative resistance of the different species of Prunus to this disease. It was soon evident that the usual methods em- ployed in discovering disease resistance would be of little value. The cause of the disease, life history and pathogenic nature of the organism had already been studied by Dr. Erwin F. Smith? and his assistants of the United States Department of Agricul- ture. They showed, by artificial inoculations, the wide range of plants susceptible to infection and.also found that some were apparently resistant. Their experiments encouraged the writer to follow with slight modifications the method of artificial inocu- lations on a number of species and varieties of the genus Prunus. It was hoped that suitable resistant stock might be discovered that would be adapted to the propagation of the stone fruits. Before considering in detail the methods employed, the general characteristics of the disease will be briefly given. The affected part of the tree, shrub or plant is generally found a little dis- tance beneath the surface of the soil at the crown or point where the roots are given off from the trunk. The disease is characterized by an enlargement or gall more or less spherical in shape and consisting of tissue that is usually much softer in texture than — normal. The surface may or may not be covered with a normal bark. This enlargement is now known to be caused by a stimulus that comes from the presence of a definite motile bacterial organ- ism known as Bacterium (Pseudomonas) tumefaciens, which lives within certain of the plant cells in relatively small numbers. / Considerable attention was given to perfecting methods for 1 Paper No. 28, Citrus Experiment Station, College of Agriculture, Uni- versity of California, Riverside, California. 2 United States Department of Agriculture, Bureau of Plant Industry, Bulletins No, 213 and 255, 1911, 1912. 47 48 THE AMERICAN NATURALIST [ Vou. LI the determination of the relative resistance of the different spe- cies of Prunus. Most of the different species and varieties were budded or grafted on other stock, the scions or bud wood being secured from several reliable sources, such as the Arnold Arbo- retum, several of the larger nurseries. The methods used are somewhat different from those usually employed in seeking for disease resistance among plants. The plan was to artificially inoculate with pure virulent laboratory cultures the different kind of Prunus under experimentation. A number of suscep- 40 TE + e a <0" sae w mee iat a a Cats u D - 4 P -M o =a ta . 1. A bunch of galled peach stock as they often occur in the nursery. Many of the other trees probably had incipient infections unrecognizable at the time of digging, which later developed galls. No. 601] RESISTANCE OF PRUNUS TO CROWN GALL 49 tible hosts were always included in the experiment, to act as a check upon the virulence of the culture and any unfavorable climatic condition. In each series of inoculation 5 or 10 punc- tures were made upon vigorous growing twigs of the current a Fig. 2. Natural gall on English walnut in nursery. This gall appeared at the point where the English walnut stock was grafted. It is common practice to bank the dirt about the scions to prevent drying. The California black root is much more resistant. year’s growth. During the experiments of 1914, ten punctures were always made. This number, being the same in all the ex- periments, was of material aid in the final compilation of results. Other check punctures were made in the same way as in the in- oculations except that none of the organisms were placed in the tissue. An ordinary steel needle in a cedar handle was used in making the puncture inoculations. This was first flamed, then used to convey some of the bacterial growth from the test tube to the twig to be inoculated, the puncture being made through the bark and wood of twig. The organism was grown in a medium made as follows: 4 per cent. glucose, 4 per cent. sodium chloride, $ per cent. meat extract, 1 per cent. peptone, 1 per cent. agar. 50 THE AMERICAN NATURALIST [ Von. LI Fic: 3. Two aT roots naturally infected with numerous galls. Such stock ould only make inferior trees The tubes were incubated from twenty-four to thirty-six hours before being used, at which time there was a vigorous, pearly white, raised growth where the medium was inoculated. The series of inoculations were made a week apart from May 1 to about September 1, 1914. -The work of 1913, while similar in nature, was not so extensive as that of 1914. The experiments thus extended over the period of the year when the trees are making their most rapid growth, and should be in their most susceptible condition for infection. The trees were well cared for and made rapid growth during the period the experiments were in progress, and hence were in favorable condition for the development of the disease. No effort was made to protect in any way the punctures, as the use of wax or other covering stim- ulates callus formation which could easily be confused with the beginning stages of a young gall or with one that has not ma- tured rapidly, as is often the case on inoculated trees showing resistance. The genus Prunus gives a wide range for investigation be- cause of the large number of species and varieties. The follow- ing are the species thus far tested by artificial inoculations: No. 601] RESISTANCE OF PRUNUS TO CROWN GALL 51 Prunus Allegheniensis, P. Americana, P. Amygdalus, P. ander- sonit, P. Armeniaca, P. Armeniaca, var. Mikado, P. Avium, P. Besseyi, P. Caroliniana, P. cerasifera, P. cerasifera, var. divari- cata, P. cerasifera, var. Planteriensis, P. domestica, several dif- ferent varieties, P. eriogyna, P. glandulosa, P. hortulama, P. ilici- folia, P. integrifolia, P. Japonica, P. maritima, P. Mahaleb, P. Mitis, P. monticola, P. Mume, P. munsoniana, P. nigra, P. ortho- sepala, P) Pennsylvanica, P. Persica, several varieties, P. platy- carpa, P. pumila (Linn.), P. serotina, P. Simoni, P. Watsoni. All the above hosts gave positive results from artificial inocu- lation, except P. pumila, P. ilicifolia and P. Caroliniana. These three hosts were inoculated during 1913 and 1914 and have always showed negative results. P. ilicifolia and P. Caroliniana were on their own roots and were not making very rapid growth at the time of the experiment, but it seems almost impossible that they should not have been in a susceptible condition some time during the period from May to September. P. pumila made rapid growth, as it was grafted on peach stock, but never showed the least indication of gall formation, nor has it during the ex- E Fic. 4. Fie. 5. Fig. 4. Artificial gall produced on pepper tree, Schinus Moll a 5. rtific ag a made by inoculating sour orange seedlings. Citrus ck is only rarely cted naturally with gall, but galls have been artifi- cially produced on Ee ssa orange and shaddock. Fn 52 THE AMERICAN NATURALIST [ Vou. LI periments of 1915. Only a part of the kinds of Prunus just mentioned were thoroughly tested out and these only are included in tabulated results. It will be noted that Tables II and III represent two varieties of Prunus domestica. Not all varieties, however, are equally resistant, but, in general, members of this group are much more Fie. 6. Artificial galls on Prunus. (a) German prune, P, domestica, a resistant variety. (b) Myrobolan, P. cerasifera, a priate: eda ir the difference in size of galls and how in es susceptible kind the galls eventually surround the twig No. 601] RESISTANCE OF PRUNUS TO CROWN GALL TABLE I 53 A TYPICAL EXPERIMENT WITH Bacterium tumefaciens, CULTURE No. 753 C Y TO NOVEMBER 15, 1914), TO ILLUSTRATE GENERAL METHODS (JUL USED IN SEARCHING FOR A RESISTANT STOCK. A SIMILAR EXPERIMENT TO THIS WAS MADE EACH WEEK FROM MAY TO SEPTEMBER, 1914. Experiment 20 Experiment No. of Positive Size of Galls Serial No. |Inoculations Inoculations Host in Inches 745 10 2 German prune gt 746 10 0 German prune (old) 747 10 5 am 3-4 748 10 10 P. triflora w- 749 10 10 | Wic 43 750 10 9 Burbank 4-4 751 10 10 Myrobolan 3-3 752 10 10 P. Munsoniana 4-3 753 10 10 | P. davidiana 1 754 10 5 P. maritima 3—4 755 10 4 P carpa i-t 756 10 0 Golden drop (silver prune) 757 10 8 Reine Claude (green gage) 3s- 758 10 10 | P. Simonii 14 759 10 10 Royal apricot 4-4 760 10 10 Elberta peach 1-1} 761 10 0 Bitter almond 762 10 9 P. hortula 3-3 763 10 5 P. i as} 764 10 10 Schinus Molle 1-14 765 10 Positive, Oleander 766 10 0 P. pumila 767 10 0 P. Watsoni 768 10 8 | P. nigra 3-4 769 10 3 P. serotina ve-3 770 10 10 P. institia pendula 3-3 771 10 9 | P. mitis vs small 772 10 10 P. Mume 3-7 773 10 7 | P. Anders t 774 10 6 Duane (Tribble) a 775 10 2 nteriensis 776 10 8 Miyrobolan (Arnold) ys-t ars: 10 9 M.yrobolan (sprouts) $-1 778 10 10 Viyrobolan (young sprouts) 3-3 779 10 10 dl Paso 3-7 780 10 0 Golden beauty 781 10 10 Arkansas 4-4 782 10 1 P. Virginiana % 783 10 7 P. cerasifera divarica -4 784 10 2 P. Besseyi ve 785 10 0 P. ilicifolia 786 10 0 P. Caroliniana 787 10 9 | P. orthosepala + 788 10 0 Olive 789 10 2 P. Armeniaca Mikado 3-3 790 10 5 Italian prune -$ 54 THE AMERICAN NATURALIST [ Vou. LI so than most other species of the genus. By comparison with other tables, it will be found that the galls are of a much smaller size than on most other hosts. The number of positive inocula- TABLE II SUMMARY OF ARTIFICIAL INOCULATIONS ON GERMAN PRUNE, Prunus domestica. CONCLUDED NOVEMBER 15, 1914 Positiv: ize of Gal EI eae Date No. of I lati focalationi a7 bt as æ 500° 5/ 3/14 woo 0 532 5/25/14 10 0 559 6/18/14 10 0 zx 560 6/18/14 10 0 617 6/29/14 10 6. }-4 x 618 6/29/14 0 667 6/29/14 Check 0 679 7/ 6/14 6 3-4 zx 680 7/ 6/14 10 0 745 7/13/14 10 2 4-4 xz 746 7/13/14 10 0 æ 792 7/20/14 10 0 870 7/27/14 10 0 æ 871 7/27/14 10 0 934 8/ 3/14 Check 0 z 8/ 3/14 k 0 8/ 4/14 10 2 i- | 4/14 10 0 A 15 8/10/14 10 0 A 87 8/15/14 10 0 2A 88 8/15/14 10 0 A125 8/17/14 10 0 A207 8/24/14 10 8 B- 2A208 8/24/14 10 0 242 8/31/14 10 0 2A243 8/31/ 10 0 240 24 tions as given in Tables II and III, is probably somewhat greater than it should be, as in making the estimate of the number of galls on these resistant stocks, any small enlargement was counted, and subsequent examination has shown that many of these small enlargements have not further increased in size.. When a gall becomes established in a resistant variety, it makes rapid growth and eventually forms one of good size. These large galls differ from similar galls on peach and many other hosts in that the gall is attached to a relatively small circum- ference of the infected twig. The gall growth is often nearly at 8 Numbers that are preceded by an x were made on rapid growing twigs of the current year of a seven-year-old tree. The other inoculations in a young tree two years old from the nursery. No. 601] RESISTANCE OF PRUNUS TO CROWN GALL 55 TABLE III SUMMARY OF ARTIFICIAL INOCULATIONS ON ITALIAN PRUNE (FELLENBERG) Prunus domestica. CONCLUDED NOVEMBER, 1913 m vi Size of Galls pegase me Date No. of I lati Porn in Inches 211 5/10/13 5 0 216 6/1 5 2 e e 252 6/13/13 5 0 264 6/16/13 5 0 270 6/17/13 5 0. 282 7/14 15 5 Tsis ie- re-te 292 7/13/13 5 0 315 7/19/13 6 2 Ts-ts 328 7/21/13 6 1 te 332 7/21/13 5 0 343 7/22/13 5 0 355 7/25/13 10 4 b-i- ik 366 7/30/13 10 0 259 6/ 9 0 375 7/30/13 10 ry 390 | 9/ 10 0 406 8/14/13 10 3 Ye-is-ts 420 8/14/13 5 0 127 17 right angles to the twig which makes these galls stand out for considerable distance from the branch. It is of interest to note that both of these are prunes that have been under cultivation for many years. The German prune is described as being one of the plums longest under cultivation and the oldest of the prune type. Seedlings also come reasonably true to type which might be of importance if grown from seed for a stock. The Italian prune (Fellenberg) is the popular prune of Oregon and has a history of over a century’s cultiva- tion. Further experiments among varieties of the domestica group are being carried on. The damson which is sometimes included among the domesticas, shows considerable resistance to artificial inoculation. Prunus cerasifera, var. Planteriensis, Table VIII, is described as a double flowering shrub and is the most gall resistant of any of the tested varieties of cerasifera, although this resistance should be again determined. Inoculations in the Arnold Ar- boretum trees, Table IV, did not develop as many galls as those of the larger Myrobolan tree, either because the former were not 4 A local commercial stock, propagated in California from sprouts, not the true Duane eimh but a small blue plum having the flavor of a Damson, but differing in 56 THE AMERICAN NATURALIST [Von. LI TABLE IV SUMMARY OF ARTIFICIAL INOCULATIONS ON Prunus cerasifera. CONCLUDED NOVEMBER, 1914 Prunus cerasifera, Arnold Arboretum Type "aan me ear Date No. of Inoculations ee: a Saar 595 6/20/14 10 10 4-4 646 6/29/14 10 8 is-is 709 6 10 9 i$- 776 7/13/14 10 8 gst 7/20/14 10 5 Is 7/27/14 8 4-4 963 / 3/1 Check 0 Á i 8/10/14 1 4 44 Alll 8/15/14 10 2 ie A149 8/17/14 10 4 ts A225 8/24/14 10 10 vs-' A253 8/31/14 10 2 ‘5 110 70 Prunus cerasifera (large four year old seedling). SAREH Datë Nö. ot Tioruiatiaan neers Z ao cula- Size Slaen in 5/ 5/14 10 10 li- 4 538 5/25/14 10 10 3 66 6/18/14 10 10 62 | 6/29/14 10 10 — t 6/29/14 10 10 ae / 6/14 10 10 -3 751 7/13/14 10 10 =% 821 7/20/14 10 10 waf T 901 7/27/14 1 10 i 962 8/ 3/14 Check 0 A / 4/14 10 10 4-3 8/10/14 10 10 4-4 A110 8/15/14 10 10 4 $ A148 8/17/14 1 10 3-4 A214 8/24/14 10 10 A251 8/31/14 10 10 150 150 growing as rapidly, or, judging from their shrub-like growth, because they are of a different type possibly nearer to the wild type than are those commonly imported from France by nursery- men. The variety known as Golden Beauty, P. hortulana, has thus shown more marked resistance than other varieties of the species thus far tested. It is interesting to remember that this variety is supposed to have originated in western Texas some- No. 601] RESISTANCE OF PRUNUS TO CROWN GALL 57 what out of the natural range of the species. P. hortulana and P. americana are used as a stock for the native plums in the middle west and east. P. hortulana does not sucker, fruits abun- dantly and has a number of excellent qualities that would recom- TABLE V SUMMARY OF ARTIFICIAL INOCULATIONS ON GOLDEN BEAUTY, Prunus hor- tulana. CONCLUDED NOVEMBER, 1914 i ize rier ee i Date No. of Inoculations | Minn son pit g lag 600 6/20/14 10 | 7 651 6/29/14 10 | 3 vs very small 714 7/ 6/14 10 | 5 ys-t 780 7/13/14 10 0 826 7/20/14 10 4 ie 906 7/27/14 10 1 is 967 8/ 3/14 Check 0 Att 8/ 4/14 10 3 + A 44 8/10/14 10 All4 8/15/14 10 1 ys A152 8/17/14 0 1 is A250 8/29/14 10 0 110 25 mend it as a stock. Further experiments with varieties of this Species are being made and its adaptability to various of our stone fruits carefully studied. | TABLE VI SUMMARY OF ARTIFICIAL INOCULATIONS ON Prunus pumila Linn. CoN- UDED NOVEMBER 15, 1914 Experiment Serial No. Date No. of Inoculations Positive Inoculations 6/20/14 10 0 635 (29/1 10 0 766 7/13/14 10 0 849 7/20/1 10 0 88 7/27/14 10 0 94 / 3/1 Check 0 Rees 8/ 4/ 10 0 A 34 8/10/14 10 0 A106 8/15/14 10 0 Al44 8/17/14 10 0 A220 /24/ ; 10 0 A254 8/31/14 10 0 110 Further inoculations of P. pumila (thirteen experiments of ten inoculations each or 130) made during the present year, 1915, t 58 THE AMERICAN NATURALIST [ Vou. LI in vigorous growing seedlings, gave negative results, agreeing with the results of the two previous years. The experiments thus far conducted show that the species is entirely resistant to arti- ficial inoculations. P. Bessey, closely related to P. pumila, also shows considerable resistance. The two other species of Prunus referred to as being resistant are P. ilicifolia and P. Caroliniana. They are evergreens and are not now considered as strictly belonging to the genus Prunus. They do not readily unite by grafting or budding with varieties of the stone fruits. P. pumila is a shrub and while this stock readily unites with many of the varieties of the stone fruits it probably would dwarf the tree more or less and might sprout. It is, however, readily grown from pits or cuttings. TABLE VII SuMMARY OF ARTIFICIAL INOCULATIONS ON Prunus. CONCLUDED OCTOBER Species Variety | isins | imta Per Cent. g Bas" P. domestica........ German prune 75 2 23 1-4 P. domestica........ German prune 50 5 10 fir P. domestica........ | Italian prune 127 17 13 34-4} P. domestica........ Green gage 17 4 23 Is-4 P bn oen 138 51 37 3-3 vba care Duane 37 15 40 P. persica Elberta 61 57 93 3-24 P. triflora X P. Wickson (hy- Simoniti ve i 56- 64 98 4-1 P. triflora.........- Burbank 31 31 100 4-4 P. cerasifera. ccs: Myrobolan 8 8 100 4-1 In each of the inoculation experiments, ten punctures were made, hence the number of inoculations divided by ten will give the number of separate experiments made with the various hosts. In the two tables, VII and VIII, where the inoculations of the years 1913 and 1914 have been summarized, there is a reasonable degree of consistency between the percentages shown for the dif- ferent hosts. Other varieties of P. persica possibly should be further tested, although such varieties as Elberta, Saucer or Peento, Salway, Lovell and Muir seedlings have not shown any marked resistance. Any of the stock as listed in Tables VII and VIII that show less than 50 per cent. infection are more or less promising, for in the experimental work with artificial inoculations from virulent No. 601] RESISTANCE OF PRUNUS TO CROWN GALL 59 TABLE VIII SUMMARY OF INOCULATIONS MADE ON Prunus. CONCLUDED NOVEMBER 15, 1914 Species Variet oes phoned Cent. oț|Size of Galls r ‘i lations | lations | Galls | in Inches P. pumila.. ....| 2 varieties, Arnold 110 0 0 Arboretum P. domestica,...... Italian prune 140 10 7 ps3 P. cerasifera....... P. planteriensis 40 3 73| tet P. domestica....... German prune 240 24 10 | ye-4 P. insititi .....| Damson 120 13 10 3-4 P. Bessey 258: — 50 5 10 4-4 P. hortulana....... Golden Beauty 110 25 22 Is-4 P. amygdalust ....| Bitter almond 100 22 25 3—3 P. domestica....... Reine Claude (green gage) 90 25 26 | 354 P. Armeniaca...... Mikado 40 11 27| b P. angustifolia.. ... Watsoni 50 15 30| 3—4 P. maritima....... Arnold Arboretum 140 48 34. i4 P. aema E apa Eei Arnold Arboretum 130 55 42| %2} P. Miti .| Arnold Arboretum 32 53| ps4 P. pivot E ESN Arnold Arboret 110 70 63 | s+ P. Munsoniana.... ai eri SS 70 48 68 4-35 P. Munsoniana.... 90 70 77 | pe? P. Americana...... poner y retum 100 83 83 | 3- P. hortulana....... Arnold epee 130 | 108 83 4-3 PP. Ansa a Pendula 90 77 85 | 7-2 P. davidiana....... — 110 96 88 4-3 Schinus Molle... .. Pepper tree 110 97 88 ił- t PSAP Rk Burbank 120 | 109 90 $- Paata A a R eh Arnold Arboretum 60 56 90 4-3 P. orthosepala...... Arnold Arboret 80 72 90 | 3-4 aA eS Arnold Arboretum 100 9 91 1 P. Munsoniana....| Pi 140 | 130 92 | y-1} P. cerasifera....... P. divaricata 00 94 $- rE Piia. on Elberta 130 122 94 4-3 P. Armenica....... Royal apricot 120 117 97 E Pi tiflora oe Arnold Arboretum J40. 137] OT -4 P. Munsoniana. ...| El Paso 100| 97| 97| 44 P. cerasifera....... Sprouts : 120 | 117 97 4-13 P. triflora X P a Pre Ft 7 (hybrid) 140 138 98 4-3 cerasifera....... 150 | 150 : 100 14 P. monticola....... eon gerson Sta. 40 40 | 100 14 P. Simonii........ Arnold Arbo 130 | 130 | 100| 4-13 cultures, the different stocks were subject to a more severe test than obtains under the usual field conditions. The more promis- ing of these, the first seven in Table VIII, with the exception of the damson, have already been considered after the various table in which the results were summarized. The damson has often 5 Prunus Amygdalus was not growing well during the last part of the season and no infections showed after that of July 6. Before this time a large percentage of the inoculations were positive. This will be repeated ye 60 THE AMERICAN NATURALIST [Von. LI been used as a stock, but is not popular on the Pacific coast be- cause of its slow growth in the nursery and the difficulty of working it with many of the stone fruits. Duane, P. domestica, Table VII, is being used to some extent as a stock in California and shows resistance to gall in old vineyard land where it has been grown for six years. It makes as large a tree as the popular Myrobolan stock. The seedlings are grown from suckers, which give a root likely to sucker. Reine Claude (green gage) variety has shown resistance and would without doubt be a good stock for the domestic type of plums. P. Armeniaca, variety Mikado, is an apricot that differs somewhat from the one commonly grown in California. It should be tested out experimentally as a stock for apricots to replace the susceptible one now being used. The almond, from field observations, is one of our most sus- ceptible stock and this is fully confirmed by the following inocu- lation experiments: Fourteen different varieties of almond seed- lings not summarized in the following tables were inoculated in April of 1913 at the University Farm at Davis, California. These in all cases showed a high percentage of infection. So far the peaches and almonds have shown only slight resistance. It will be noted that our most popular stocks as Myrobolan, peach, apricot and almond are very susceptible, which only goes to confirm field observations that the stock used for the stone fruits are very susceptible to crown gall. The work so far conducted shows that seedlings of the German and Italian prunes might be promising stock for certain of the stone fruits, probably those of the domestica type. However, no definite recommendations can be given, as the work is now only in its preliminary stages. SHORTER ARTICLES AND DISCUSSION EUROPEAN FOSSIL FISH-SCALES In European Cretaceous deposits fish-scales have been found at various times, and occasionally have been named and described by paleontologists. Dr. A. S. Woodward, in his great ‘‘Cata- logue of the Fossil Fishes,’’ has carefully and accurately listed all the names so given, but has made little or no attempt to ex- amine the records critically, assuming that they were valueless, or nearly so. More recent work on fish-scales brings out the fact that these materials are of great value for the understanding of Mesozoic fish life, but unfortunately they have been described with little knowledge of their significant characters. An im- portant pioneer work was that of Geinitz,! describing scales from the Turonian of Saxony. Geinitz realized that it was necessary to make comparisons with scales of recent fishes, and gave a plate of ‘‘Schuppen von lebenden Fischen,’’ but unfortunately chose species which had little relationship, for the most part, with the fossils studied. The three plates of fossil scales appear to have been very carefully drawn, and from them it is possible to gather a number of facts not brought out in the text. Cyclolepis agas- sizi appears to be Salmonoid, agreeing quite well with the modern Salmo. Aspidolepis steinlai is like the scales of living Stroma- teide, as Poronotus. Osmeroides divaricatus evidently has nothing to do with the genus to which it is assigned, but is of characteristic Albulid type. Osmeroides lewesiensis (Mantell), as determined by Geinitz, consists in the main of scales agreeing with those of the remarkable living genus Pterothrissus. The genus Cladocyclus presents some serious difficulties. The type is a Brazilian species (C. gardneri Agassiz) from the Upper Cre- taceous of the Province of Ceará. I am indebted to Dr. D. S Jordan for material referred to this species, and it appears that the large scales have extremely fine circuli, while those of the lateral line possess branching canals of the same general type as those of the living S. American genus Hydrolycus.? It does not 1*“Die Fossilen Fischschuppen aus dem Plinerkalke in Strehlen,’’ 1868. 2 These canals are in the apical, not the te field, as erroneously stated by me in Annals Carnegie Museum, IX, p. 1 61 62 THE AMERICAN NATURALIST [Vou. LI seem certain that this is the true Cladocyclus of Agassiz; it may represent a new genus ancestral to the Neotropical Characoid fishes. European ‘‘Cladocyclus”’ is in any event surely distinct from the Brazilian. The C. strehlensis of Geinitz includes scales approaching those of Potamalosa, but the species is founded in the main on an entirely different type, which is evidently close to the English C. lewesiensis. It is a question what generic name should be used for the lewesiensis-strehlensis type, which must be removed from Cladocyclus. It was formerly included in Hyp- sodon Agassiz, but that generic name appears to belong properly to the fishes usually called Portheus, the type being lewesiensis Mantell — mantelli Newton. The name lewesiensis also belongs strictly to the Portheus, and the English so-called Cladocyclus, if distinct from the Geinitz species, seems to need a new name. These matters will be taken up more fully elsewhere at a later date. Beryx ornatus of Geinitz, properly called Hoplopteryx lewestensis (Mantell), appears to be a primitive Berycoid type, having scales such as might be expected in an ancestor of the modern Berycide. Hemilampronites steinlai Geinitz consists of scales differing little from the living Hyporhamphus. The scales figured by Geinitz as those of Macropoma mantelli Agassiz? have no resemblance to that species; from the fine transverse circuli, basal radii, and apical teeth like those of Pomacanthus, they appear to belong to some Teleost more or less related to the Berycoids. Thus we find that although Geinitz knew little about the affinities of his scales, they had excellent characters, remind- ing us in certain cases of modern genera, and indicating the great antiquity and constancy of peculiarities of scale structure. In 1878 Anton Fritsch* undertook to describe the fish-seales of the Upper Cretaceous of Bohemia, and believed that he had a number of the species of Geinitz. His Cladocyclus strehlensis and Cy- clolepis agassizi are perhaps correct, but the others are evidently different from the Geinitzian forms. His Macropoma speciosum Reuss is a genuine species of that genus, with quite characteristic seales. His Macropoma forte, on the other hand, appears to be a Celacanthus. His Osmeroides lewesiensis has regular trans- verse circuli between the radii, instead of the minute tubercles (markings like the surface of a strawberry) of the Geinitz scales 3 Macropoma mantelli should be called Arm lewesiensis (Mantell), based on the Amia(?) lewesiensis of Mani 4‘*Die Reptilien und Fische der ‘hated Kreideformation.’’ No. 601] SHORTER ARTICLES AND DISCUSSION 63 and of Pterothrissus. His O. divaricatus is wholly distinct from the scale described by Geinitz under that name, and has not the Albulid characters. The Beryx ornatus scales are printed up- side down, and the artist has added ctenoid structures (small . teeth) above, on what is really the basal margin. In 1874 T. C. Winkler published a paper® in which he described two species from the scales. His Osmeroides belgicus appears to be con- generic with the Osmeroides lewesiensis as understood by Fritsch. His Cycloides incisus, supposed new genus and species, of which he says that he knows no fish, living or fossil, with such scales, is apparently worthless. It may not be a fish-scale. In America the fossil scales of Teleosts have received very little attention, but a large collection accumulated by the U. S. Geological Survey is now under review, and will undoubtedly yield much of value for the understanding of Mesozoic fishes, and at the same time throw light on the ancestry and relationships of modern families. T. D. A. CocKERELL UNIVERSITY OF COLORADO, BOULDER, July 8, 191 5‘‘ Mémoire sur quelques restes du Poissons du systéme heersien,’’ Arch. Mus. Teyler., IV. THE AMERICAN NATURALIST VOL. LI. February, 1917 No. 602 THE SELECTION PROBLEM? RAYMOND PEARL Or all the supporters of the doctrine of natural selec- tion as the chief factor in organic evolution, August Weismann was preeminent. He stood shoulder to shoulder with Wallace in his entire willingness to attempt . the explanation by selection of any biological phenom- enon whatsoever, and he far outstripped the latter in the keenness and subtlety of his logical powers when an espe- cially difficult bit of exegetic activity was called for. Both, it hardly needs saying, left Darwin far behind in the extent of their advocacy of the Allmacht of selection. Certain it is that if any one could speak authoritatively, and without suspicion of either hostility or doubt, about the selection theory, Weismann could. For that reason it seems desirable to take as the starting point of this dis- cussion a statement made by that distinguished biologist So lately as seven years ago. At the Darwin Centenary meeting in Cambridge, Weismann, discussing the ade- quacy of the selection theory to explain the initial steps of evolutionary change, said:? To this question even one who, like myself, has been for many years a convinced adherent of the theory of selection, can only reply: “ We must assume so, but we can not prove it in any case. It is not upon 1 Papers from the Biological Laboratory of the Maine Agricultural Ex- periment Station No. 109. This paper constitutes the address of the retir- ing President, read in abbreviated form at the dinner of the American So- ciety of Naturalists in New York, December 29, 1916. ; 2**Darwin and Modern Science,’’ Cambridge, 1909, p. 25. The italics are in the original. 65 66 THE AMERICAN NATURALIST [ Vou. LI demonstrative evidence that we rely when we champion the doctrine of selection as scientific truth; we base our argument on quite other grounds.” Even since 1909 a good deal of water has flowed under all our bridges, and particularly under the evolutionary ones. Among other changes in viewpoint there is evident a marked disinclination in science nowadays to regard as “scientific truth’? anything which is not based upon demonstrative evidence. But it is also a fact, perhaps at first thought to be regarded as curious, in view of the opinion of Weismann which has been quoted, that there are here with us to-day those who assert, with great zeal and pertinacity, that in selection is to be found the chief cause of evolutionary change. These things being so, it has seemed that possibly it might be profitable to spend a little time upon the selection problem, trying to deter- mine whether the case is any better now than Weismann conceived it to be seven years ago, from the viewpoint of tangible objective evidence. It is to be hoped that it is, for among working geneticists just now any theory which has to depend for its sole support upon its ‘‘inter- pretative value’’® is sure to receive scant attention. So then what I shall try to do is to review briefly some of the real evidence about the selection problem which has been accumulating since biologists turned definitely to. the experimental study of evolution, and definitely away from the glorious, but on the whole unproductive, attempt to solve its problems by @ priori reasoning. From such a review it is to be hoped that we may get some light as to the directions in which further research and new evidence are most urgently needed. I. NATURAL SELECTION In considering the whole selection problem we may well begin with an examination of the theory of natural selection. The mere fact of natural selection, in the sense solely and strictly of a process leading to the elimination of some individuals and the survival of others, is no 3 Weismann, loc. cit., p. 50. No. 602] THE SELECTION PROBLEM 67 longer questioned by any one who takes the trouble either to think or to observe living things. It is a process which goes on constantly and affects all organisms. In this sense it is no more than the resultant of the observed absence of individual, mundane immortality among living things. The fact that individuals die implies that those not yet dead are a selected lot, in at least one respect, namely survival. This mere fact of elimination and survival is, how- ever, not in itself particularly illuminating. The first question before us is whether such a process is capable of bringing about evolutionary changes of a progressive sort. Obviously it is capable of doing so, in theory at least, if we add two assumptions, or better rules accord- ing to which the Dance of Death is to be performed. The first of these rules is that the individuals alive at any time shall be different from those dead, in some other respects than that of survival merely. In other words, the elimination shall be selective. The second rule is that the survivors shall transmit to their progeny those differences which mark them off from the eliminated. The theory that these two rules are always and every- where in operation, taken together with the observed fact that living creatures do die, is the Darwinian theory of natural selection as a factor in organic evolution. If the premise be granted that the game of survival is in fact played by these rules, the conclusion is then logically irresistible that evolutionary progress is bound to occur in the direction of those differences which distinguish the survivors. Here many have been content to let the matter rest. In the minds of an astonishingly large number of people, which number includes some rather great names in the world of science, it is precisely the same thing to show that something logically must be so, as it is to show that It is so. If the formal rules of logie are satisfied, truth Seems to them to be thereby established. No further evi- dence is demanded. As every one knows, this attitude led practically to the intellectual bankruptcy of the whole 68 THE AMERICAN NATURALIST [ Vor. LI evolution theory in the late nineties, from which it was rescued only by the active movement towards an ob- jective, experimental accumulation of facts about the sub- ject. But the danger which lurks in formal logic is always threatening the progress of science. In the field of sci- ence in which we are interested the most recent conspicu- ous example of it is found in the vicious attacks on Men- delism, which upon analysis can be seen to have their only basis in a formally logical, so-called ‘‘proof’’ that it can not be true. The danger is so insidious, and takes such diverse forms, that one feels justified in quoting a brief statement made by Professor F. C. S. Schiller, which might well, in sufficiently large type, be hung upon the wall of every biological laboratory, as a constant reminder that the foundations of scientific truth lie in experiment and observation, not in logic. Schiller says: The proof that any logic, which declines to consider the question of the real truth of the reasonings it attempts to deal with, necessarily condemns itself to utter formality is easily given, and very instructive. It is a formal characteristic of every assertion that it claims truth, ab- solutely and without reservation or suggestion of fallibility. Hence it follows both if (a) the question of the actual value of this claim is ruled out of order, and if (b) the assertion is accepted at its own estimation, that the distinction between true and false must, in fact though not in name, disappear from Logic. For all assertions will be held true be- cause they formally claim truth; because none profess to be false, error no longer exists—for Logie. Thus the logical form of an assertion affords no means of deciding upon the real value of its claim to truth, and hence any logic which restricts itself to the study of this form inevitably accepts a truth-claim as the equivalent of real truth. It is like a bank which does not distinguish between promises to pay and hard cash. Then clearly the question to which we want an answer is not whether natural selection can cause evolutionary changes, but rather whether it does cause such changes in any significant degree or extent. In other words, we shall prefer the ‘‘hard cash’’ of objective experimental evidence to any logical ‘‘promise to pay,’’ however tight and compulsive its reasoning. 2 4Schiller, F. C. S., ‘‘ Formal Logic. A Scientific and Social Problem.’’ London, 1912, pp. 6-7. No. 602] THE SELECTION PROBLEM 69 The tale here is not a long one. Indeed it is surpris- ingly brief, considering the mass of literature which the theory of natural selection in its more formally logical aspects has engendered. We have first the pioneer work of Weldon® with Carcinus, in which a selective elimina- tion of individuals different physically from the survivors was first demonstrated numerically, the eliminating en- vironmental factor being the silt in the water. This was followed by a number of investigations of a more or less similar character, notably those of Poulton and Sanders® with Vanessa, and of di Cesnola” with Mantis, in which different colored forms of these insects were exposed to elimination by natural enemies, chiefly birds, with the result that there was found to be some relation between the chances of elimination and the degree to which the insect matched its background. Bumpus® studied sur- viving and eliminated English sparrows after a severe winter storm. Crampton® measured the surviving and eliminated pupe of Philosamia, the elimination having been produced by wholly natural causes. Davenport,” in a very small lot of chickens, found that those killed by crows were colored differently from those eliminated. Lutz!! found in Drosophila some differences in type be- tween survivors and eliminated. Harrist? has shown that among seedling beans abnormal types perished more fre- quently than strictly normal types under the same field conditions. The same author? has also made extended 5 Weldon, W. F. R., Proc. Roy. Soc., Vol. XLVII, pp. 360-379, 1894. Also see Brit. Assoc. Rept., Bristol (1898), pp. 887-902, 1899. ê Poulton and Sanders, Rept. Brit. Assoc. (Bristol), gid ni 1899, 7 di Cesnola, A. P., Biometrika, Vol. III, pp. 58-59, 8 Bumpus, H. C., Biological Lectures, Woods Hole, pae pp. 209-226, Boston, 1899. 9 Crampton, H. E., Biometrika, Vol. III, pp. 113-130, 1904. 10 Davenport, C. B., Nature, Vol. LXXVIII, p. 101, 1908. 11 Lutz, F. E., Bulletin Amer. Mus. Nat. Hist., Vol. XXIV, pp. 605-624, 1915. 12 rg J. A., Science, N. S., Vol. XXXVI, pp. 713-715, 1912. 18 Harris, J. A., Science, N. 8., Vol. XXXII, pp. 519-528, 1910. Pop. Sci. Honki, Vol. LXXVIII, pp- 521-538, 1912, and numerous other papers in Biometrika, AMER. Nat. and elsewhere. 70 THE AMERICAN NATURALIST [Vou. LI researches on the elimination of organs in a series of dif- ferent plants. The critical value of these different investigations is not in every case equal. Some are distinctly fragmentary, and in others the differences between eliminated and sur- viving are so small as to be of extremely doubtful sig- nificance. If one examines critically the actual biometric constants in the more extended of these studies (e. g., Crampton’s and Bumpus’s as analyzed by Harrist) he can not but be impressed with the doubtfulness of many of the differences. However, if we take all these re- searches at their face value, and give all the benefit of the doubt to the weak, then they agree in indicating that the survivors are of somewhat different type physically than the eliminated. But nearly as many investigations have been made which show that, on the whole, the survivors are not physically different from those naturally eliminated. Again the studies of Weldon'® on Clausilia come first. Closely related to these is di Cesnola’s'® work on Helix. All three of these investigations agree in showing no sig- nificant difference in physical type between the general population before elimination and the selected survivors from that population after elimination. There was in Helix and in Clausilia laminata some reduction in vari- ability, but even that failed in another species of Clau- silia. Kellogg and Bell!" were not able to find any evi- dence that survivors and eliminated were different in respect of either type or variability under natural con- ditions, in the case of bees, or of the lady-bird beetle Hip- podamia. Pearl,’® in a much more extended series of ob- servations than those of Davenport, found no relation between the colors of chickens and their elimination by 14 Harris, J. A., AMER. NAT., Vol. XLV, pp. 314-318, 1911. 15 Weldon, W. F. R., Biometrika, Vol. I, pp. 109-124, 1901, and Ibid., Vol. IIT, pp. 290-307, 1904. 16 di Cesnola, A. P., Biometrika, Vol. V, pp. 387-399, 1907. 17 Kellogg, V., and Bell, R. G., Proc. Washington Acad. Sci., Vol. VI, pp. 203-332, 1904. 18 Pearl, R., AMER. NAT., Vol. XLV, pp. 107-117, 1911. No. 602] THE SELECTION PROBLEM 71 natural enemies. This result, it may be said, has been confirmed in subsequent years. Reighard,'® in one of the most beautiful experimental studies of natural selection which has ever been made, found that there was no rela- tion between the colors of coral, reef fishes and their elimination by natural enemies. While the researches which have been mentioned do not exhaust the literature, they are all for which time can be spared now, and they are fairly representative of the whole of the distinctly meager experimental and quanti- tative evidence regarding selective elimination. On the whole, the net result is not so clear-cut and outstanding as could be wished. If a scientific person came here from some other planet with an earnest desire to inform him- self about selective elimination, of which he had not before known anything, and read all the available real evidence on the point, he would be sure to come to some such conclusion as this: that in some cases natural elim- ination is certainly in some degree selective, while in other cases it certainly is not; and in the most favorable cases of all the selection is apparently not very rigorous. Grossly teratological abnormalities are eliminated. But the smaller deviations from type, which in theory ought to furnish the basis of selection, appear upon quantitative study less generally and sharply determinative of sur- vival than might reasonably have been expected theo- retically. The case regarding this first element of the theory of natural selection certainly seems far less strong, under the critical eye of experiment and measurement, than those of us who were nourished on Weismann, Romanes, and their like, would have supposed possible twenty years ago. Still the writer has no desire to be controversial about the matter and if any one is disposed to draw the opposite conclusion from the facts he is en- tirely welcome to. Let us now turn to the consideration of our second rule, which must be fully enforced if natural selection is to be an important factor in the causation of evolutionary 19 Reighard, J., Carnegie Institution, Publication 103, pp. 257-325, 1908. tz THE AMERICAN NATURALIST [ Vou. LI change. This, it will be recalled, was that the survivors must produce offspring which bear characters like those which had led to the survival. Or, to put the matter crudely, the survivors must transmit their characters to their offspring. In pre-Mendelian days this phase of the subject was always neatly and summarily disposed of by stating, as one of the facts on which the theory of natural selection rested, that ‘‘variations are inherited’’ or ‘‘like produces like.” Times have changed. We are a great deal less certain about that particular brand of inheri- tance which the theory of natural selection demands than we were before any one had taken the trouble to make experiments on heredity. The essential difficulty lies here. The differences upon which natural selection di- rectly operates are somatic differences, by hypothesis and in fact. Every worker in genetics has learned since the truly epoch-making researches of Johannsen”? to be ex- tremely cautious in assuming a priori that any particular somatic difference is so inherited. The writer has lately been experimenting with a char- acter which very well illustrates this point. A not infre- quent variation of the single comb in poultry is the ap- pearance, on one or both sides of the comb, of a small excrescence, known technically as a side-sprig. Indi- viduals exhibiting this variation have been selected for breeding purposes. But, so far as the experiments have yet gone, it does not appear that the offspring of such animals are any more likely to exhibit the variation than are the offspring of any random sample of single-combed fowls. Now suppose, for a moment, that in a state of nature the possession of a side-sprig on the comb gave a bird a distinctly better chance for survival than did a plain single comb. Those lacking the variation would then by hypothesis tend to be eliminated, but there is not the smallest indication that there would result any pro- gressive evolution towards a side-sprigged race. Now one might go on and review a great accumulation 20 Johannsen, W., ‘‘Ueber Erblichkeit in Populationen und in reinen Linien.’’ Jena, 1903. No. 602] THE SELECTION PROBLEM 73 of evidence from the work of de Vries,2! Dewar and Finn,*? Bateson,?* Lloyd,?4 and many others, showing the failure of the theory of natural selection to account satis- factorily for various observed happenings in evolution. It is not my purpose to do this. These facts are all familiar, and indeed have become commonplaces of bio- logical literature. We may, however, with some chance of profit try to generalize all this evidence. If we do so, the writer believes that the conclusion will be reached that natural selection is no longer generally regarded as the primary, or perhaps even a major, factor in evolution because of three general groups of facts, each well estab- lished by the common observation of many biologists. The first of these large facts is that all organisms possess in varying, but usually very large, degree the power of per- sonal, immediate, individual, somatic adaptation to the en- vironment. In consequence of this power of personal adap- tation the survival expectation of anindividualis not gen- erally and regularly a function of any static, single-valued relation between its somatic structure, habits, or physi- ology, on the one hand, and the impinging environmental Stresses on the other hand. Yet such a relation is im- plicitly assumed in that part of the theory of natural selec- tion which affirms a selective elimination on the basis of somatic characteristics. The second broad fact is that, even when selective elimination on the basis of somatic characteristics does occur, it does not follow generally and regularly that the somatic differences on which the selection acted will reappear in the progeny, or in short be inherited, actual experience having abundantly demon- strated that a very great many of such somatic differences are not inherited. The third large fact is that observation indicates that in many cases evolutionary changes have come about by relatively large, discontinuous steps, the 21 de Vries, H., ‘‘The Mutation Theory.’’ Chicago. 22 Dewar, D., and’ Fi inn, F., ‘‘The Making of Species. ’’ London. 23 Bateson, W., ‘í Problenis of Genetics.’? New Haven, 1913. 24 Lloyd, R. E., , ‘‘The Growth of Groups in the Animal Kingdom. ”? Lon- don, 1912, 74 THE AMERICAN NATURALIST [ Vou. LI new form being not merely fully differentiated at its first appearance, but also fully able to survive. Natural selection is, from the point of view of modern genetics, a somatic theory. It begins and ends in somatic differences. Except under the most unusual and rare of conditions natural selection can not possibly operate directly upon germ-cells. It must, from the very nature of the case, work only indirectly upon the germ through the soma. Now it is a historical fact that just so long as the study of heredity confined itself solely to the somatic results of the process, substantially no advance in knowl- edge was made. Only when it was clearly perceived that heredity is primarily and fundamentally a problem in the physiology of germ cells, and that the soma is, as some one has said, only the mechanism which the fertilized egg uses to produce another fertilized egg like itself, did we begin to make progress, Can we suppose that these con- siderations have no meaning or application in our at- tempts to solve the larger problem of organic evolution? If natural selection only could act directly upon germ cells we should have a different story to tell. The writer has lately been experimenting” with an agent which acts with extraordinary precision and definiteness in a selec- tive manner upon the germ cells, killing or inactivating the weak, and leaving only the strong and resistant to produce zygotes, and somata. This substance is alcohol. In the case at least of the domestic fowl the evolutionary effects produced by this substance are remarkable in their magnitude and definiteness. It is not possible here to go into details, but briefly it may be said that the first general result of the continued administration of ethyl or methyl alcohol, by the inhalation method, to the parents in poul- try, is to diminish progressively the fertility. But the smaller number of offspring formed by the surviving germ cells, after selective elimination in the gonads has occurred, are in every measurable respect distinctly and markedly superior to the normal individuals of the races 25 Pearl, R.. Proc. Amer. Phil. Soc.. Vol. LV, pp. 243-258, 1916; Proc. Nat. Acad. Sci., Vol. 2, pn. 380-384; Ibid., Vol. 2, pp. 675-683, 1916; Jour. Exper. Zool., Vol. 22, 1917. No. 602] THE SELECTION PROBLEM 76 from which they come. Here is real selection making real evolutionary progress, because its point of applica- tion is the germ and not the soma. That the same agent may produce an evolutionary change in the opposite direction in another organism, the guinea pig, as has been shown by Stockard?® to be the case, seems to mean, from the point of view of the present discussion, nothing more than that the direction and amount of any evolutionary change is, fundamentally, a function of two variables, the organism and the environment. If the same environ- mental stress produced the same evolutionary effect upon all organisms, then it would follow that all organisms in the same environment must necessarily be alike at the end of the process, which is, of course, not the case. II. THE EXPERIENCE or PRACTICAL BREEDERS Let us at this point leave our discussion of natural selection and turn to the other great aspect of the prob- lem, artificial selection. Here we shall tread on surer ground, first, because it is where the great bulk of the experimental work on the selection problem has been con- centrated, and second, because the considerable mass of reliable historical material about the origin and improve- ment of domestic animals and plants becomes available as a source of pertinent and critical evidence on the prob- lem. At the outstart it may be recalled that it was on the Supposed results of artificial selection, as set forth in the experience of practical breeders, that Darwin chiefly re- lied for objective evidence in favor of natural selection. In general this evidence has been accepted very un- critically by followers of Darwin. This is not strange in view of the fact that there have been, and are now, rela- tively few trained biologists who know anything at first hand about the practical breeding of animals for the show ring, advanced registry test, or any other purpose which involves necessarily the production of élite specimens which shall rank measurably with the best of the breed. 26 Cf. Stockard, C. R., and Papanicolaou, G., AMER. Nart., Vol. 50, pp. 65-88, 144-177, 1916, 716 THE AMERICAN NATURALIST [ Vou. LI This fact has led to some entirely unwarranted inclusions in the technical literature of biology. Statements in the agricultural press which were intended by their breeder- authors merely as harmless generalities on a subject about which they had not the slightest intention of being specific, have been accorded, by the laboratory evolu- tionist, the dignity and authority of detailed reports of actual breeding operations, and cited as valuable evidence on the problem of evolution. This confusion has played particular havoc in discus- sions of the selection problem because of the general and usually quite irresponsible use of the term ‘‘selection’’ by practical breeders. In the literature of live stock breeding the word ‘‘selection’’ has been, and is being used to-day, to designate, upon occasion, every known kind of breeding operation. To illustrate: a fancier who bred a new variety of poultry started with a mongrel male bird which happened to possess just the combination of characters which he wanted in his new breed, as the result of a previous series of indiscriminate crossings. This male was crossed with a female of a well established breed which possessed some of the desired characters. The daughters from this mating were back-crossed to their sire, the original male bird, and so in turn were his granddaughters. The granddaughters’ progeny consti- tuted the new breed, full blown and breeding tolerably true. This was an entirely legitimate, and indeed usual, way of making a new breed. But the point lies in the fact that the breeder who did all this always refers publicly to the series of matings which has just been described as **this process of selection’’! Since Darwin selection has been a word to conjure with amongst the practical breeders. In most cases without any comprehension whatever of the exact technical sense in which the term was originally used breeders have taken the word itself as a sort of fetish. Darwin?’ himself speaks of artificial selection as ‘‘the accumulation in one direction, during successive generations, of differences 27 Darwin, C., ‘‘ Origin of Species,’’ Chap. I, p. 26. No. 602] THE SELECTION PROBLEM 77 absolutely inappreciable by an uneducated eye.” Men in whose breeding operations selection in this sense demonstrably plays no significant part whatever, attribute to its magic power all improvement in their animals and plants. This is harmless so long as one understands that the ‘‘selection’’ is verbal and not biological. But unfor- tunately it has not always been so understood. Conse- quently we find the literature of evolution cluttered with a lot of utterly preposterous statements about domestic animals and plants, masquerading as valid evidence for the selection doctrine. So fixed in the minds of most biologists not acquainted with agricultural matters at first hand is the idea that the vast majority of improved varieties of plants and animals owe their origin, or their improvement, or both, to cumu- lative selection of slight differences, that it appears de- sirable to review briefly a few of the actual facts. One can not hope to do more than touch here and there in a great body of evidence, but at least representative cases may serve to indicate the general tenor of the whole. We may consider first the cultivated grape, of which there are over 1,300 recognized varieties grown in this country. Hedrick?* and his assistants have made a very careful and systematic study of the origin of these forms. After discussing the early history of the vine in this coun- try he says (p. 52): We have found that the wild grapes of the country, valued but uncul- tivated for two hundred years, became through mere transplanting from the woods into the vineyards . . . one of our most important fruits. Again in commenting upon the growth of the grape- growing industry he says (p. 62): The results achieved seem all the greater when one considers that many of the best varieties now grown are the first, and searcely any are further removed than the second generation from wild plants. But we should not be content with these general state- ments. Let us examine specifically the history of a well- 28 Hedrick, U, P., ‘The Grapes of New York.’ Albany, 1908. 78 THE AMERICAN NATURALIST [Vou. LI known and very excellent ‘‘improved’’ variety, the Con- cord. Of this grape Hedrick says (p. 219) The Concord is known by all. The most widely grown of the grapes of this continent, it also represents the dominant type of our native species and with its offspring, purebred and crossbred, furnishes seventy- five per cent. or more of the grapes of eastern America. He speaks of ‘‘the preeminently meritorious character of Concord which has enabled it to take first place in Amer- ican viticulture.’’ Now for the origin and history of this paragon. The seed of a wild grape was planted in the fall of 1843 by E. W. Bull of Concord, Massachusetts, from which fruit was born in 1849. The wild grape from which the seed came had been transplanted from beside a field fence to the garden in which there was at least another grape, the Catawba, and the wild vine was open to cross-pollination. One of these seedlings was named Concord and the variety was exhibited before the Massachusetts Horticultural Society in the fall of 1852. The new grape was introduced in the spring of 1854 by Hovey & toe ie of Boston. From the time of its introduction the growth of this variety in popularity was phenomenal. In 1865 it was eae a prize by the American Institute . . . as the best grape for cultivation. But where in the history of this ‘‘ best’’ American grape is the gradual accumulation of minute variations by selec- tion? The story is of the same sort for all varieties of grapes about whose origin anything is known; either they were chance seedlings, or F, hybrids, or F, segregates, grown under good conditions. The same general situation obtains in regard to the origin and improvement of other fruits besides the grape. One need only mention various cases briefly to recall them to mind. New and improved varieties of apples, plums, cherries, strawberries, etc., have originated either as chance seedlings, or as bud variations, or as hybrids. In their production selection, in the sense of the accumula- tion of minute favorable variations, has had no part. In the case of many fruits the mere fact of domestica- tion accounts for all the improvement over the wild type. A striking example of this is found in the case of the most recently domesticated wild plant, the blueberry. Mr. F No. 602] THE SELECTION PROBLEM 79 V. Coville? succeeded in discovering some years ago the two essentials for the successful cultivation or domestica- — tion of the swamp blueberry (Vaccinium corymbosum). These were found to be (1) an acid soil, and (2) a root fungus that appears to supply the plant with nitrogen. When these essentials are supplied, and the plant brought under cultivation, a marked improvement in the size and quality of the berries at once occurs. Further improve- ment is being made by hybridization, and by seeking superior natural variations to be used in such hybridiza- tion and for asexual propagation. In his latest paper Coville makes this statement, the significance of which in the present connection will be noted. Seedling plants, even from the largest berried wild parents, produce small berries as often as large ones. Selection, in the Darwinian sense, is playing no part in the improvement of the blueberry, so far as we can learn from the published records. The essential factors in the improvement are better conditions, hybridization, and the asexual propagation or superior natural variations. Let us now turn to other sorts of plants. In 1910 my colleague, Dr. Frank M. Surface, originated a variety of oats, known as Maine 340, which is superior to any variety which we have been able to find in the market and test. It is now widely grown in Maine. For the conditions of soil and climate in that state it is certainly to be regarded as the most highly ‘‘improved’’ variety known. Its origin and history are fully known, indeed are originally re- corded in the archives of this laboratory.® All of the thousands upon thousands of bushels of this Maine 340 oat which were grown this year are the lineal, unchanged descendants of one particular oat plant which Dr. Surface isolated in 1910. That original plant has simply been multiplied, by seed, year after year without selection of 2° Coville, F, V., U. S. Dept. Agr. Bulletin 193, 1910; Circular 122, 1913; Bulletin 334, 1915. 80 Cf. Surface, F. M., and Zinn, J., Me. Agr. Expt. Stat. Ann. Rept. for 1916, pp. 97-148. : 80 THE AMERICAN NATURALIST [Vou. LI any sort or kind, after the first isolation of the plant _ which originated the variety. The commercial variety of oats which comes the nearest under our conditions to equalling Maine 340 in yield and other desirable qualities is one known as the Early Pearl. The history of this variety has been given by Surface and Barber,*! who got it from the originator, Mr. R. L. Copeland. Mr. Copeland says: The first seed was obtained from a bunch growing by the roadside some twenty years ago, presumably from one seed. It was examined and showed such merit that it was cut and preserved for seed. Although the first seed was not secured by me personally, it soon after came into my possession. The oat seemed to possess excellent qualities and as it matured fairly early and had a pearly tint to the hull I gave it the name of Early Pearl. Since the beginning the oat has simply been grown by Mr. Copeland unmixed with other sorts. Here again that painstaking and laborious selection, by which the practical breeder is supposed to make the wonderfully improved and valuable varieties which we have, is conspicuous by its absence. And one must not for a moment suppose that this oat is not a wonderfully superior one. Any oat which will yield, year in and year out, good seasons and bad, not less than about 80 bushels to the acre, and at the same time possess a whole series of other desirable qualities, which are too technical to go into here, is in the front rank of the products of the breeder’s art. We find it to be surpassed only by some of our own new varieties. It outranks all other com- mercial varieties under our conditions. Many other examples of the same sort of histories of varietal origin and improvement might be given. But de Vries®? has covered the ground thoroughly and we need not stay longer over the plant side. The writer wishes to make clear before leaving the subject the reason for calling attention to these well-known matters. It is to emphasize that the ‘‘experience of practical breeders’? 31 Surface, F. M., and Barber, C. W., Me. Agr. Expt. Stat. Ann. Rept. for 1915, pp. 137-192. 82 de Vries, H., ‘‘Plant Breeding.’’ Chicago, 1907. No. 602] THE SELECTION PROBLEM 81 shows that the principle of the gradual accumulation by continued selection of minute somatic variations has had no essential part in the origin or amelioration of cer- tainly a great many of the best varieties of agricultural plants which we have to-day. The essential factors which have been involved in the production of our best fruits, grains, vegetables, flowers, etc., have been (1) the im- proved conditions of domestication, (2) mutations, lead- ing at once to new and better forms, (3) hybridization, which by new combinations of characters and as a result of heterosis®* has led to amelioration, and (4) the puri- fication of previously mixed races or varieties by selective sorting. It is to the overwhelming importance of one or a combination of these factors that the ‘‘experience of breeders’’ points and not to Darwinian selection. But what of the animal side? Here the true facts are much more difficult to get at; in part for reasons which have been developed earlier in this paper, and in part for the reason that the making of new breeds of domestic animals is no longer going on to any extent except in the smaller sorts such as poultry. As has been pointed out elsewhere,** this is primarily a result of the great devel- opment of the system of pedigree registration, which puts a ban on cross breeding in cattle, horses, ete. _ So then let us take as our first example one from poul- try breeding where unequivocal facts are available. In his ‘‘Organic Evolution’? Metcalf% makes the following statement: The extent of the modification produced by artificial selection is very _ great in many eases. Notice the common domestic chickens, in which the different breeds differ from one another to such a degree that if they occurred in nature the several kinds would be referred not only to different species, but to different genera. Compare the slender “ game” which most closely of all resembles the ancestral “jungle fowl,” with the heavy “Brahma” or “ Cochin-china,” or with the ting tailed “ Jap- anese” cocks, or with the little “bantam.” _ 83 East, E. M., and Hayes, H. K., U. S. Dept. Agr., Bur. Plant Industry, Bulletin 243, 1912, 34 Pearl, R., ‘€ Modes of Research in Genetics.’? New York, 1915. se Metcalf, M. M., ‘‘Organie Evolution.’’ New York, 1904. 82 THE AMERICAN NATURALIST [Von. LI All these forms, according to Metcalf’s idea, have been produced by selection from the jungle fowl. For a num- ber of years past the writer has been interested in col- lecting evidence regarding the making of new races and varieties of bantams, for the purpose of seeing what part selection, in the Darwinian sense, probably plays in the matter. The problem was given enhanced interest and significance by the appearance of the important paper by Punnett and Bailey, which showed that in respect of body weight bantams are differentiated from large fowls by at least three genetic factors, and that in a cross be- tween a bantam breed (Seabright) and a larger breed (Hamburg) the inheritance of body weight is typically Mendelian. This latter result the writer had found to be true, using different breeds from those employed by the English workers, but had not worked out in detail the exact mechanism of the inheritance. New varieties of bantams are all the time being produced and exhibited at poultry shows. Broadly speaking, it is nearly true that for every different variety of large fowl a corresponding bantam variety either has been produced or is sure to be soon. In view of this great activity in the making of new varieties it seemed that an excellent opportunity was offered to find out how the expert bantam fancier really turns the trick. So the writer has corresponded - with bantam fanciers in all parts of the world and in this way has accumulated a large amount of interesting ma- terial. One question was always asked, and always in the same form. This was: Do you know of any case in which a stable Tace, variety or breed of bariams, which bred true indefinitely in respect to bantam size of body, was originated or created solely by selection of small sized individuals of a large race, variety, or breed of fowls, without any crossing in of bantam blood? If so, please give a detailed account of the cireum- stances. ; 36 Punnett, R. C., and Bailey, P. G., Jour. Genetics, Vol. IV, pp. 23-39, 1914. No. 602] THE SELECTION PROBLEM 83 The answers to this question were negative in every case except one. In that case the correspondent was unable to cite any specific instances illustrating his con- tention, but thought on general principles that it must be so. In other words, his belief in the selection theory, like that of Weismann, was based on wholly other grounds than demonstrative evidence. The general tenor of the answers to this question is well indicated by a statement of Mr. J. F. Entwisle,*7 who is probably the greatest authority on bantam breeds of poultry now living. He stated that it was thought by some that proper little bantams could be bred from large fowls without any admixture of bantam blood. . . . If such ean be done, then our thirty- odd years’ experience of bantam “manufacturing” counts for very little. We’ have lived to see the manufacture of some forty varieties, and none without crossing, so far. If time permitted, the citation of the detailed history of the making of several breeds of bantams would show very clearly that Darwinian selection plays an extremely minor and unimportant part in the process as it is actu- ally performed. Large breeds of poultry show the same thing. Two of our most important breeds, the Wyan- dottes and the Orpingtons, are of comparatively recent origin. The facts as to their origin are well known and the essential biological factors concerned in both cases are the same, namely, hybridization as a start, followed by close inbreeding of desired segregating types. In other cases new varieties of poultry have appeared as sudden mutations by loss of factors. Such is the origin of the White Plymouth Rock and the White Cornish, for example. III. SELECTION EXPERIMENTS ; We now come to the third class of evidence on the selec- tion problem, namely, that afforded by controlled ad hoc 31 Entwisle, J. F., Poultry (London), Vol. 30, P 767, 1912. 38 Thé “we”? hate’ is editorial. Mr. Entwisle is speaking of his own per- Sonal experience. 84 THE AMERICAN NATURALIST [ Vou. LI experimentation. Here the facts are so recent and so well known that a detailed review is unnecessary. I shall, therefore, only attempt to deal with them, for the most part, in a general way. Careful and critical selection ex- periments may be said in general to have given opposite results, according to whether the hereditary factors for the character which formed the basis of the selection were or were not positively known to be in a homozygous condition in all the individuals of the race experimented with. If the form used constituted a ‘‘pure line’’ in the strict sense of Johannsen’s conception, so that every in- dividual was surely known to be strictly homozygous for some state or condition of the character selected, and the mode of reproduction was such as automatically to retain and continue this condition, then the results of con- tinued selection have in most cases been wholly negative so far as the production of any change in type is con- cerned. This is shown by the work of the Svalof Sta- tion®® with various cereals, of Johannsen*® with beans, of Jennings*! with Paramecium, of Hanel*? with Hydra, of Vilmorin? with wheat, of Ewing‘! with Aphis, of Surface and Pearl** with oats, of Fruwirth*® with lentils, peas, soy-beans and lupines, of Lashley‘? with Hydra, of Agar* with Simocephalus, and of others. 39 Cf. de Vries, H., loc. cit. passim, and Newman, L. H., ‘‘ Plant Breed- ing in Seandinavia.’’ Ottawa, 1912. : 40 Johannsen, W., ‘‘Ueber Erblichkeit in Populationen und in reinen Linien.’’ Jena, 1902. 41 Jennings, H. S., Proc. Amer. Phil. Soc., Vol. 47, pp. 393-546, 1908; and Amer. NAT., Vol. 46, pp. 487-491, 1912. 42 Hanel, E., Jenaische Zeitschr., Bd. 43, pp. 321-373, 1908. 43 Cf. Hagedoorn, A. L., and Hagedoorn, A. C., Zeitschr. ind. Abst. Ver. Lehre., Bd. XI, pp. 145-183, 1914. 44 Ewing, H. E., Biol. Bul, Vol. 26, pp. 25-35, 1914; and Ibid., Vol. 27, pp. 164-168; and Ibid., Vol. 31, pp. 53-112, 1916. 45 Surface, F, M., and Pearl, R., Me. Agr. Expt. Stat. Ann. Rept. for 1915, pp. 1-40. 46 Fruwirth, C., Zeitschr. f. Pflanzensiichtung, Bä. III, pp. 173-225, 1915. 47 Lashley, K. S., Jour. Exp. Zool., Vol. 19, pp. 157-210, 1915; and Ibid., tics 20, pp. "19-26, 1916. gar, W. E., Phil. Trans. Roy. Soc., London, B, Vol. 205, pp. 421-489, No. 602] THE SELECTION PROBLEM 85 So far as the writer is aware the only important ex- ceptions to the general rule exemplified by the above- cited researches are those obtained by Jennings“ in Dif- flugia, and by his students, Middleton®® with Stylonychia, and Stocking®! with abnormalities in Paramecium. The facts in the case of these exceptions are beyond question. Just what their correct interpretation is does not seem to be so clear. Jennings himself (loc. cit., p. 529) has ex- pressed some doubt as to the significance of Miss Stock- ing’s results so far as concerns their relation to normal reproduction. Morgan®? has suggested that Jennings’s Diffiugia results may possibly be due to a sorting out of genetic diversities in the germ plasm, which came about from earlier conjugations, the material thus not repre- senting a strictly homozygotic pure line. If this sugges- tion should prove to be valid the Diflugia work would fall at once into the same category as the cases of sorting out of pure lines from a mixed population by selection, with which the studies of Johannsen, de Vries and Jen- nings himself have made us familiar. There is another point in connection with this extremely interesting and important investigation on Difflugia which seems to the writer, in the light of his own experi- ence in breeding, of really extraordinary significance. One of the longest and most crucial selection experiments. in the whole series was that for diverse numbers of spines in Family No. 326. During the first six periods after Selection was begun in this experiment (p. 488) ‘‘no progress was made by selection.’’ Then the basis of selection was changed. This change is described by Jen- nings in the eee words (p. 489): After this time selection was based to a considerable extent on past performance. By this time many of the existent individuals had pro- 49 Jennings, H. S., Genetics, Vol. 1, pp. 407-534, 1916. 50 Middleton, A. R., Jour. Exp. Zool., Vol. 19, pp. 451-503, 1915. 51 Stocking, R., Jour. Exp. Zool., Vol. 19, pp. 387-449, 1915. 52 Morgan, T. H., ‘ʻA Critique of the Theory of Evolution.’’ Princeton, 1916, 86 THE AMERICAN NATURALIST [Vou. LI duced several offspring. Where a parent of the low group had been found to bring forth high progeny, that parent was removed. Simi- larly, if a parent with a high number of spines is found to produce offspring with low numbers, this parent was removed. Thus in the low group we gradually tend to accumulate a set of individuals (1) which in the past have produced progeny with low numbers of spines; (2) whose ancestors for several generations back are individuals with low numbers of spines. In the ligh group the reverse conditions are fulfilled. From the time this change in method was made until the end of the experiment, selection produced a marked effect, differentiating clearly high spine groups and low spine groups. The significance of this result seems to me to lie in the fact that with the new method of selection described by Jennings the ultimate basis of selection was changed from the soma to the germ. Because after the change in method what primarily determined the selection of any particular individual for further reproduction was not its own spine number, but instead its demonstrated ability to produce offspring with a particular spine number (or nearly that number). A low spine parent was allowed to survive and reproduce not alone or primarily because _ it had few spines but only when in addition it was surely known to produce low spine progeny. This is indeed a different basis than that which merely selects individuals because of their somatic spine number and nothing else. It abruptly and completely transfers the selection from soma to germ. As has already been pointed out earlier in this paper (p. 5), there can be no question about the efficacy of selection which operates on a directly gametic rather than a somatic basis. This idea of making the basis of selection the ability to transmit to the progeny the desired quality is one which the writer has for a good many years strongly advocated as the only really useful or hopeful method in the practical breeding of higher animals. In his own work with poultry? it has been crowned with the highest practical success. Beyond 53 Pearl, R., AMER. Nar., Vol. XLIX, pp. 306-317, 1915. No. 602] THE SELECTION PROBLEM 87 doubt or question the reason for the success is because by this method we select directly the kind of gametes we want. In-ordinary Darwinian selection we select the kind of somata we want, and trust blindly that a wise providence has implanted in them the sort of gametes we need in order to get further somata like those we selected. And as in so many other troublesome affairs in this vale of tears, too often we find in the outcome that our trust has been misplaced! It seems to me that whether Morgan’s suggestion re- garding the highly interesting and important results with Diflugia is true or not the fact dealt with in the pre- ceding paragraph objectively paraHels exactly the phe- nomena which one sees when he isolates pure lines from a mixed population (e. g., in oats). He finds some individuals which are somatically what he wants but which fail to produce the sort of progeny he is looking for. These he discards. Others produce progeny which are like themselves. They transmit** their qualities and hence are retained. Having considered the results of experiments on selec- tion in pure lines we may turn to similar experiments with sexually reproducing organisms. Here the facts are, in the main, no less clear, but they are, on the whole, exactly opposite in their sense, at first sight at least. For in most of the experiments of this sort selection has been attended with an alteration of the type in the direction of the selection. This has been the case in the writer’s*® experiments on egg production, in the domestic fowl since 1908, in the work of Smith5* and others with maize in 54 Lest there should be any misunderstanding I may say that I am fully aware that the old idea of heredity as a direct and material transmission of personal qualities from generation to generation is wholly incorrect, out-of- date, and pedagogically pernicious. I find it, however, extremely çonte- nient, saving of breath and good white paper, and, with this explanation, I hope permissible, to use ‘‘transmit’’ as a technical genetic term meaning ‘*to possess gametes of such sort as will produce in the progeny.’’ 55 Pearl, R., AMER. Nat., Vol. XLIV, pp. 595-608, 1915. 56 Smith, L. H., Ill. Agr. Expt. Stat. Bul. 128, 1908. 88 THE AMERICAN NATURALIST [ Vou. LI Illinois, in the work of Pearl and Surface®* with maize, in MacDowell’s®* work on bristle number in Drosophila, in the work of Zeleny and Mattoon”? with the bar eye of Drosophila, in the experiments of Castle and Phillips® with hooded rats, and in the work of some others. The difficulty with all these experiments is not in the facts but in their interpretation. Before taking up this question of interpretation, how- ever, it seems desirable to point out certain objective features which the results of these experiments have in common. The first is that they all occur (with the ex- ception of the Jennings cases already discussed) in sexu- ally reproducing organisms, not certainly known to be homozygotie with respect to all the factors which may be concerned in the production of the selected characters, and subject to a mixing of germ plasms in each genera- tion. The second is that when any result at all follows selection in most if not all of the cases it comes quickly (i. e., in a few generations), is relatively large in amount, and either no further change follows further selection or if it does occur it is again sudden and large in amount. This was true in the work with high producing lines of hens, of all work with maize by the ear-row method, and of MacDowell’s with Drosophila, and also, as MacDowell’! has so clearly demonstrated, in Castle’ S experiments with rats. A third point which strikes one is that in many of these successful cases of selection the basis of the selec- tion has been fundamentally gametic, that of ‘progeny performance,” rather than solely somatic. Individuals are selected for further multiplication which have demon- strated their ability to produce progeny bearing the de- sired somatie qualities. This. was certainly the case in 57 Pearl, R., and Surface, F. M., Maine Agr. Expt. Stat. Ann. Rept. for 1910, pp. 249-307. 58 MacDowell, E. C., Jour. Exp. Zool., Vol. 19, 1915. 59 Zeleny, C., and Mattoon, E. W., Jour. Exp. Zool., Vol. XIX, pp. 515- 529, 1915. 60 Castle, W. E., and Phillips, J. C., Carnegie Institution, Publ. 195, 1914. 61 MacDowell, E. C., Amer. Nar., Vol. 50, pp. 719-742, 1916. No. 602] THE SELECTION PROBLEM 89 the Illinois corn work, as Surface’? very clearly demon- | strated. The writer? showed a number of years ago that it was the basis of success in selecting poultry for egg production. One strongly suspects it to be true in the other cases, though because the importance of the point has not been perceived, evidence in the published ac- counts is lacking on which to make any positive statement. What is the correct interpretation of these favorable results? Two opposed opinions are held. Fortunately the heat of the controversy has been intense enough to produce some distillation and we at least have the issue very clearly and sharply defined. On the one hand it is held, because there has been an alteration of type in point of time coincident with successive selections, that selec- tion on the basis of personal somatic qualities only, as such, in and of itself, has altered hereditary factors in the germ plasm. This view makes selection a cause of genetic variation,®* a total reversal of the position held by Dar- win and most of his followers. The opposing view is that selection can only be successful in altering the type when hereditary determiners to produce the desired somatic qualities are already present inthe germplasm. Selection, on this view, has nothing whatever to do with the causa- tion of the variation, and is wholly powerless and with- out effect on the race unless either (a) the basis of the selection is directly gametic, by means of progeny per- formance test, or (b) the soniatically selected individuals happen by good fortune to carry the necessary hereditary determiners in their germ plasm. The opposition here really goes very far back. It is the world-old fight between heredity and environment, nature and nurture, germ and soma. One side believes first, that hereditary determiners or factors fluctuate reg- e 62 Surface, F. M., IV° Conf. internat. de Génétique, Paris, 1911, pp. 221- 35. 63 Pearl, R., AMER. Nar., Vol. XLV, pp. 321-345, 1911. 64 Castle, W. E., Sci. Mo., Vol. 2, p. 91, 1916, and Castle, W. E., and Phil- lips, J. C. , Onmnegle Inst. Publ. 195, p. 31. 90 THE AMERICAN NATURALIST [ Vou. LI ularly and frequently, if not indeed usually, and in high correlation with somatic characters; second, that mixing of germ plasms in fertilization alters hereditary deter- miners mutually and hence is, in and of itself, a cause of genetic variations, and, therefore, third, that a purely ex- ternal agent, the continued selection of personal somatic qualities, will alter the germ plasm.*® The other side believes first, that the germ plasm is fundamental and remarkably conservative, basing this belief on such observations®* on the one hand as those of Walcott that pre-Cambrian annelids, snails, cr and algæ were in many cases so iky forms living to- day as to belong to the same genera, though a period of time variously estimated at from 60 to 200 million years has elapsed; and, on the other hand, those of Wheeler on ants enclosed in amber two million years ago but morpholog- ically identical with forms living to-day. It believes, second, that when the germ plasm changes it does so as a result either of wholly internal physiological causes, or of very extraordinary environmental stresses acting directly upon the germ cells; third, that mixing of germ plasms, in and of itself, does not mutually alter hereditary determiners, basing this belief on the regularity, con- stancy and cleanness of typical Mendelian segregation; and fourth, that selection only acts as a mechanical sorter of existing diversities in the germ plasm and not as a cause of alteration in it. . The alternative views have been presented. In the present state of knowledge nothing is to be gained by mere assertions of opinion as to which more nearly repre- sents the truth. But one may at least advance the view, 65 In his last paper MacDowell has justly emphasized the scientific futility of defining and codifying the opponent’s position in a controversy, so that one may the more neatly bowl him over. I am very sensible of the force of this point, and therefore have been at great pains in the preceding sentence to make only such statements as can be supported by verbatim quotations from the literature. Since, however, it seems to me equally important to the personal element out of scientific controversy, so far as may be, I do not make the specific references. 66 Cited from Loeb, J., ‘‘ The Organism as a Whole.’’ New York, 1916. No. 602] THE SELECTION PROBLEM 91 to which the whole of this paper has been leading, and which one may hope will be heartily concurred in by both sides in the selection controversy, that the great outstand- ing need in research on the problem of evolution in gen- eral, and of selection in particular, is more, and more searching, investigations as to the causes of genetic (fac- torial) variation. That both sides realize this need, and are all the time bending more and more energies to its selection, is indeed cause for congratulation and augurs well for the future of that branch of biological science in which America has taken a leading place. MENDELIAN FACTOR DIFFERENCES VERSUS REACTION SYSTEM CONTRASTS IN HEREDITY. I T. H. GOODSPEED anp R. E. CLAUSEN The literature on species crosses shows clearly that all gradations occur as regards fertility from complete steril- ity to what is apparently complete fertility. Recombina- tions may, therefore, in some cases, occur freely between species that appear to be distinct. Lotsy (1913) in par- ticular has shown this to be true for a number of species crosses in Antirrhinum and doubtless such instances might be multiplied considerably. But even in these cases the behavior of the hybrid progeny in subsequent genera- tions indicates that there is a possibility that some of the recombinations do not form functional reaction systems because of the discordant elements they possess. This is shown in the entirely new characteristics which a certain portion of the population may exhibit and in the peculiar ratios which are sometimes obtained in segregation. Such results only bear out more completely the conception that for any recombination of elements to be completely func- tional they must together form a harmonious reaction system. Even Detlefsen’s (1915) results with the cavy species cross which gave in F, sterile males and fertile females may be brought into line with such a physiologi- cal conception perhaps better than by trying to account for it on the basis of any definite number of Mendelian factors. At least in the cavy cross the attempt to account for the results in this latter manner demonstrated that the number of factors concerned must be relatively large, and this is precisely what would be expected on the basis of recombinations involving whole choromosomes or sec- tions of chromosomes carrying with them, perhaps, many discordant elements. If we recall the observation that no crossing-over occurs in the sex heterozygote (Morgan 92 No.602] | MENDELIAN FACTOR DIFFERENCES 93 et al, l. c.), presumably the male in this case, then it would be possible in the male cavy to secure only recombinations involving whole chromosomes, or in other words involv- ing the building up of reaction systems with a large num- ber of interchanged factors. In the female, however, these recombinations might include, in addition to those resulting from redistribution of whole chromosomes, cross-over gametes resulting from the exchange of sec- tions of chromosomes. Such gametes might conceivably contain fewer discordant elements or they might be dis- tributed in such a way as to distrub the reaction system thus formed less profoundly. Such data of course need first to be reexamined from this viewpoint before any definite conclusions can be reached, but the list of species hybrids which Detlefsen gives which result in sterile males and fertile females would seem to indicate that this is a phenomenon connected in some way with the ob- served lack of crossing-over in the sex-heterozygote already demonstrated in the work with Drosophila and the silkworm (Sturtevant, 1915). Until we know more about the fundamental basis of crossing-over and the factors affecting it, it is idle to speculate on such differ- ences in behavior as are shown by males and females in the species crosses mentioned. The insistent way in which the species hybrids between Tabacum and sylvestris point to the conception of the Mendelian reaction systems as units in themselves is of interest because of the broad and far reaching conse- quences which follow the application of such anidea. For if the nature of the progeny of these partially sterile hybrids as grown through several generations points to anything, it is that the abortive ovules and pollen grains represent a selective elimination of certain types of re- combinations. Obviously, then, the presence of any con- siderable proportion of sterile ovules or pollen grains in plant material may be a consequence of hybridity and that of a rather profound type involving reaction systems which are more or less specific in their nature and in part incompatible with each other. The importance of this 94 THE AMERICAN NATURALIST [ Von. LI fact has, perhaps, not been sufficiently emphasized in work involving material which displays such partial ster- ility. Obviously, however, it is impossible to regard ratios obtained from such material as of any significance, unless it be possible to demonstrate definitely the nature of such gametes as fail to function. It is, therefore, not strange that species hybrids as a class require a some- what different sort of treatment from that applicable to intervarietal crosses involving relatively few factor dif- ferences. In the extensive work which has been done with the various species and forms of @nothera this influence of partial sterility has undoubtedly played an important part, but at the same time one which has not been clearly defined. Jeffrey (1914) in particular has sought to es- tablish the hybrid nature of O. Lamarckiana on the basis of the high percentage of abortive pollen grains which are found quite generally in the genus, and Heribert-Nilsson (1912) has attempted to analyze the material from a Men- delian standpoint. These attempts have not, however, led to a consistent explanation of the results observed, although they have established certain peculiar condi- tions in Enothera which practically preclude the applica- tion of a rigid Mendelian analysis to such a type of be- havior. On the other hand, a critical examination of the phenomena displayed by the various forms of Ginothera clearly indicates that these belong to several distinct cate- gories, and not to one as is very generally assumed in discussions bearing on this subject. In this brief discus- sion we propose to classify them roughly as follows: - 1. Strict factor mutations arising from unknown causes but clearly referable to specific germinal changes in iso- lated loci in the hereditary system. These display a simple, consistent Mendelian behavior when tested with the forms from which they arise. They usually show relatively simple and definite character differences when compared with the parent forms rather than complex dif- ferences throughout. 2. Segregation phenomena of a complex type resulting No. 602] MENDELIAN FACTOR DIFFERENCES 95 in the constant and continual production of a number of distinct forms which display for the most part a compli- cated behavior when tested with the parents from which they arise and with other forms. Such forms are often widely different from the parent forms in all or nearly all of their characters. 3. Chromosome duplications resulting in duplication of one or more or even all of the chromosomes to the ex- tent of tetraploidy in some forms. The first of these categories may be rigidly distin- guished from the other two, both of which may be funda- mentally expressions of some inherent condition in the ‘‘mutating’’ individual. The existence of this first type of mutation can scarcely be denied in the face of the ex- tensive evidence accumulated both on the plant and on the animal side. In addition to this long series of past observations, we have now the extensive work of Morgan and his associates (l. c.) in which the origin of over a hundred such factor mutations has occurred under obser- vation in Drosophila cultures. These investigations indi- cate clearly that such mutations are fundamentally de- pendent upon actual changes in the germinal substance. They are the loss-mutations of genetic literature, but with the abandonment of the presence and absence hypothesis such a classification, of course, loses its significance. Practically always only one locus is involved in such a change, and the new form displays a consistent alterna- tive behavior when tested with the form from which it arose. Such mutations are usually recessive, although a few dominant ones have been secured. When obtained in pure culture they show no tendency to revert to the parent form, at least not more frequently than they tend to change i in entirely different directions. They are rela- tively rare, they do not recur in any definite considerable ratio in any strain, and they are not clearly due to any specific cause. In @nothera, rubricalyx appears clearly to be a dominant mutation of this type (Gates, 1914). There seems to be little occasion for confusing these strict 96 THE AMERICAN NATURALIST [ Vou. LI factor mutations with the more complex type of behavior included under the second and third categories. — When the second category is considered we are met with the task of harmonizing a large mass of rather con- fusing data, and it is perhaps true that this can not be done successfully at the present stage of our knowledge. There are not lacking, however, as many others have pointed out, a number of significant facts which are at least as logically explainable on the basis of hybridity as on assumptions of general germinal changes. If the re- sults which have followed species crosses are to be ex- tended to the type of behavior displayed by Lamarckiana, then it is clear that the numerical ratio in which segrega- tion occurs is of no particular significance except in so far - as its constancy indicates that it is due to a specific be- havior in gametogenesis. Lotsy (1912) in particular has pointed out that in Antirrhinum species crosses, races may be secured which behave very much like some forms of Œnothera with respect to the segregation ratios ob- tained. The significant facts with which we have to deal are apparently, first that in @nothera these ‘‘mutations’’ affect the sum total of the characters of the individuals, i. e., they are dependent on complex germinal differences when compared with the parents, and second that these ‘‘mutations’’ continually recur within certain races in fairly constant ratios. These are facts which are just as simply explained on the basis of a complex type of segre- gacion in which many of the systems formed contain dis- cordant elements and therefore fail to develop as to refer them to actual change in germinal substance. Moreover, the former explanation harmonizes the results obtained with the simpler category of strictly Mendelian phe- nomena. That this view of the @nothera situation has something in its favor beyond the known general occurrence of par- tial sterility in this genus is shown by some of the results of hybridization between the various forms of Œnothera. The frequent appearance of F, populations consisting of distinct forms is usually taken to be an expression of No. 602] MENDELIAN FACTOR DIFFERENCES 97 segregation in the gametic series of such forms. Beyond this the Gnothera phenomena display a remarkably orderly behavior, and, although complex in nature, such orderliness points strongly to some kind of definite seg- regation dependent upon the hereditary constitution of the forms involved rather than upen any change in germi- nal substance expressed in the gametic series. Lamarcki- ana, for instance, produces constantly a small percentage of nanella as well as a number of other forms. It is, therefore, necessary to assume merely that nanella ga- metes make up a small percentage of the gametic series in Lamarckiana. When two nanella gametes meet, a nanella individual is produced, and it breeds true, as might be ex- pected. When, however, a nanella gamete meets a La- marckiana gamete a Lamarckiana individual is produced, which need not necessarily differ in its behavior with re- spect to the production of nanella from other Lamarcki- ana races, since practically all of the combinations in- volving nanella elements would fail to develop, or in case some few of these combinations did develop, they might produce other characteristic forms unlike either Lamarckt- ana or nanella. This conception is in part borne out by the fact that nanella appears to differ from Lamarckt- ana not only in stature, but also in other characters, as is shown in hybridization phenomena involving this form. From this standpoint also, the occurrence of nanella in the progeny of a wide range of forms is perfectly intel- ligible. When nanella, which apparently really breeds true, is crossed back with Lamarckiana a small propor- tion of the progeny is of the nanella type and the rest are Lamarckiana. This proportion is apparently greater than that normally obtained from selfed Lamarckiana, and this is merely a consequence of the realization of the gametic ratio of nanellas in the Lamarckiana series as a phenotypic ratio. The true breeding of the forms thus resulting, to the extent that they breed true, is a natural consequence of the germinal constitution of such forms. On the other hand, when rubrinervis, which never pro- duces nanella as a ‘‘mutant’’ (Gates, 1915), is crossed a: ieee 98 THE AMERICAN NATURALIST [ Von, LI with nanella, nanella does not appear in F, and in F, appears in a fairly definite ratio in the progeny of the subrobusta forms thus produced (de Vries, 1913), whereas the Lamarckiana forms produced breed true, after their usual fashion. Obviously this behavior is simply a conse- quence of the fact that rubrinervis does not produce a -gametie series containing nanella, and, therefore, nanella can not appear in F,. The impressively orderly behavior throughout in these hybridization phenomena is an elo- quent testimony of the existence of a casual agency of a simpler nature than that called for under the hypothesis of actual, germinal changes. Those phenomena included in the third category and dealing with chromosome duplication may be referred to the same type of iregular behavior in gametogenesis as is concerned in the production of the other ‘‘ mutants.’’ They are of interest most particularly from our stand- point with regard to the structural relations which they display i in comparison with the other forms. Thus there is a series of lata forms dependent on the duplication of one chromosome in Lamarckiana. They are not identical within the series, but at least three or four are known which differ from one another in a characteristic fashion. This may well be dependent on the particular chro- mosome pair which is concerned in the duplication as Gates (1908) has suggested. According to the chromo- some view of heredity this duplication has the effect of altering the proportions of the various elements in the reaction system, and naturally in a delicately balanced system such alteration results in a change in the somato- genic processes in a definite direction, the latter depend- ent upon which particular elements are increased. Since a whole chromosome with presumably a large set of fac- tors is involved, it should follow that the entire set of characters of the plant would be affected, as appears actually to be the case. The important point in connec- tion with these phenomena is the apparent consequence that phenotypic changes may be dependent merely on an alteration in the relative quantitative proportions of the No.602] © MENDELIAN FACTOR DIFFERENCES 99 Mendelian units making up the reaction systems. Ap- parently this will account satisfactorily for the orderly resemblance of these chromosome ‘‘mutants’’ to each other and to the forms from which they arise. Conceiv- ably the great body of data on which the mutation theory is based would, for the most part, find a simpler expla- nation along the lines thus indicated. SuMMARY Summarizing briefly the content of this paper, the fol- lowing facts have, on the experimental side, been pre- sented with reference to a species hybrid: 1. N. sylvestris when crossed with various varieties of N. Tabacum gives F, hybrids which are replicas on a large scale of the particular Tabacum variety concerned in the cross. 2. The F, hybrids of sylvestris and Tabacum produce a small number of functional ovules which represent the sylvestris and Tabacum extremes of a recombination series, the great majority of the members of which fail to function because of mutual incompatibility of the ele- ments of the two systems. : 3. Back crosses with sylvestris give sylvestris and aberrant forms, and of the two the sylvestris alone are fertile and breed true. On the other hand, back crosses with Tabacum produce apparently only Tabacum forms of which some are completely fertile and continue to pro- duce only Tabacum forms. On the theoretical side the following conclusions have been drawn and their application indicated: 1. As a consequence of modern Mendelian develop- ments, the Mendelian factors may be considered as mak- ing up a reaction system the elements of which exhibit more or less specific relations to one another. 2. Strictly Mendelian results are to be expected only when the contrast is between factor differences within a common Mendelian reaction system as is ordinarily the case in varietal hybrids. 100 THE AMERICAN NATURALIST [ Vou. LI 3. When distinct reaction systems are involved, as in species crosses, the phenomena must be viewed in the light of a contrast between systems rather than between specific factor differences, and tke results obtained will depend upon the degree of mutual compatibility displayed between the specific elements of the two systems. 4, Sterility in such cases depends upon non-specific in- compatibility displayed between the elements of the sys- tems involved, and the degree of this sterility depends upon the degree of such incompatibility rather than upon a certain number of factors concerned in the expression of such behavior. 5. The consequences of the application of such a con- ception to the complex type of behavior in @nothera are pointed out, and the suggestion is specifically made that the type of behavior exhibited by Lamarckiana and its segregants in hybridization may be referred to such com- plex system interactions. LITERATURE CITED Bateson, W., and Punnett, R. C. 1912. Experimental Studies in the Physiology of Heredity. Proc. Roy. Soc. London, Rept. Evol. Com., : 6-18. Conklin, E. G. 1908. The Mechanism of Heredity. Science, N. $., 27: 89-99. Correns, C. 1913. Selbststerilitit und Individualstoffe. Biol. Centralb. 33: 389- 423. Detlefsen, J. A. 1914. Genetic Studies on a Cavy Species Cross. Publ. Carnegie Inst. Washington, No. 205, pp. 1-134, 1914. East, E. M. ; ; 1915. An Interpretation of Sterility in Certain Plants. Proc. Amer. Phil. Soc : 70-72. East, E. M., and Hayes, H. K. 1912. Heterozygosis in Evolution and in Plant Breeding. U. S. Dept. Agr. Bur. Pl. Ind., Bul. No. 243: 1-58. Gates, R. R. 1908. A Study of Reduction in Cnothera rubrinervis. Bot. Gazette, 45: 1-3 1914, Breeding Experiments which Show that Hybridization and Mutation are Independent Phenomena. Z. Abst. Vererbdl., 9- 1915. The Mutation Factor in Evolution. No. 602] MENDELIAN FACTOR DIFFERENCES 101 somes f Me 1913, a the Partial Sterility of Nicotiana Hybrids Made with N. sylvestris as a Parent. Univ. Calif. Publ. Botany, 5: 189- 1915. Parthenogeness, So AA and Agee in Nico- tiana. Unw. Mee ght ubl. Botany, 5: 249-272 Goodspeed, T. H., and Ayres, A. 1916. On “the Partial Sterility of Nicotiana Hybrids Made with N. sylvestris as a Parent. II. Univ. Calif. Publ. Botany, 5: in pres Heribert-Nilsson, N. 1912 Die Variabilität der Gnothera Lamarckiana oa: das Prob- lem der Mutation. Z. Abst. Vererbl., 8: 89-2 Jeffrey, E. C, 1914. Spore Conditions in Hybrids and the Mutation Hypothesis of de Vries. Bot. Gazette, 58: 322-336. ‘Jennings, H. S. 1914. Development and Inheritance—Its Relation to the Germ. The Johns Hopkins Univ. Circ., 1914, Lotsy, J. P, 1912. Versuche über Artbastarde und Betrachtungen über die Möglichkeit einer Evolution trotz Artbestiindigkeit. Z. Abst. Vererbl., 8: 325-332. 1913. Hybrides entre ae d’Antirrhinum. IV. Conf. Inter. Genetique, pp. 4 Miles, F. C. 1915. A Genetic and Cytological Study of Certain Types of Albin- ism in Maize. Jour. Genetics, 4: 193-21 Morgan, T. H. 1914. Two Sex-Linked Lethal Factors in Drosophila and Their In- fluence on the Sex Ratio. Journ. Exper. Zoology, 17: 81-122. 1915a. The Rôle of the Environment in the Realization of a Sex- inked Mendelian Character in Drosophila. AMER. NAT., 49: 385-429, 1915. 1915b. The Constitution of the Hereditary Material. Proc. Amer. il. 3-153 Morgan, T. H., Sturtevant, A. H., Mallet, HJ. and Briages OB: 1915. The Mican of Mendelian Herviity: Pearl, R. 1915. Modes of Research in Genetics. Setchell, W.A. 1912. Studies in Nicotiana. I. Univ. Calif. Publ. Botany, 5: 1-86. Sturtevant, A. H. 1915. No Crossing-over in the Female of the Silkworm Moth. AMER. Nar. : 42-44. de Vries, H. 1913. Gruppenweise Artbildung, p. 215. Wilson, E. B, 1914. The Bearing of nage se Research on Heredity. Proc. Roy. Soc. London, B88: 333-352. SHORTER ARTICLES AND DISCUSSION PIEBALD RATS AND MULTIPLE FACTORS In the Naturauist for December, 1916, MacDowell has pub- lished an extended criticism of experiments in the modification of the hooded pattern of piebald rats by selection, in which my colleagues and I have been engaged for some years. This is not the first time that readers of the Narurauist have had their at- tention called to these experiments by similar adverse criticism and they are possibly quite weary of the subject. In so far as MacDowell merely offers in new form arguments which have already been presented by Muller and Pearl and answered by me, I shall make no reply. But as regards two points which may fairly be considered critical, one of which actually is so designated by MacDowell, I desire to present some evidence which I regard as conclusive but which MacDowell has not dis- cussed, evidence possibly not accessible to many readers of this journal. MacDowell’s criticism is based on the data presented in Publication 195 of the Carnegie Institution (Castle and Phil- lips, 1914), and in a brief paper in The Scientific Monthly (1916). Many additional data are given in Publication 241 of the Carnegie Institution, but this is not considered by Mac- Dowell, although it was issued in September, 1916, as Paper No. 26 of the station with which he is connected, nearly two months in advance of his own paper. Had he considered carefully the evidence contained in this later publication, I am sure that he would have modified his criticism materially. In 1914 Phillips and I offered two alternative explanations of the progressive changes observed under selection in the hooded pattern. These were (a) variability of the unit-character (‘‘fac- tor”) hooded, and (b) multiple modifying factors affecting the hooded character. We found it difficult to decide between these two interpretations on the basis of evidence then avail- able. For this hesitancy we were promptly taken to task by Muller, who championed the multiple factor interpretation now adopted also by MacDowell. MacDowell elaborates in much de- tail a dozen points which show compatibility between our ob- servations and a multiple factor hypothesis, but without consid- 2 No. 602] SHORTER ARTICLES AND DISCUSSIONS 103 ering whether they are also compatible with the alternative hypothesis of a single varying factor. Modification on crossing, decreasing variability, regression, greater variability in F, than in F,, effective return selection—these are all phenomena to be expected equally on either hypothesis. To cite them is no argu- ment for one hypothesis rather than the other. This point has wholly escaped both Muller and MacDowell, who seem quite un- able to conceive any but the single explanatory principle of mul- tiple factors. Putting aside these irrelevant arguments, there remain. two points in MacDowell’s discussion which require further consid- eration. They are the same two points which led us in 1914 to hesitate between the alternative interpretations which we con- sidered, but on which we now have fuller evidence. But before we go into this new evidence one or two minor points may be noted in which the accuracy of our generalizations is questioned. MacDowell has gone over our 1914 publication in great detail, devoting as many pages to its destructive criti- cism as we to its original exposition, and recalculating the sta- tistical coefficients. It is gratifying to know that he has de- tected in these no serious errors, though his figures differ from ours slightly in some cases. Whether his calculations are more accurate than our twice-checked ones, I am unable to say with- out detailed reexamination of the data. As these are public property, the curious reader may satisfy himself on the point. I do not consider it necessary to go into the matter anew since the substantial correctness of our findings is not challenged. MacDowell thinks that we did not sufficiently emphasize the decreasing variability (standard deviation) and the decreasing rate of divergence of the selected races, observed as the selec- tion progressed. These to his mind imply that selection would sooner or later cease to be effective. In this opinion I can not concur, since in neither the plus nor the minus selection series has the standard deviation decreased by half, although sixteen Successive selections had been made and the hooded character had been so modified as to be scarcely recognizable longer. Whether one considers the decrease in variability large or small depends principally upon how much importance he attaches to the values found for the first two generations of the experiment, when the numbers of individuals observed were still small and methods of grading them had not yet been fully standardized. Mac- 104 THE AMERICAN NATURALIST [Vor. LI Dowell emphasizes the high variability of these early generations, - few in individuals, and attaches importance to the relatively smaller variability of later generations. It seems to me fairer to compare the first half of the series with the second half. Con- cerning the point I have said (Publication 241, p. 172) : The amount of the variation as measured by the standard deviation is less in the last half of the experiment than in the first half. It is also steadier, owing in part doubtless to the fact that the numbers are larger, and in part to a more stable genetic character of the selected races. But the genetic variability is plainly still large enough to per- mit further racial modification and there is no indication that it will cease until the hooded character has been completely selected out of existence, producing at one extreme of the series all-black rats, and at the other end of the series black-eyed white rats. It should be noted that these conditions have already been approximated in individual cases. THE New EVIDENCE 1. The progeny of plus selected crossed with wild rats. (Quoted without change from Publication 241, pp. 163-168.) In 1914 Castle and Phillips published a report on breeding experi- ments with hooded rats, in which it was shown that the hooded color- pattern—itself a Mendelian recessive character in crosses with the entirely colored (or “self ”) coat of wild rats—is subject to quantitative variation, and that different quantitative conditions of the hooded pattern are heritable. (Compare fig. 36, plate 7.) It was also shown that by repeated selection of the more extreme variations in the hooded pattern (either plus or minus) it is possible gradually to modify the racial mean, mode and range as regards these fluctuations, without eliminating further fluctuation or greatly reducing its amount. We concluded that the unit-character, hooded color-pattern, is a quanti- tatively varying one, but were at that time unable to decide whether the observed variability was due simply and exclusively to variation in a single Mendelian unit-factor or partly to independent and subsidiary modifying Mendelian factors. Since publication of the above I have been engaged in further experi- ments designed to show which of the alternative explanations is the correct one, and these are now sufficiently advanced to indicate definite conclusions. Previous experiments had shown that when a race of hooded rats, whose character has been modified by selection (either plus or minus), is erossed with wild rats, the extracted hooded animals obtained in F, as recessives show regression toward the mean condition No. 602] SHORTER ARTICLES AND DISCUSSIONS 105 of the recessive race before selection began. This result suggested that the regression observed might be due to removal by the cross of modifying factors, which selection had accumulated in the hooded race. If this view was correct, it was thought that further crossing of the extracted hooded animals with the same wild race should result in further regression, and that if this further regression was not observed a different explanation must be sought for the regression already noted. The entire experiment has accordingly been repeated from the be- with the same result as regards regression in the first F, gen- eration, but with no regression of the same sort in a second F, contain- ing twice-extracted hooded animals. So far from observing further regression as a result of the second cross with wild rats, we have unmis- takable evidence that the movement of the mean, mode and range of the hooded character has been in the reverse direction. So the hypothe- sis of modifying factors to account for the regression and for the pro- gressive changes observed under selection becomes untenable. In repeating the experiment of crossing hooded rats of our selected races with wild rats, great care has been taken to employ as parents individuals of the greatest racial purity and to inbreed the offspring brother with sister, thus precluding the possibility of introducing modifying factors from other sources. In making the second set of crosses, the extracted individual has, wherever possible, been crossed with its own wild grandparent. In the few cases in which this was impossible, wild animals of the same stock have been used. This stock consisted of a colony of wild rats which invaded the basement of the Bussey Institution apparently from a near-by stable. Owing to faulty construction of the building they were able to breed in spots inaccessible to us, and it took many months of continuous and persistent trapping to secure their extermination. During this period we trap a hun- dred or more of them, all typical Norway rats, colored all over, without even the white spot occasionally seen on the chest of wild rats. Two generations of rats from this wild stock have been reared in the labora- tory, and all have this same self-colored condition. The hooded animals used in the experiments to be reported on in this connection consisted of 4 individuals of the plus selected series, a male and 3 females, as follows: TABLE 140 Individuai Grade.! Generation. | | 9 Spig... +4} 10 48.. +4 10 600. . + 4} 12 | | 969056... ; +4 12 | 1 See figure 35, plate 7 for significance of the grades. 106 THE AMERICAN NATURALIST [ Vou. LI Each of these animals was mated with a single wild mate, and their children were weaned directly into breeding cages containing a male and two or three females (brother and sisters). In the case of two matings, F, males of the same parentage were at the time lacking and males froma different cross were used. The results of such matings are tabulated by themselves and serve a useful purpose as controls. The F, animals all closely resembled their wild parents, but many of them had a white spot on the chest. They ranged from grade +5% to +6 (self). The F, animals are classified in table 141, where it appears that 73 of them were hooded and 219 non-hooded (i. e., like F,), an exact 1:3 ratio. More than half of this F, generation consists of the grand- TABLE 141 Table 141 shows the classification of extracted hooded first F, young ob- tained from crossing hooded rats of the plus-selected series with wild rats. Grade of hooded grandchildren. Total | Total | Means - Hooded grandparents. hooded. |_ nOn- of 14| 12| 2 | 2}| 24| 22) 3 | 34| 34| 3ł| 4 hooded. | hooded. 9 5513, + 44, gen. 10....| 1/..| 3) 2} 1| 7| 8| 6! 5| 7| 1 41 107 | 3.05 d 6348, + 4, gen. 10..... .-|--| 1}..| 1} 2] 4) 3] 4) 6 I) 22 68 | 3.28 Q 6955, + 4, gen. 12..... Ros iar Rae rer E se e Ys |e Po ee 27 | 3.51 9 5513, + 44, an 6600; 48, gen. 13.056 b ech stots ead Bed des 3 D Sn Q 5513, + 44, and 9 6955, + 4, gen. 12..:.. Pas Pree Fans Pay By bere aes Wes T 2 5 | 3.37 Teele eS 1)..| 4| 2| 2} oj14l11|12lt6| f 73 | 219 | 3.17 children of 95513, produced by breeding her children brother with sister, those children all having been sired by the same wild rat. Her grandchildren include 41 hooded and 107 non-hooded young. The hooded young range in grade from +114 to +-4, their mean grade ing + 3.05, a considerable on from the grade of the grand- mother, which was 4.25. Hooded rats of the same grade and generation as the grandmother, when bred with each other, produced young of mean grade + 3.84. (See table 10, Castle and Phillips.2) The mean of the extracted hooded grandchildren in this case (being 3.05) shows a regression of 0.79 from that expected for the uncrossed hooded race. From the extracted hooded grandchildren of 95513, produced as just described by a cross with a wild male, 7 individuals, 2 males and 5 females, were selected for a second cross with the wild race. They ranged in grade 2 Comparison should have been made with generation 11 offspring, whose mean was 3.91 not 3.84. This would make the regression 0.86, instead of 0.79. No. 602] SHORTER ARTICLES AND DISCUSSIONS 107 TABLE 142 Table 142 shows the classification of extracted hooded second F, young obtained from crossing first F, hooded rats (table 141) with wild rats. The hooded grandparents were themselves grandchildren of 95513, + 44, gen- eration 10, on the side of both parents. Grade of hooded grandchildren. | Total | Total | Means Hooded grandparents. | : i | hooded. |_ non- | of 23 | 2} | 2} | 3 | 34} | 3} | 34| 4 | | . | hooded. i | | } | | “eS CA Se AR ae As Sp a, a) at Shah oe 8 | 3.37 g 9686, + 23 ir Peet oD | 28 | 3.40 F 9620, + 23....... H ieS ai arna aake 24 | 3.06 9 9729, + 23 vob bi) Oo 21-2110 22 62 ENM acin bist O1 OT Mtl) 71 at OO 1) Oe) Oa 2 Ore, E eee BC: ea ae Se ee 68 | 3.55 9 9621, +34... nia ny Ba Tae 1... tai B12: 18 42 | 3.70 Tome. oi yi" 3) 3 4 | 6|13|28|30|11| 98 | 296 | 3.47 from +2 to +314. (See table 142.) They produced several litters of young of the same character as the first F, young, all being similar to wild rats in appearance, except for the frequent occurrence of a white spot on the belly. These second F, young were at weaning time mated, brother with sister, in breeding-pens, precisely as had been done with the first F,’s. They produced 394 second F, young, of which 98 were hooded and 296 non-hooded, a perfect 1:3 ratio. The hooded young varied in grade from + 2 to +4, as shown in table 142, the data there being given for each family separately as well as for all combined in the totals. One family was very like another as regards the character of the hooded young, except that the higher-grade grandparents had grandchildren of slightly higher grade. Thus the average of all the 98 hooded young was + 3.47, but the average of those descended from the 3 grandparents of lowest grade was less than this, while the average of those descended from the 3 grandparents of highest grade was greater. This is just what had been observed throughout the entire selection experiments. (See Castle and Phillips.) If we weight each of the grandparents in table 142 in proportion to the number of its hooded grandchildren, then the mean grade of all the grandparents is + 2.95. Since the mean grade of all the 41 first F, hooded grandchildren, from which these 7 were chosen, was -+ 3.05, it will be seen that these 7 are, so far as grade is concerned, fair repre- sentatives of the 41, being in fact of slightly lower mean grade. It is therefore all the more striking that their grandchildren, the second F, hooded young (table 142), are of higher grade. They regress in an Opposite direction to that taken by the first F, hooded young. Thus the original hooded ancestor (25513) was of grade 4.25, The grade of hooded young expected from such animals is 3.84. What she produced 108 THE AMERICAN NATURALIST [ Vou. LI in F, following a cross with the wild male, was young of mean grade 3.05. Seven of these of mean grade 2.95 produced a second F, contain- ing hooded young of mean grade 3.47. This is a reversed regression of 0.52 on the grade of their actual hooded grandparents, or of 0.42 on the group from which their grandparents were chosen. Their mean lies about midway? between that which would have been expected from the original hooded female (5513) had no crossing with wild rats oc- curred and that which was observed in the first Obviously these facts do not harmonize with the assumption that the regression observed in the first F, was due to loss of modifying fac- tors accumulated during-the ten preceding generations of selection; for no further loss occurs in the second F, On the other hand, a par- tial recovery is made of what was lost in the first F, This suggests the idea that that loss may have been due to physiological causes non- genetic in character, such as produce increased size in racial crosses; for among guinea-pigs (as among certain plants) it has been found that F, has an increased size due to vigor produced by crossing and not due to heredity at all. This increased size persists partially in F,, but for the most part is not in evidence beyond F,. I would not suggest that the present case is parallel with this, but it seems quite possible that similar non-genetie agencies are concerned in the striking regression of the first F, and the subsequent reversed regression in the second F,. Whatever its correct explanation may be, the fact of the reversed re- gression in a second F, is very clear, as other cases than those already discussed will show. A hooded rat of grade + 4 and generation 10, $6348, had by a wild female several young of the character already described for the young of 95513. These, mated brother with sister, produced a first F, (table 141) of 90 a 22 of which were hooded, 68 being non-hooded, again a good 1:3 ratio. The hooded young ranged from + 2 to + 4 in grade, their mean being 3.28. Of the 22 hooded individuals, 1 male and 7 females were mated with wild rats to obtain a second F,, and the second F, animals were then mated brother with sister to obtain the esired second F, The character of this is shown family by family in table 143. It contained 497 individuals, of which 121 were hooded and 376 non-hooded, a ratio of 1:3.1. The weighted mean of the 8 selected grandparents is 2.93, which is 0.35 below the mean of the 22 first F, hooded animals which they represent. The mean of the second F, hooded young is 3.22, which indicates a reversed regression of 0.29 3In The Scientific Monthly (Jan. 1916) I have stated that a second cross showed ‘‘a return to about what the selected race would have been had no crossing at all occurred.’’ This is obviously inaccurate and should be corrected. It rests on a comparison with the combined average of both the older and the more recent experiments. [MacDowell devotes half a page to demolishing the statement already corrected here.] No.602] SHORTER ARTICLES AND DISCUSSIONS 109 on the grade of the grandparents, but shows no significant difference from the mean of the grandparental group (3.28). TABLE 143 Table 143 shows the classification of extracted hooded second F, young “ obtained from crossing first F, hooded rats (table 141) with wild rats. The hooded grandparents were themselves grandchildren of ¢ 6348, + 4, gen- eration 10, on the side of both parents. Grade of hooded grandchildren. “saan Total | Means o Hooded d ts. non] of ae 12] 2 | 23 | 23 | 22 | 3 | 34] 33 | 33] 4 hooded. | hooded. | hooded g 9639, + 2..... BS Lie St 6) ari G6 ll oe 400 e Wie tE a ahal Linh Oh eal. 6 16 | 3.17 Q 9765, +3..... Sheds Yea ee Fic ake 1 10 | 3.50 OT aly OF OES el BA Ped Dok a ae A 7A 76 | 2.90 WW, PSG lt Pee 2h Sel: Bt 2 2-16 47 | 3.28 Q 9705, + 33 SED ice t ee el ee 74 | 3.48 F 9748, + 34 VIA WG Bi 9 40 | 3.36 © oe; 4 aikoo des Aecuteoseaehus at ed 3 32} BaF Totale; onua. 2|10| 2| 2| 8|23| 8|35|24| 7| 121 | 376 | 3.22 All except one of the 8 families classified in table 143 show unmis- takably the reversed regression. This exceptional family consists of the grandchildren of 99747. They have a mean grade of 2.90, sub- stantially the same as that of the entire group of grandparents, but con- siderably lower than that of their own hooded grandmother. Appar- ently she did not come up genetically to her oe e grade. This the other grandparents of the group did. For those of lowest grade (2, 234) produced lower-grade hooded pater than did the grandparents of highest grade (314, 4), as was found to be the case also in table 142 We may next trace the inheritance of the hooded character through a third but smaller family produced by two successive crosses with wild rats, the hooded character in this case being derived from ? 6995, grade + 4, generation 12. The character of her first F, descendants is shown in table 141. They consist of 5 hooded and 27 non-hooded individuals. The mean grade of the hooded young is 3.51, but the number of these young is too small to make this mean of much signifi- cance. One of the hooded young (49660, + 334) was mated with a wild female to secure a second F, generation and from this in due course was produced the second F, generation (table 144). It consisted of 21 hooded and 44 non-hooded young. The hooded young showed the usual ge (2 to 4). Their mean grade was 3.50, substantially identical with that of the first F, animals, but 0.25 below that of the actual hooded grandparent. This family history is less satisfactory than the two already diseussed because of the smaller numbers which it includes. It 110 THE AMERICAN NATURALIST [Vou. LI contains nothing contradictory to the sabi ame already given, though reversed regression is not in this case in evidence. TABLE 144 Table 144 shows the classification of extracted hooded second F, young obtained from crossing first F, hooded rats with wild rats. The hooded grandparent, $9660, + 33, was a grandson of 96955, + 4, generation 12, on the side of both parents. The hooded grandparent 419711, + 34, was a grandson, on the side of one parent, of 95513, + 44, generation 10, and on the side of the other parent, of 9 6955, + 4, generation 12. (See table 141.) Grade of hooded grandchildren. Total | Total | Means Hooded grandparents. | naoded | non- of 2 | 2}|24]2ł}| 3 |34/3}|3f] 4 `| hooded.| hooded. T O Vine. TERE OL Ce eee Gee See ge eke Weeder eae oii, Far... rete Mt el 4 Sh Abe ie bere 33 3.28 Tou i 5s diss a tiet 1+ S 94 6} OPilt Ot o 97 27] AO In two eases F, females could not be mated with brothers and so mates were taken from other families. Thus “mixed F, matings” were made between children of 5513 and 6600 and children of 5513 and 6955. (See table 141.) The former mating produced 3 hooded and 12 non-hooded “ first” F, young; the latter produced 2 hooded and 5 non- hooded “ first” F, young. The grade of the hooded young produced by these mixed matings was not different from that of brother-sister matings, so far as the small numbers permit one to judge. One of these mixed matings was carried into a second F, generation. The first F, hooded & 9711, + 31⁄4, was mated with a wild female, and the young were bred, brother with sister, producing 16 hooded and 33 non- hooded young. (See table 144.) The mean grade of the 16 hooded young was 3.28, substantially the same as that of the first F, hooded grandparent. No additional regression through loss of modifiers (or other agency) is here in evidence. The result is the same as that ob- attention has been given to individual pedigrees, is particularly di- rected to the foregoing case. Having now discussed each family history separately, we may com- bine all the second F, families in one table, in order to get a clearer impression of the results as a whole. (See table 145.) The second F, generation thus combined includes 256 hooded and 749 non-hooded individuals, a ratio of 1: 2.9, an unmistakable mono-hybrid Mendelian ratio. The mean grade of the hooded individuals is 3.34. The weighted mean grade of their hooded grandparents was 3.02, which indicates a reversed regression of 0.32 for the entire second F, group of hooded animals. No. 602] SHORTER ARTICLES AND DISCUSSIONS 111 TABLE 145 Table 145 is a combination of tables 142 to 144, in which the second F, young are classified according to the grade of their first F, hooded grand- parent. i Gaat sehen | Grade of hooded grandchildren pukka Total Means .toresenaoaeese (18 | 2 | 2 | 24 | 28| 3 | 3t| 38 st] 4 meg TE | nied 2 apep a}...1 akej 5}16| 6| ail 41 | 118 | 3.25 23 ee R ee We” a Aan Sa ets ADE eB 1 34 90 | 3.29 3 FERA E dd Od kame et eae LO PD 53 182 | 3.48 34 POT ee ad BIOS IO AS ee 59 151 | 3.22 34 coner Dip def Ode SEES ad 46 161 | 3.39 3ł TR EE A bin kee oO oe E 21 44 | 3.50 4 ae ole cata cer web ed i 2 3 3.87 | 3.02 2 | 12| 4) 6|15|32|27|72|65)21} 256 749 | 3.34 Classified according to the grade of the (first F,) grandparent, they show a correlation between grade of grandparent and grade of grand- child. The lower-grade grandparent has lower-grade hooded grand- children, and the higher-grade grandparent has higher-grade hooded grandchildren. This shows that the variation in grade is (in part at least) genotypic. As the experiment yields no evidence that the varia- tion in the hooded character is due to independent modifying factors, there remains no alternative to the coneltision that the single genetic Mendelian factor concerned fluctuates in genetic value. Fluctuation accordingly is not exclusively phenotypic, as DeVries and Johannsen have thought, but may be genetic also. Hence racial changes may be effected through selection by the isolation of genetic fluctuations, as well as by the isolation of mutations. Moreover, genetic fluctuation makes tion attaining results which it would be quite hopeless to seek by any other means. . The progeny of ‘‘mutant’’ crossed with wild rats. (From Piast. 241, p. 173.) Castle and Phillips described, under the name of “ mutants,” 2 rats of the plus-selection series of extremely high grade. They proved to be eterozygotes between the average condition of the plus-selected race at that time, about + 3.75, and a new condition, not previously known in our hooded races, but resembling that seen in “ Irish” rats, which are black all over except for a white spot on the belly and would be classed on our grading scale as about +514. In later generations we secured animals homozygous for the darker condition just deseribed (that of Trish rats). The homozygous “ mutant” race proved to be very stable in color pattern, varying only from 514 to 534, with a majority of ani- mals graded 514. Attempts to alter the modal condition of the race 112 THE AMERICAN NATURALIST [Vor. LI by selection have thus far proved futile because of our inability to in- crease the race sufficiently to afford a basis for selection. Its inbred- ness and its feebleness are perhaps causally related. The suggestion was made that the change from our plus-selected race, which had occurred in the mutant stock, might be due to some supplementary modifying factor, not to a change in the hooded factor itself.* If so, a cross with a race lacking the hooded factor or its “ modifiers ” might serve to demonstrate their distinctness by separating one from the other. A wild race seemed best suited for a test of this hypothesis, since it would be free from suspicion on the possible ground of harboring either the hooded pattern or its supposed modifier, which had converted the hooded pattern into the mutant. It was to be ex- pected, if the hypothesis were correct, that the mutant character was hooded plus modifier; that then a cross with wild should produce in F, hooded young (lacking the modifier) as well as mutants and selfs. But if the mutant race had arisen through a change in the hooded factor itself, then the cross should produce only mutants and selfs, without hooded young in F, Crosses have now been made on a sufficient scale to show beyond question the correctness of the latter alternative, which is entirely in harmony also with the results described in the preceding parts of this paper Six homozygous “ mutant” females of grade + 51⁄2 were mated with wild males of the same race described in Part I. They produced 46 young, all gray like wild rats and of grades as follows: Exactly half of the 46 F, rats bore no white spot, i. e., were of grade Seven more bore only a few white hairs (grade 57%). The re- mainder were very similar to the mutant parent in grade. Several matings were made of the F, rats, brother with sister, which produced 212 F, young. About a quarter of these were black (non- agouti), the rest being gray (agouti). Both sorts included about equal numbers of individuals with and without white spots. No difference was observed in this respect between the progeny of spotted and of unspotted parents. Table 158 shows the F, young grouped family by family according to grade. Three of the four families are descended from a single mutant grandparent; the fourth family is descended from two different mutant grandparents which were bred simultane- 4 This also is MacDowell’s view. He says, p. 734: ‘‘The newly discovered factor acts independently of the other factors, is not modified by them, and does not modify them. Being the one difference between the mutant and the plus race at the time the mutant appeared, this factor affords a crucial test for the interpretation of the modifications that result from crosses.’’ No. 602] SHORTER ARTICLES AND DISCUSSIONS 113 ously to the same wild male in the same cage. The 10 F, young of this family may have been produced either by full brother and sister, or by half-brother and half-sister; it is uncertain which, All other F, young were produced by brother-sister matings. TABLE 158 Table 158 shows the classification of the F, young obtained by crossing homozygous ‘‘mutant’’ with wild rats. - Grade of offspring. Mutant grandparents. Sad a Totals. 5 | 64 | 53 | 53] 53] 6 : 2 0630, De ie oe ot | 81a2t 91 2} 2 46 ee E OE ee te ae 1} 2/22/29| r] 59 114 A A O E ed 333) 481 Bet 9 42 9 0630, + 54, or 0636, + 53............ 11 Scab es 10 OE ee ee ee 1| 6|49/50| 3 103] 212 It will be observed that the F, young (table 158) which are white- spotted are in no case hooded. Their range of variation does not fall beyond that of the uncrossed mutant race. It is certain, therefore, that the “mutant” condition is not hooded plus an independent Mende- lian modifier. It is a changed form of white-spotting, alternative to the form of spotting found in the race from which it was derived (the plus- Selection series, generation 10). It is, without much doubt, also alter- native to the self condition of wild rats, though fluctuation in grade ob- Secures the segregation, which may, very likely, be imperfect. This serves to confirm the general conclusion that throughout the entire series of experiments with the hooded pattern of rats we are dealing with quantitative variations in one and the same genetic factor. CONCLUSION I have now presented the évidence which has led me to reject the hypothesis formerly held tentatively that modifying factors were largely concerned in changes produced in the hooded pat- tern of rats under repeated selection. This evidence seems to me to admit of only one consistent interpretation, that a single variable genetic factor was concerned in the original hooded race, that a changed condition of this same factor was produced in the minus race, and another changed condition in the plus race, and a third appeared in the mutant race. All are allelo- morphs of each other, and of the non-hooded or self condition found in wild rats, yet all tend to modify each other in crosses. The character has a high degree of genetic stability, yet is sub- 114 THE AMERICAN NATURALIST [ Vou. LI ject to continuous genetic fluctuation. I have been unable to produce or to discover any race of spotted animals which is free from genetic fluctuation, though I have made an extended search. If MacDowell or any one else has discovered such a race, let it be produced. It does not, of course, follow that because white spotting in rats is capable of indefinite modification through selection, there- fore all heritable characters are equally capable of modification. Physiological limitations no doubt often limit the modifiability of characters. A sugar-beet can not be produced which is all sugar or much over 25 per cent. sugar. There has to be retained a plant mechanism for the production of sugar, a beet. Neither is it to be expected that the thorax of Drosophila can be deco- rated with an indefinite number of extra bristles. The bristles have to be attached to something, and the thorax of Drosophila is finite in size. It is not necessary to suppose that hypothetical modifying factors have been used up simply because variation ceases to progress in a particular direction. For no one, I sup- pose, would contend that variation is equally easy in all direc- tions and in all characters. De Vries has taught us the signifi- cance of one-sided variation and we have become familiar with recurrent types of variation which are encountered first in one species and then in another. Such cases show that different kinds of germplasm are similar in structure and likely to under- go similar changes. But what happens to these spontaneous variations ‘when once they have put in an appearance depends on external agencies, man or other factors in the struggle for existence. The modern study of evolution has indeed empha- sized the importance of spontaneous internal changes in pro- ducing variations, but we still have to reckon with selection, nat- ural and artificial, in determining the survival of variations as well as in controlling their magnitude and the direction of their further variation. W. E. CASTLE BUSSEY INSTITUTION, T HILLs, MASS., December 20, 1916 No. 602] SHORTER ARTICLES AND DISCUSSIONS 115 FIELD TESTS OF THEORIES CONCERNING DISTRIBU- TIONAL CONTROL? THE conditions of animal distribution and the causes of these conditions are facts which concern intimately the problems of the persistence and of the evolution of species. The present writer believes that the field naturalist is in a position to con- tribute in large measure toward the solution of these problems, and it is the purpose of this paper to show how comparative studies in the distribution of species may throw light not only upon the nature of the environmental complex, but also on the relative importance of its various component factors. Some simple facts of distribution which are of common obser- vation, and which were early recorded by the systematic zoolo- gist, are: (1) that each animal occupies a definite area, that 1s, has a habitat or range, which is distinctive enough to be included among the characters of the species and described along with its habits and the features of its bodily structure; (2) that some Species (and even some of the higher systematic groups) range widely, and cover great extents of country, while others are ex- tremely local or restricted in their distribution; and (3) that, notwithstanding considerable variation in this degree of distri- butional restriction, many species (or higher groups) are found nearly or entirely to coincide in range, so that sets of species, of varying ranks, may be recognized distributionally, as constitut- ing realms, zones, faunas, subfaunas, associations, etc. Perhaps the most prominent delimiting factor, and the one which has been emphasized through repetition in the early sys- tematic writings, is the obvious one of physical barriers—repre- sented by bodies of water in the case of the terrestrial species and by land in that of the aquatic. The majority of animals inhabiting islands and seas are specialized in such a manner as to be hemmed in by the limits of their respective habitats. In- dividuals overstepping the barrier in either case are subject to prompt destruction. This obvious type of distributional control has always been and will remain an important one for consid- eration; but with the acquisition of detailed knowledge regard- ing the distribution of animals on large continental areas, naturalists have been led to propose many other factors which 1 Contribution from the Museum of Vertebrate Zoology of the University of California 116 THE AMERICAN NATURALIST [ Vou. LI have seemed to them to prevent the random and unrestricted spreading of animals over the surface of the land. The fol- lowing is a list of the factors which various writers have nomi- nated as affecting the distribution of the higher vertebrate ani- mals. This list is complete only to the extent that my own exam- ination of the literature is so. Many of the items have been found in dissertations upon bird migration, which is, of course, but one phase of the general subject of distribution. Vegetation. Food supply, kind and quantity. Rainfall. Humidity of the air (relative or absolute). Wetness or dryness of the soi Barometric pressure, or altitude: Atmospheric density. Safety of breeding places. Availability of temporary refuges. Water (to land species). Land (to aquatic species). Nature and availability of cover, or shelter from enemies. Nature of the ground (coarse or fine soil, or rock). Insolation, or light intensity. -Cloudiness. Temperature: in general; mean annual; of winter; of period of reproduction; of hottest part of year. Interspecific pressure, or competition, or race antagonism. Parasitism individual, or racial, preferences. It is at once plain that some of the items enumerated are ex- tremely complex, and that the most superficial analysis will | show some duplication among them. For example, the factor of vegetation as influencing the distribution of different mammals resolves itself principally into the elements of food-supply and shelter, and, subordinately in most cases, into those of tempera- ture, humidity, and nature of the soil. As some of the suggested factors may really never function in any vital degree as sup- posed, the total number of really critical factors is probably smaller than the total of the items just listed. Time could not here be taken to discuss the intrinsic nature of each elemental factor, even if the writer were equipped to handle such a variety No. 602] SHORTER ARTICLES AND DISCUSSIONS il? of subjects; for such a discussion would in most instances lead directly into physics and chemistry, and into a study of the physiological processes of the animals affected by each of these factors. I should, however, like to dispose at once of one of the ‘*factors’’ listed, and which I hear and see repeatedly cited as. a cause of restriction in distribution—particularly in that of birds. Many people claim to see in the facts of distribution only the operation of a preference on the part of each animal—by virtue of which, if a heterogeneous lot of animals were introduced into an area presenting diverse conditions, each species would choose its ‘‘natural’’ surroundings and rapidly allocate itself in a nor- mal way. I grant that such a choice would almost certainly be made. In fact the hypothesis is being proved continually all over the country in connection with the migration of birds. Seores of species travel north in the spring to countries for a preceding interval unoccupied; and while, roughly speaking, they travel together, and arrive together, they segregate them- selves, immediately on their arrival, and repair to separate sorts of ground, each species by itself: the pipits to the prairie, the water-thrushes to the streamside thicket, the black-poll war- blers to the spruce forest, and so on. We have here an obvious choice exercised in the selection of habitats. But does this segregation of species by exercise of ‘‘individual preference’’ in a uniform direction change the nature of the problem in any fundamental way? Should we not here recognize merely a char- acter in the cerebral equipment of each race, which, like every external peculiarity in its structure, is in considerable measure the result of protracted impress upon the organism from the environmental complex of factors to which the race has been subject through past time? There is no other additional factor than those environmental ones (plus the intrinsic fixedness of the species, within certain limits of plasticity, and the ‘‘evolu- tionary momentum’’) to be called into account. As to the mechanism of geographic limitation, the adjustments to the various critical factors are inevitably forever in process, though reduced to a minimum at times of slow environmental change. The refined method of individual ‘‘preference’’ or ‘choice’? is superior to the wasteful process of wholesale de- struction which would be experienced by individuals finding themselves out of place as the result of a haphazard selection of 118 THE AMERICAN NATURALIST [ Vou. LI locality. The frontier individuals, those on the margin of the habitat of the species, may not prosper as greatly, or reproduce as prolifically, as those in the metropolis of their species ; but they certainly do not, as a rule, beat themselves to death individually against their limiting barrier, of whatever nature it may be. To resume the main topic of this discussion, I shall attempt to show that it is possible from field observation to indicate in the case of certain species, some, at least, of the factors which con- trol their distribution ; and further that we who live in Califor- nia have splendid opportunities to gather and examine data by means of which the general laws of animal distribution can be determined. An area within comparatively easy reach presents a wide diversity in topographic and climatic features. Occupy- ing this area is an abundant complement of the higher vertebrate classes. Within the political limits of the state, systematists now recognize the presence of 388 species of mammals, 543 of birds, 79 of reptiles and 37 of amphibians. We have plenty of ma- terial to work with. I shall proceed to discuss a few selected species about which we seem to have knowledge enough to war- rant provisional inferences. THE CASE OF THE OREGON JAY The Oregon jay (Perisoreus obscurus), a close relative of the Canada jay, or whisky-jack, occurs in California only in the northern third of the state. Even there it is very local in its oc- currence and absolutely non-migratory. On the Warner Moun- tains, Modoe County, it ranges from the highest parts down to > 7,000 feet altitude. On Mount Shasta it ranges from near tim- berline down to about 6,000 feet altitude. It is absent for a long distance to the west, through the Trinity mountain mass, but it recurs along the seacoast of Humboldt County, within fifteen - miles of the ocean. And here is the curious point: along this coast strip it does not range higher than 300 or 400 feet above sea level, although there are mountains not far inland which rise to an altitude of several thousand feet. Let us look into this ease for the purpose of determining the factors responsible for this interrupted range. , The Oregon jay, like most members of the crow family, is not restricted in diet. It eats a great variety of both vegetable and animal substances; its food varies in character according to season and local conditions. -The supply of any particular kind No. 602] SHORTER ARTICLES AND DISCUSSIONS 119 of food is not likely, therefore, to be a controlling factor in its distribution. he bird is a forest dweller. Its equipment as regards man- ner of flight and course to take in case of attack by enemies is adjusted to a forest habitat, and nowhere within the writer’s knowledge does this jay extend its range beyond the limits of woods of some sort. Although somewhat predaceous itself, it has regular enemies among hawks and owls, for protection from which it makes use of forest vegetation. This factor of forest cover, then, must be counted as essential. But the range of the bird is not continuous wherever forests extend. In the interior of California it does not descend below a cer- tain altitude. Now three other factors in its distribution are quite obviously connected with that of altitude, namely, baro- metric pressure, atmospheric density, and temperature. But when we take into account the fact that the Oregon jay exists at or close to sea level around Humboldt Bay, the first two fac- tors, those of pressure, and air density, are instantly eliminated, because of the obvious fact that the bird successfully maintains itself in localities of widely differing altitude where these factors are thus extremely diverse. With reference to temperature, we know without recourse to instrumentation that there is a decrease upwards at an average rate of 3 to 4 degrees F. per thousand feet. If, then, the bird is limited downwards at a critical point, the inference appar- ently follows that temperature is the determining factor, and this conelusion is inevitable if we consider only Mount Shasta and the Warner Mountains. But the bird’s oceurrence at Hum- boldt Bay complicates the problem. In order to reconcile these facts of distribution we must look into the situation with refer- ence to season. On doing so we discover that the home of the Shasta and Warner jays is subject to severe winters with heavy snow, very much colder than the winters at Humboldt Bay, where the climate is equable and snow rarely falls. But the summer temperature at Humboldt Bay is well known to be much cooler than that of even somewhat higher regions in the interior, up to an altitude of at least 4,000 or 5,000 feet, because of the eastward moving air-currents, which are coolest where they first leave the sea surface and warm up as they pass farther and far- ther inland. We are therefore led directly to the final inference that the summer temperature at sea level about Humboldt Bay 120 THE AMERICAN NATURALIST [ Von. LI closely approximates the summer temperature at from 6,000 to 9,000 feet on Mount Shasta and above 7,000 feet on the Warner Mountains. In these three areas, the air is cooler in summer than in the interlying areas and thus better adapted to the finely adjusted requirements of the Oregon jay. Summer tempera- ture, between certain degrees, is one critical factor. Three more factors present themselves for consideration in connection with the Oregon jay, those of humidity, rainfall and cloudiness. Humboldt Bay lies in the most humid and continu- ously rainy section of California. Mount Shasta and the War- ner Mountains are relatively arid, the latter most notably so. It would appear, therefore, that humidity, rainfall and cloudi- ness had little or nothing to do with cutting off the range of this bird, though one or other of these factors may have been respon- sible for the very slightly darker tone of color which distin- guishes the coast jays (subspecies Perisoreus obscurus obscurus) from those in the interior (P. o. griseus). But, however this may be, it is clear that temperature must dominate greatly over the three factors named in checking dissemination. In summary, we may therefore dispose of the following fac- tors as having little or no effect on the distribution of the Oregon jay as a species: the nature or quantity of its food supply, at- mospherie density and pressure, cloudiness, rainfall, humidity of the air or soil, and winter temperature. This eliminates all but the two factors: shelter of a sort provided by the forest habitat, and temperature of the summer season. THE CASE oF THE Cony The cony or pika is a mammal represented in California by four quite similar races (Ochotona taylori, O. schisticeps schistt- ceps, O. s. muiri, and O. s. albatus) , which agree distributionally in occupying a very restricted habitat along high mountain crests. I know of no place in central California where conies range below an altitude of about 8,000 feet, and they range upwards to fully 12,000 feet in the vicinity of Mount Lyell. They thus occupy an altitudinal belt between extremes 4,000 feet apart. With regard to zones of vegetation conies live from considerably below timberline to considerably above timberline. Exten observation shows that their existence is in no way correlated with that of trees or shrubs of any sort. Like their relatives, the rabbits, they feed entirely on low vegetation, biennials mostly ; No. 602] SHORTER ARTICLES AND DISCUSSIONS 121 but unlike most kinds of rabbits they are strictly dependent for safety from enemies upon rocks, especially where these are loosely piled as in talus slopes and so afford deep retreats within their interstices. The whole equipment of a rabbit is clearly adapted to foraging in the open, its keen hearing and eyesight quickly warning it of the approach of enemies, and giving it time to escape by means of its unusual running powers. But the cony is equipped in a very different way, as it has relatively small ears and eyes, and small hind legs. It is compelled to forage close to or beneath cover. In fact in field observations it is rarely seen on the move except momentarily, and then only between or beneath angular granite blocks, where it grazes on such little patches of vegetation as are within immediate reach. Tt is clear from numerous observations that the cony is sharply restricted in a large part of its range by the rock-pile habitat. Even at favorable altitudes it is not found away from this refuge. There are obviously, however, one or more additional factors in its distribution. In many parts of the Sierras, talus slopes occur from near the highest summits down to the foothills. As examples of these, one may cite the vast earthquake taluses of the Yosemite Valley proper, which occur almost continuously down to and below the 4,000-foot contour. These taluses have been searched diligently both by trapping and hunting, without our naturalists finding a trace of conies below 8,000 feet. The animals are easy to detect, by reason of their characteristic ery, uttered at any time during the day, though more particularly in the morning and the evening, and by the accumulations of their feces, the pellets constituting which are, in size, shape and tex- ture, unlike those of any other mammal. What is it, then, that limits the conies downward on the western flank of the Sierras, where their necessary rock habitat is continuous, and where food of the right sort is also continuous? Let us try barometric pres- sure, and atmospheric density, which may properly be consid- ered together. These conditions change sensibly with altitude and, if we take into account California alone, the facts would Seem to entitle them to serious consideration as active delimitors of the conies downward. But as we trace the range of the conies far to the northward we are led to a different conclusion. The altitudinal limits of their range is found to descend quite regu- larly towards the north, until, in the case of one race, even sea level is reached, at Bering Sea. Clearly, conies, generically, are t22 THE AMERICAN NATURALIST [Vou. LI thus proven not to be affected by atmospheric pressure, or by atmospheric density, at least in as far as it is modified by alti- tudes ùp to 12,000 feet. The same fact—depression of range towards the north—discloses a third concomitant of altitude, which is also a concomitant of latitude, namely, temperature, and this is beyond doubt the determining factor. As the iso- therms dip toward sea level to the northward so does the range of the genus Ochotona. We have, therefore, by study of geo- graphical distribution in this case established two important con- trolling factors, namely (1) safety refuges of a sort provided by talus slopes and glacial moraines; (2) temperature, at least downward below the degree, correlated in the mountains of Cali- fornia by a mean annual or summer computation or for a briefer period at the time of reproduction, with an altitude of eight to twelve thousand feet, according to latitude, slope exposure and air currents. It is not possible for one to say from the data in hand what the direct controlling factors of the upward limits of the cony’s range may be. Taluses extend up to the highest peaks, but there is no growth of grass above about the 12,000-foot contour even on ‘the most favorable slopes. As the disappearance of the cony in the higher altitudes is coincident with the disappearance of its food, it appears as if failure of food alone were the delimitor ere; but we have no way of showing that even if food did con- tinue the cony would be restricted upward, as it certainly is downward, by a change in temperature beyond some critical point. The cause of its delimitation downward, however, re- mains clear. THE CASE OF THE Rosy FINCH In the case of the bird called generically Leucosticte, or rosy finch, we find a condition astonishingly similar to that of the cony. In fact almost the entire preceding account could be måde relevant here, by merely substituting the term rosy finch for cony. The ranges, altitudinal and geographical, of the two animals are almost identical. The only obvious differences ap- pear in their ecologic relations, and consist in the lesser de- pendence of the bird upon shelter and in the dissimilar nature of its food. The rosy finch forages gregariously on the open slopes, near timberline and above, though its nest is hidden _ away in the clefts of rock ledges and taluses. It shuns the trees No. 602] SHORTER ARTICLES AND DISCUSSIONS 123 and bushes even where it ranges well below timberline. It feeds winter and summer upon seeds of dwarfed vegetation, including those of grass and herbs of various sorts. As far as I can see, its food and feeding habits are identical with those of such other fringillids as goldfinches and siskins. Yet the leucosticte, by the same tests as were used with the cony, is beyond any conten- tion limited downward by an increase of temperature. We fin the bird to possess various adaptive features in common with cer- tain arctic finches, such as tufts of bristle-like feathers over the nostrils to prevent fine snow from entering. These enable the bird to spend the long winter on the cold wind-swept ridges, but at the same time would hardly prevent the bird’s dropping to warmer climes if the heat were not a strongly deterrent factor. Cases of coincidence, as instanced by that of the cony and leucosticte, among animals of widely different powers of locomo- tion and ecologic position, are the rule, not the exception, and impel the observer to belief in the efficacy of the controlling fac- tor above mentioned. THE CAsE OF THE Repwoop CHIPMUNK The redwood chipmunk (Eutamias townsendi ochrogenys) is an animal confined to a very narrow but exceedingly long dis- tributional area extending south from the Oregon line as far as Freestone, Sonoma County. Throughout this belt it is conspicu- ously numerous, and is usually the only species of chipmunk present, so that the limits of its range have been easy to mark definitely along the several lines explored. This rodent, by Various geographic tests similar to those I have recounted for other birds and mammals, is clearly delimited away from the - coast at the bounds ‘of the well-known fog-belt to which the red- wood tree and numerous other plants as well as animals belong. The chipmunk, however, depends in no way upon the redwood or any other one plant species as far as I can see, but feeds upon a great variety of seeds and fruits, like many of its con- = geners elsewhere. - That temperature is also a delimiting factor is shown in parts of the range of the redwood chipmunk. But atmospheric humid- ity or cloudiness or rainfall, factors which I have in this case failed to dissociate, together constitute or include the chief controls. 124 THE AMERICAN NATURALIST [Vou. LI THE CASE OF THE BELTED KINGFISHER It is to be observed that specialization for getting a particu- lar kind of food invariably brings with it restriction of range to the territory providing that kind of food. The northwestern belted kingfisher (Ceryle alcyon caurina) is a good example of this. In California we find this bird present at various times of the year both along the seacoast and along various fish-support- ing streams, from the Colorado River to the Klamath River and up the mountain streams to at least as high an altitude as Yosem- ite Valley. The kingfisher is seen during migration in many places away from streams, but it tarries at such times only where its natural diet can be procured, as, on occasion, at fish ponds in city parks. There is a unique instance of a kingfisher observed on the desert catching lizards, but exceptional occurrences of this kind are of course not to be given consideration in making generalizations. It is observable further in regard to this species of kingfisher, that it must have earth banks in which to excavate its breeding tunnels. Lack of these along any stream, otherwise favorable, prevents the bird from staying there through the season of re- production. Furthermore, there is also obvious temperature re- striction ; for, given a fish-producing stream, with banks appar- ently well suited for excavation of nesting places, such as is the Colorado River and its distributaries, and the summer tempera- ture must be at least below that of southern California south of the 35th parallel. That all such streams are well supplied with kingfishers in winter, and are forsaken only during the hot sum- mer, seems to show that a relatively cool temperature is for them in some way or another essential to successful reproduction. We find, then, in the case of the belted kingfisher, that the fac- tors of a requisite kind of food, and a requisite kind of nesting place, both having to do with the structural powers and limita- tions of the bird, together with the factor of the temperature of the summer season, are those that account for the distribution of the species within the state of California, as we find it. THE CASE OF THE MEADOWLARK The western meadowlark (Sturnella neglecta) is a bird of relatively omnivorous diet. Note that I say relatively, for the word omnivorous unmodified would apply only to such an ani- No. 602] SHORTER ARTICLES AND DISCUSSIONS 125 mal as would eat the sort of food that any animal eats, and this is an obvious impossibility for the meadowlark when we con- sider such uncommon articles of diet as wood and petroleum. Compared with many other birds, the meadowlark does use as food a very wide range of plant and animal objects. This food, however, is restricted to a particular habitat source, namely to the meadow. The bird’s entire equipment specializes it for successful food-getting and for escape from enemies upon a grassy plain or meadow. And it is a matter of common obser- vation that its range is sharply delimited in most directions at the margin of the meadow habitat, as where this is interrupted by forest, brushland, marsh, rock surface or sand flat. This is a conspicuous example of what we may call associational restric- tion. But it is not the only way in which the meadowlark is hemmed in. In this connection California again provides crit- ical distributional evidence. e find meadowlarks occupying practically every appropriate meadow, large and small, from the Mexican line to the Oregon line and from the shores of the Pacific to the Nevada line, ez- cept above a certain level on the higher mountains. In travel- ing up the west flank of the Sierras, and this I have now verified along three sections, meadowlarks cease to be observed at ap- proximately the 4,500-foot level, and this in spite of the fact that above that altitude meadows are found which are to all appear- ances ideal for meadowlark requirements. I need only refer to such seemingly perfect summer habitats as Monache Meadows and Tuolumne Meadows. And though, in the winter these would be uninhabitable, so are other meadows (as those in the Modoc region, for instance), which are in summer warm and at that season abundantly inhabited by meadowlarks. By the elim- ination then upon proper grounds of various factors from the list, we have left only three possible factors in this upward de- limitation, namely, decreased atmospheric pressure, decreased air density and decreased temperature of the summer season. Since meadowlarks exist at corresponding altitudes in the warmer though elevated Great Basin region, and since it has been pos- Sible to eliminate positively and in a similar way the first two factors in the cases of many other birds and mammals, these fac- tors are presumably sana pence on the meadowlark; and there is left but one—tem Within the state of APOE meadowlarks, without the 126 THE AMERICAN NATURALIST [Vou. LI slightest detectable subspecific modification, thrive under both the cloudy, humid conditions of the northwest coast belt and under the relatively cloudless, arid conditions of Owens Valley. Factors of humidity, of air and soil, cloudiness, and light in- tensity, seem to avail nothing in checking their spread. With such a degree of associational specialization as is exhibited by these birds there is little chance of a serious competitive struggle with other vertebrates, and no evidence of such has been ob- served. As far as California is concerned, the meadowlark’s range is thus only limited associationally and zonally, that is by the extent of its particular meadow habitat and. by dimin- ished summer temperature below some critical point. The meadowlark well illustrates some further facts with re- gard to distribution. In California it is unquestionably on the increase as regards total population. This is due chiefly to the great extension of habitable territory resulting from man’s oc- -eupancy and cultivation of the land, bare plains, brushlands and even woods being replaced by irrigated alfalfa and grain fields. These the meadowlarks find suitable and invade because of their expansive reproductivity, and soon populate to the fullest ex- tent permitted by the minimum annual food supply. In other words, associational barriers have moved, to the advantage of this particular bird, though at the same time to the disadvan- tage of endemic species of different predilections. I should esti- mate that the total meadowlark population in the San Joaquin- Sacramento basin is now fully three times what it was thirty years ago. Animal distribution is not fixed. It changes with the shifting of the various sorts of barriers, and doubtless also as a result of a gradual acquisition by the animals themselves of the power to overstep barriers, as by becoming inured to greater or lesser de- gree of temperature. The power of such accommodation, or in- herent plasticity, evidently varies greatly among different ani- mals; and at best its operation is very slow. Many species have proved stubborn and have been exterminated, as the factor-lines, or barriers, shifted. By the shifting of, say, two critical factor- lines towards one another, the existence of a species may have been cut off as by a pair of shears. No. 602] SHORTER ARTICLES AND DISCUSSIONS 127 SUMMARY In this paper I have enumerated various factors thought to be concerned with the control of the distribution of vertebrate animals. A number of birds and mammals have been cited to show how we may use our more or less detailed knowledge of their ranges so as to demonstrate the operation of one or several out of the many possible factors as limiters to distribution. The method employed is one of examination, comparison and elim- ination, applied to all parts of the margin of animals’ ranges. The range of any one animal must be examined at all points of its periphery in order that all of the factors concerned may be detected. One factor may constitute the barrier in one section of the periphery of the range of a species, a totally different fac- tor in another section. The results of the geometric ratio of reproduction would bring about areas of occupancy in the shape of perfect circles. But we never find such symmetrical ranges. The very fact that the outlines of the ranges of animals are extremely irregular is significant of the critical nature or inexorableness of the factors which delimit them. These factors have to do with the evolu- tion, persistence and extermination of species. Note that we always have to take into account, in attempting to discern factors of limitation, the animal’s own inherent struc- tural equipment. This prescribes restriction at once in certain regards. Referring again to our list of suggested factors, we find the long-emphasized ones of land to aquatic species and bodies of water to terrestrial species really presenting an extreme mani- festation of associational restriction. Food source, methods of food-getting and safety refuges are involved. It is to be noted further that the factors are various and that the most important factor for one species may prove of little effect with another species. Species do not react uniformly to the same environment. It is undoubtedly always a combination of factors which accounts for an animal’s geographic range in all parts of the periphery of that range. It is most certainly never one factor alone. No one will claim that temperature is the only delimiting agent in controlling vertebrate distribution ; nor could this claim be made for humidity alone, or for food supply alone, or for safety of breeding-places alone. Given a large continuous area, however, as upon the North 128 THE AMERICAN NATURALIST [ Von. LI American continent, one single factor does happen to loom up as being the most frequent delimiter of distribution, or even the ultimately effective one, in greater or less degree, even though other factors be effective also. This factor is temperature. The cases cited illustrate the tenet that in some direction or an- other, temperature beyond certain limits, up or down, cuts off further dissemination. This is part of the basis of the life-zone idea. But, as I have tried to bring out above, this fact is in no way antagonistic to the claim that other factors, as of humidity, food supply, and shelter, also figure critically, giving a basis for recognizing faunal areas and associations. Finally, if our discussion of the subject has been sound, it is evident that data secured through field observation can be so employed as to bring results essentially similar to, and as conclusive as, these secured through laboratory experimentation. JOSEPH GRINNELL MUSEUM OF VERTEBRATE ZOOLOGY, UNIVERSITY OF CALIFORNIA, November 1; 1916 THE AMERICAN NATURALIST Vot. LI. March, 1917 No. 603 THE BEARING OF SOME GENERAL BIOLOGICAL FACTS ON BUD-VARIATION?! PROFESSOR E. M. EAST Bussey INSTITUTION, HARVARD UNIVERSITY I taxe it no one denies that in the Angiosperms vari- ations may be produced in connection with reproduction by means of buds and that these variations may be per- petuated by the same method. Practically, as horticul- turists and plant breeders, we care little about the occur- rence of bud-variations elsewhere in the organic world. Nevertheless, it may help in the orientation of our ideas if we remember that budding is not a rare or unconven- tional method of reproduction. In a generalized form, the earliest method, it has persisted throughout the plant kingdom from the most primitive to the highest and most Specialized types. Sexual reproduction has not replaced it, but has been added to it. Even in the animal kingdom, though eliminated among the higher forms, it still exists. as an occasional alternate method in three fourths of the phyla. Such being the case, it would seem logically to follow, that variation must have been within its possi- bilities. The cause, the frequency, the type, the constancy, the mechanism, of these variations are more debatable, how- ever, and on these questions many biological facts which superficially seem unconnected, have a direct bearing. In 1 Read before the meeting of the Society for Horticultural Science, De- cember 28, 1916. 129 130 THE AMERICAN NATURALIST [Vou. LI fact, on certain phases circumstantial evidence is the only evidence at hand. The exact nature of the cause or causes of bud-variation can hardly be discussed profitably. We may imagine irregularities of cell division directed by combinations of unknown factors, but to describe these factors in concrete terms is at present impossible. At the same time, cause can not be neglected entirely even at present, for cause in a generalized sense is intimately connected with frequency in that vigorous perennial the question of the inheritance of acquired characters. The data on this subject are so voluminous that each for himself must give them careful conscientious consideration. Here no more can be done than to point out some of the conclusions to which I, per- sonally, have been driven, and their connection with the subject in hand. These conclusions are: 1. Broad and varied circumstantial evidence indicates unmistakably that the inheritance of acquired characters has played an extremely important rôle in evolution. 2. Numerous experimental investigations designed to test the possibility of such inheritance directly have either failed utterly or have been open to serious destructive criticism. Direct proof of the inheritance of acquired characters is therefore lacking. 3. If conclusions 1 and 2 are to be harmonized, either modifications are fully inherited so rarely that proof that they do not belong to the general category of chance changes in constitution of the germ-plasm is impossible, or the imprint of the environment is so weak that ex- tremely long periods of time—perhaps geological epochs —are necessary for its manifestation. . Diametrically opposed views on the inheritance of ac- quired characters are held tenaciously and unequivocally by equally eminent biologists. Those who concur with the Lamarekian position are nearly always the students of evolution who approach the subject from the historical or the philosophical side and who rely almost entirely on circumstantial evidence; those who adhere to the side No. 603] BUD-V ARIATION 131 of Weismann are usually experimentalists whose evi- dence is indeed direct, but often questionable, usually capable of various interpretations, and always fragmen- tary. I have been bold enough to grasp both horns of the dilemma, and to plead that each is right from his point of view. My confession of faith is, the environment has been an immense factor in organic evolution, but its effects are shown either so infrequently or after the elapse of so great a time, that for the practical purposes of plant breeding we can neglect it as we would neglect an infinitesimal in a calculation. As Bergson, I think, said: We have been trying to prove that the hour hand moves, in a second of time. A few words will make clear the general arguments in favor of this position, although adequate support to the thesis would require considerable time. In the first place, it seems to me the possibility of the in- heritance of acquirements must be admitted. Weismann’s general contention that the chromatin of the germ-cells is the actual hereditary substance, and that the germ-cells themselves may be regarded as one-celled organisms re- producing by fission and conjugating at certain times, while the body must be considered simply an appendage thrown off from and independent of the germ-cells, is not supported merely by the embryological researches of Boveri, Kahle and Hegner on two or three animal forms, or by the ingenious ovarian transplantations made by Castle and Phillips on guinea pigs, but by all of the recent pedigree culture and cytological genetic work, botanical as well as zoological. Nevertheless it has not been and logically can not be proven that there is no way for en- vironmental forces to produce germ-plasmic changes. Memory is just as strange a phenomenon and Semon has done biology a service by pointing out the analogy be-. tween the mechanical requirements for memory and for the inheritance of somatic modifications. This possibility being admitted, one may well concede the plausibility of the arguments of the numerous pale- 132 THE AMERICAN NATURALIST [Vou LI ontologists, taxonomists and ecologists in favor of La- marckian principles, in spite of the fact that their evi- dence is circumstantial. They take a comprehensive view of the actual conditions that exist among organisms, which is impossible to the experimentalist. It will not do simply to say that the manifest convergence of analogous organs in all parts of the organic world, or the wonderful adaptations of the social insects may be explained in some other way. Of course there may be other explanations for these phenomena; but until more satisfactory ex- planations are forthcoming it is rightfully a custom in science that the adequate interpretation at hand should be accepted. On the other hand it is equally wrong for the ardent devotees of Lamarckism to clutch at every isolated case, every inadequate and abortive experiment, when judicial consideration shows not a single unassailable instance of the inheritance of a somatic modification. Many of these experiments have a direct bearing on bud-variation, and I shall attempt to show where they lead us. 1. Inheritance of Mutilations.—The most radical La- marckians of the present day only go so far as to sup- pose that mutilations are inherited on very rare occa- sions—and they are always zoologists. Ethnology has furnished us with so many histories of mutilations of ears, of lips, of feet, of reproductive organs, long con- tinued in the folkways of a people, that new laboratory experiments have been deserving of the ridicule they have received. Botanists have seldom had any delusions on the subject. Plants are so continually mutilated in the buffetings they receive during life, with no resulf in the next generation, that the non-inheritance of the effects of such injuries is taken as a matter of course. Yet there is occasionally one whose reason fails at the critical moment, and who holds that cuttings from the chrys- anthemum with the large flower resulting from the re- moval of lateral branches, will produce larger flowers in the next generation than will an untreated sister plant. Tf not this, some equally indefensible doctrine. No. 603] BUD-V ARIATION 133 2. Effects of Changed Food Supply.—This last ex- ample was really one of changed food supply induced by mutilation. Change of food supply by other methods has been the basis of scores of experiments, particularly on -insects. Many insects are so very whimsical about what they eat that it seems possible their selective appetite may be an inherited instinct impressed by the environ- ment of countless generations. But the total result of all experiments on them is merely to prove that a second generation may be influenced in the start they get in life by the nutrition of the mother. ' The same thing is true in plants. We fertilize a pop corn to get a bumper crop of good plump healthy seeds, but we don’t expect a dent corn as the next year’s result. We very properly endeavor to give our potatoes a bal- anced ration, in expectancy of a larger yield of well- matured, healthy tubers, but we should not expect these tubers to affect our next season’s supply other than by their health. Similarly we take scions from well-lighted parts of the tree where growth has been good. In such twigs the graft union heals easily and properly, and a fit channel for conveying nutrients is established. In doing these things we are practising sanitation or preventive medicine, as it were, a laudable proceeding. But the hor- ticulturist who promises a different variety by such means is illogical and misleading. Yet we find Bailey so imbued with the idea of making out a perfect case for Lamarckism that he lends the weight of his authority to the following statement among others :? Whilst these ‘ ‘sports?’ are well known to horticulturists they are generally considered to be rare, but nothing can be farther from the truth. As a matter of fact, every branch of a tree is different from every other branch, and when the difference is sufficient to attract attention, or to have com- mercial value, it is propagated and called a ‘‘sport.’’ We may admit the differences between the branches of a tree without cavil. What is more serious is the impli- 2 ‘í Survival of the Unlike,’’ p. 72. 134 THE AMERICAN NATURALIST (Vou. LI cation to the reader that all variations have the same co- efficients of heredity, that a bud-variation is simply a wide fluctuation imposed by external conditions. If this were true the whole organic world would be chaos. But species and varieties do exist. They may be ‘‘judgments”’ in one sense, but in another they are concrete things. In fact we learn this further on in this volume when it suits Bailey’s purpose to have asexually propagated varieties very constant. He says (p. 353): At first thought this fact—that varieties may be self-sterile—looks strange, but it is after all what we should expect, because any variety of tree fruits, being propagated by buds, is really but a multiplication of one original plant, and all the trees which spring from this original are ex- pected to reproduce its characters. 3. The Effects of Disease.—The influence of disease is in many ways like that of malnutrition, in that it is wholly an effect on the physiological efficiency of the reproducing cells. This fact is fairly clear when dealing with diseases with outstanding symptoms. In many instances, how- ever, diseases are not easily diagnosed. There may even be no suspicion that disease is present. In such cases it is rather hard to believe that selection is not accomplish- - ing a positive and radical improvement. A good ex- ample of this is the selection of potato tubers. No one consciously selects a seed potato infected with blight. In- dependent of the probability of reinfection, there is the likelihood that the diseased tuber will not be able to pro- duce a normal plant because of the effect the fungus has had on its own cells. One doesn’t usually believe, how- ever, that rejection of this tuber and selection of the healthy sister is going to lead to the formation of a new race. Yet numerous experiments on potatoes in which it is shown that successive selections have raised the average yield over that of the unselected tubers, are prob- ably of just this type. The race is kept up by the re- jection of diseased tubers, but there is no evidence what- ever that it is improved. I am not going to argue that desirable asexual variations may not occur during this time, and be retained. I say only that any improvement No. 603] BUD-V ARIATION 135 indicated by the raw data must be discounted by the amount of deterioration shown by the unselected variety under similar conditions. Such deterioration is very common, and is due to disease, I believe, rather than to any supposed disadvantage of asexual reproduction per se. This category of facts has been cited under the discus- sion of the inheritance of acquired characters, because such phenomena have perplexed other than botanists. Belief in the transmission of disease, or the effects of dis- ease, by sexual reproduction was current for many years. It is only since the possibility of infection in the egg itself was demonstrated for various diseases, that the true state of affairs has been known. Many other types of experiments designed to demon- strate Lamarckism might be cited, but they have no direct bearing on bud-variation except in so far as a positive case would affect our general attitude on the frequency of their occurrence. They are all similarly negative or questionable, however, so that we must conclude with Weismann that no case of inheritance of acquirements has been proved beyond a reasonable doubt. In other words we grant such a possibility but believe it to be so rare or so gradual that practically it may be disregarded. In reality one could hardly have expected any other conclusion from the type of experiment by which the question has been attacked. Generalized they are some- thing like this. Species X having been grown under en- vironment A for numerous generations is removed to en- vironment B. An adaptive change occurs which persists during several generations. Later the descendants of the original plants are returned to environment A and the change is reversed. When the reverse change occurs more slowly than the original change, it is argued that Lamarckian inheritance is shown. The logic used to draw such a conclusion is indefensible, even if the difficulty of correcting properly for changes due to normal heredity is left out of consideration. If acquired characters are inherited and the changes 136 THE AMERICAN NATURALIST [ Vou. LI induced are reversible, the long period under environ- ment A should have produced a deep impression on species X. Change under environment B should be slow. Reversal should be rapid, however, because of the slight impression environment B must be supposed to have made during the very few generations in which its influ- ence was possible. If acquired characters are not inherited, precisely the same changes should occur, owing to somatic adaptation, the only differences being that the total amount of change in each case would be reached in the second generation after the environment had acted during the earliest stages of the life history. If, on the other hand, the changes induced by environ- ment B are not reversible, judgment must be based on the percentage of individuals changed by B and not re- changed by A. One can readily see how a just judgment would be clouded by probable reversible somatic effects in such cases. Instances of the inheritance of acquire- ments, unless they were very frequent, which from our general evidence is unthinkable, would be indistinguish- able from ordinary chance variations. | Such methods of attack on the subject being almost predestined to failure from the inherent difficulties of the problem, it would seem wiser to seek for a more hope- ful methodology, and in the meantime to accept the only conclusion justified by the data at hand; namely, the inheritance of acquired characters is either so rare an occurrence or so slow a process, that by plant-breeders it may be assumed to be non-existent. One realizes of course that the problem of sexual transmission of somatic acquirements is not necessarily the same as that of asex- ual transmission, but the experimental results have been the same in both eases. Let us, admit, therefore, that. one can not hope to obtain real improvement in asexually propagated varieties merely by selecting buds from plants or parts of plants which have developed under especially favorable conditions. This does not mean that radical environmental changes No. 603] BUD-V ARIATION 137 may not be the direct cause of such a modification. Dr. H. J. Webber once informed the writer that immediately after the great Florida freeze of the early nineties bud- variations in the citrus fruits of that region were greatly increased. Such variations may have been induced by the freezing, but they were not adaptive variations. The conclusions reached thus far have not involved a point of theory which practically is difficult to separate from the one just discussed. It is this. If we disregard adaptive variations, is there not still a reason for select- ing fluctuations? Are there not internal factors which so act that there is a narrow but appreciable variability in an asexually produced population which may offer a basis for selection? In other words, how constant is an asexually propagated race? We can make an effort to compute the frequency of marked bud-variations. But have we any right to assume that these represent the sum total of all bud-variations? Are not bud-variations and perhaps all inherited vari- ations like residual errors, the small ones frequent, the large ones rare? This may be the case, but I should like to emphasize the fact that we have no true criterion for determining the size of a variation. A variation that ap- pears large by visual criteria may be an extremely small change in the constitution of the plant, and vice versa. In view of this fact together with the practical consid- eration that commercially valuable variations must be measurable within a reasonable duration of time—say a lifetime—it is by no means certain that we are going far astray in calculating the frequency of bud-variations by the so-called marked jumps or mutations. Furthermore the range of the fluctuations of asexually propagated varieties of most species is very small even when broadened—as it always is—by the addition of the effects of variable external conditions. It is not hard to recognize a Winesap apple, a Clapp’s Favorite pear or a Concord grape, even though these varieties have heen srown extensively for a considerable number of years. Certain local subvarieties of the pome fruits are said to 138 THE AMERICAN NATURALIST (Vou. LI exist, but they are so extremely rare that one may admit all cases of disputed origin and still have very little asexual variation to account for. I have never seen a published calculation of the fre- quency of bud-variation, and presume it would be of little value anyway, since the general evidence indicates a dif- ferent frequency for different species and even for the same species at different times. It may be mentioned, however, that in personal examination of over 100,000 hills of potatoes belonging to several hundred varieties, 12 definite bud-variations have been seen, a frequency of 1 in 10,000; while just as careful a scrutiny of about 200,000 plants belonging to the genus Nicotiana has brought to light but 1 case. Probably a more practical and just as satisfactory an estimate of the frequency of bud-variations in economic plants is the record of varieties that have been produced in this manner. Naturally such a record contributes little — to theory because only a portion of the variations arising are observed, and only a fraction of those observed are propagated. Further, the origin of comparatively few commercial varieties is known. Yet we may get some idea of what to expect in the future, by noting what has oc- curred in the past. Data gathered in this manner will appear to give us different values depending on how we approach the matter. For example, in Cramer’s wonderful monograph on bud-variation, the grape is cited as one of the species that often varies in this manner. He cites some 25 or more such varieties. Yet in the large list of American grapes in Hedrick’s ‘‘Grapes of New York’’ only one doubtful case of bud-origin is reported. When one re- members that hundreds of varieties of grapes are grown and millions of vines are examined each year, improve- ment by this method seems rather hopeless. And ex- amination of the list of present-day apples, pears, plums and cherries, of the bush-fruits, or of potatoes—all groups of considerable horticultural importance—is still more disappointing. for I venture to say that the varieties of No. 603] BUD-V ARIATION 139 these types in cultivation which have originated as bud- variations can be counted on the fingers of one hand. At the same time it would be wrong not to attribute any importance to bud-variation as a plant breeding ad- junct. Cramer lists several hundred chrysanthemums and over a hundred roses as of bud-origin, as well as a smaller number of varieties in species where bud-varia- tion appears to be less prevalent. Further, Shamel is said to have found bud-variation in the citrus-fruits to be sufficiently common to be worthy of an extended inves- tigation. These species, heen with perhaps the banana and the pineapple—the origin of whose varieties is little known—are the outstanding examples of comparatively frequent bud-variation, picked from our whole long list of cultivated plants. The first two examples, moreover, are species belonging to the domain of floriculture, where rather superficial characters such as color are valuable. In very few other species have bud-variations been re- corded in sufficient numbers to justify us in employing any other adjective than ‘‘rare’’ in describing them. And of the sum total of these varieties only an extremely small percentage are of such a nature that agriculture would suffer a material loss if they were eliminated. Perhaps these last statements appear to imply a very limited type of bud-variations. This is not true. Bud- variations are wholly comparable to seed-variations in their nature, but they are handicapped because recom- binations of variant characters are possible only in sexual reproduction. N bud-variations in a species are simply N variations, but N seed-variations may become 2" seed- variations provided they are not linked together i in hered- ity. An immense advantage thus accrues in favor of seminal reproduction because by far the greater number of commercially valuable characters are complex in their heredity, i. e., they are represented in the germ-plasm by several factors independently inherited. Cramer divides bud-variations into the same classes that de Vries has used for sexual mutations: progressive, 140 THE AMERICAN NATURALIST [Vou. LI where new characters arise; retrogressive, where a char- acter becomes latent or lost; and degressive, where latent characters become active. In this important monograph practically all recorded bud-variations to the date of pub- lication, 1907, are discussed. Yet not a single case of progressive variation is listed. They are all catalogued as retrogressive or degressive. Their classification is correct, however, only when a progressive variation is defined as the addition of a character wholly unknown in the previous history of the species. As examples of what bud-variation does produce we may well study Cramer’s painstaking work. There are losses of thorns, hairs and other epidermal characters, together with an occasional degressive change of the same kind. There are changes in color in vegetative parts. Green becomes red or ‘‘aurea’’ yellow, or a loss of anthocyan occurs. Sometimes the changes are such that the plants remain striped or otherwise variegated. Flowers and fruits exhibit the same types of color varia- tions in considerable numbers. They are mostly losses, with the appearance of what in Mendelian terminology is called hypostatic colors, but once in a great while epistatic colors recur anew. Monstrosities appear. Other parts of the flower take on the appearance and form of petals or of sepals. Dou- bling occurs in several different ways. Fasciations arise. Changes in the character of the reproductive apparatus are not uncommon, sometimes giving us seedless fruits. Plants change their habit of growth. They become dwarf. They retain juvenile characters. They become laciniate, or develop the trait known as ‘‘weeping.’’ Thus we see that bud-variation is not limited in its manifestations; and what is more important, we realize that bud-variations are very comparable to seminal varia- tions, there being hardly a type of change known in sexually reproduced plants that has not been duplicated asexually. What then is the difference, if any, between true somatic changes and true germinal changes in con- stitution? We can get clues which indicate a fairly satis- No. 603] BUD-V ARIATION 141 factory solution of this problem from three different lines of research, pedigree cultures, graft-hybrids and cell- studies. It is a noteworthy fact that the character of the progeny produced sexually by bud-variations has been studied in a comparatively few cases, and in most of these instances self-pollinations were not made. Nevertheless Cramer believes the following conclusions are justified: 1. In a vegetative Mendelization, of the progeny of a branch with the positive character 75 per cent. have the character and 25 per. cent. are without it, while the prog- eny of a branch without the character all lack it. 2. In a vegetative ‘‘Zwischenrasse’’ by which he gen- erally means a variegated race, of the progeny of each type (original and variant), a part retain and a part lack the character, the percentage being variable. 3. In a vegetative mutation, by which he means any change not a ‘‘Zwischenrasse’’ and which did not appear to him to be Mendelian in type, of the progeny of a branch retaining the positive character, either all possessed it or a part were with and a part without it, while the progeny of a branch without the character were all of the same type. If we allow for some deviation due to cross-pollination, I believe that Cramer’s records support this view, and that modern genetic research suggests the interpretation. In the first place, the ‘‘Zwischenrasse’’ are evidently of the type studied principally by Correns and by Baur in sexually reproducing races. They are due to chro- matophore changes, and in many cases at least are not the result of nuclear activity. This being true, one would expect in neither asexual nor sexual reproduction the same type of inheritance for variegated races that obtains for other types of variation. Inheritance will parallel cytoplasmic rather than nuclear distribution; an expecta- tion apparently realized for both types of reproduction. Omitting the ‘‘Zwischenrassen’’’ therefore, we have two phenomena to explain, both of which are similar to cases of inheritance in sexual reproduction where chromatin 142 THE AMERICAN NATURALIST [Vou. LI distribution parallels the facts. In each instance the negative variant—may we call it the recessive—breeds true. In one case the positive variant breeds true, in the other case it gives a simple Mendelian ratio. The mechanism necessary for such phenomena is not difficult to picture. Bud-variations are many times more frequent in hybrids, that is, in plants heterozygous for one or more characters, than they are in pure species. This is the view of Cramer, this was the view of Masters, the eminent English student of bud-variations and tera- tological phenomena, this was the conclusion drawn by the present writer in several articles published some years ago. Such results would be obtained either when the proper germinal change occurs in the chromosome whose mate lacks a character for which the plant is hetero- zygous; or, when there is a dichotomy in which the chro- mosomes of such a pair are not halved but pass the material basis necessary for the production of the posi- tive character to one daughter cell and not to the other, provided the daughter cell lacking the character gives rise to a branch. A bud-variation in a character for which the plant was homozygous would be obtained only when simultaneous like changes occur in both chromosomes of a homologous pair, or when the material basis necessary for the pro- duction of the positive character all passes to one daugh- ter cell, as described above. This hypothesis would account for the fact that hetero- zygotes give rise to bud-variations more frequently than homozygotes, since a germinal change seldom gives rise to a new positive character, and a change in one chromo- some of an identical pair tending toward the production of a recessive, would not show in the latter case. T am not certain that this hypothesis may not with reason be applied to variations that are usually consid- ered seminal. There is no particular ground for assum- ing that such variations occur only at the maturation of the germ-cells. We know that progressive variations of whatever origin are extremely rare. Why then may not No. 603] BUD-V ARIATION 143 most variations be produced in cell divisions previous to the formation of the germ-cells? When recessive we should not note them as bud-variations unless the plant is heterozygous and the mutating cell gives rise to a branch; when dominant we should only note them in the latter eventuality. But if these mutating cells should later give rise to germ-eells, the change would become apparent in the progeny. We have still one other hypothetical case to consider. It is said that some bud-variations are not transmitted by seed. I have not been able to trace an authentic case, but such is the general belief, fathered, I think, by Dar- win. The usual citation is the nectarine, which sometimes is said to give nectarines but at other times gives only peaches. Whether trichome characters only behave thus I do not know. But if that be true, we can understand why if we refer to Winkler’s work on the so-called graft- hybrids. Winkler found that the most interesting of these pecul- lar phenomena are caused by the tissue of one species growing around the tissue of the other. He therefore gave them the euphonious name of periclinal chimeras. Cytological examination showed that the epidermal tis- sues only are from one race, the remaining tissues being from the other. It is really a symbiosis and not a union. Now as the germ-cells are formed wholly from subepi- dermal and never from epidermal tissues, the seeds of these plants always produced seedlings like the type forming the inner cell-layers. It seems probable that the production of the nectarine may be analogous. If the change producing the nec- tarine occurs after the epidermal tissue has been segre- gated from other tissues, the cells which are ancestors of the germ-cells should not be affected and the nectarine seedlings would give peaches, If, on the other hand, the change producing the nectarine, has occurred before any such segregation, the progeny sexually produced should in part be nectarines. THE PROBABLE ERROR OF A MENDELIAN CLASS FREQUENCY? DR. RAYMOND PEARL 1. Wir the increasing volume of Mendelian experi- mentation there is an ever-growing need for adequate and clearly understood tests for the statistical significance of differences between observed results and expectation. A number of different methods of making such tests have been proposed and used by different workers. For ex- ample, early in the discussion of Mendelism, Weldon? made use of the ordinary, and frequently inadequate, ex- pression for the standard deviation of a sub-frequency c= Vnpq. Johannsen? has also made much use of this method. It has several defects. In the first place it as- sumes the Gaussian distribution of the errors, an assump- tion not often strictly warranted, as Pearson‘ has clearly shown, and in many cases grossly in error. In the second place it is not even approximately adequate under certain extreme conditions (frequently realized in Mendelian work) of class frequency. Harris® has proposed the x’ ‘‘xoodness of fit’’ test for comparing observed and ex- pected Mendelian distributions. There are several fea- tures of this method which greatly limit it for such use. Among these are (1) its failure to make correct allowance for ‘‘tail’’ frequencies (it is just this class of very small ecm which one most often wants to test in prac- Papers from the ai Laboratory of the Maine Agricultural Ex- sien Station No. 108. e author is greatly indebted to his assistant, Mr. John Rice Miner, for Ki laborious arithmetic involved in Section 6 of the paper per. 2 Weldon, W. F. R., ‘‘ Mendel’s Laws of Alternative Inheritance in Peas,’’ pies bo Vol. I, pp. 228-265, 1902. 3 Johannsen, W., ‘‘Elemente der exakten Erblichkeitslehre,’’ Zweite Aus- gabe, Jen 1913. 4 Pearson, K., ‘‘On the Influence of Past Experience on Future Expecta- en ed ner Maj, verpnti 1907, pp. 365-378. rris, J. A., ‘‘A Simple Test of the Goodness of Fit of Mendelian hie? ? AMER. Nar, Vol. XLVI, pp. 741-745, 1912. Cf. also Pearson, K., and Heron, D., ‘‘On Theories of Association,’’ Biometrika, Vol. IX, pp. 159-315, 1913. 144 No. 603] MENDELIAN CLASS FREQUENCY 145 tical Mendelian work), and (2) the fact that the test takes into account only the magnitude of the error and not its direction (i. e., whether in excess or defect) in any par- ticular case. (3) It gives a result not particularly well adapted to the actual needs of Mendelian research. The x? test gives a measure of the goodness of fit of the whole distribution, and only that. Now besides being in- terested in that point the Mendelian worker quite as often wants to know, in addition, something about the prob- ability that particular classes observed are significantly different from the expected. To that sort of knowledge the x? test helps him not at all. It is an ‘‘all or none” sort of method.® 2. It has seemed to the writer that it would be useful to discuss methods of determining the true probable error of each class frequency in Mendelian distributions as a supplement to the x? test, and for use in cases where it is not applicable. The fundamental theorems have all been given by Pearson’ in a very important paper, which is apparently almost entirely unknown to biologists. The purpose of the present paper is first to show the appli- cability of these theorems to the problem in hand, and second to point out some matters regarding the practical use of the method likely to be helpful to biologists with but little mathematical training who may attempt to use it. 3. In the paper referred to, Pearson, starting from Bayes’ theorem, shows that the distribution of chances of an event occurring in a particular way in a second €I have earlier pointed out other objections to the x2 test in Mendelian work, in particular its total failure to deal with cases where experiment yields a small frequency on classes where the expectation is zero, and need not further discuss them here. I have never thought it necessary to make any rejoinder to Pearson’s Ne on bitter reply to my criticism, nor do I yet. The x2 test leads to this absurdity: if I perform a Men- delian experiment in which I get ni thousand million offspring agreeing perfectly with expectation save for one lone individual (perhaps a mutation, Perhaps a mistake in the record, or what not) which is of a sort not ex- pected, then Pearson and the x2 test agree that the probability is infinitely eet that the ten thousand million offspring do not follow Mendelian law! earson, K., loc. cit. 146 THE AMERICAN NATURALIST [Vou. LI sample from a population from which a first sample has produced a certain value is given, not by the ordinates of a normal curve of errors, as is commonly assumed in writ- ings on the theory of probability, but by the successive terms of a simple hypergeometrical series. In an earlier paper the same author had solved the problem of the momental properties of the hypergéometrical series. Combining the two results he was able to derive the neces- sary equations for the complete solution of the problem of probable errors in sampling. We may proceed at once to the exposition of these results, referring the reader for the proofs to the papers of Pearson cited. Let it be supposed that a first sample of n= p + q be drawn from the population, p denoting the number of times the event dealt with occurs in the v trials, and q the number of times it fails. Write whence of course p+q=1. We then have for the chief constants of the error dis- tribution for a second sample, of magnitude m, drawn from the same population the following values: ae H o eg é Mean’ = mp + 749 — P) (i) Mode = the integral ees of mp + p, (ii) Standard Deviation = | m (z p +: wF 1-2) x (a-pa) +g) ao These values are entirely general, and independent of the values of n, m, p and q. Under certain circumstances, 8 Pearson, K., ‘‘On Certain Properties of the Hypergeometrical Series, and on a "Fitting of such Series to Observation Polygons on the Theory of Chance,’’ Phil. Mag., Feb., 1899, pp. 236-246. ə From origin at the lower range end, or r= 0. No. 603] MENDELIAN CLASS FREQUENCY 147 as when n is very large as compared with m, and neither p nor q are very small, (i) and (iii) are obviously capable of being put in much simpler form and still giving a sufficiently close approximation to the true result. For Mendelian work, however, where frequently neither of these conditions are even approximately realized, it will in general be better to use the full expression as given above. The ordinates of the error distribution (the chances of different occurrences) are given by the successive terms of the hypergeometrical series m p+1 , m(m—1) (p+1)(p+2) cas -0147 ‘qtmt 2 UGA ap ME 2) (p+1)(p+2)(p+3) 3 (q-+m)(q-+m—1)(q+m—2) m(m—1)(m—2)(m—3) j 4 x (p+1)(p+2)(p+3)(p+4) (q-+m)(q+m—1)(q-+-m—2)(q+m—3) (iv) +etc. l, where T(q+m+1)0(n+2) Co= G4 DT tm+2)" (v) As we shall presently see, the calculation of the terms in (iv) becomes a tedious and laborious matter when the number needed is at all considerable. Under such cir- cumstances, and when in addition m and are even mod- erately large, equation (iv) may be greatly simplified, without significant loss of accuracy, by the use of Ster- ling’s theorem (to the bracket) or by Forsyth’s approxi- mation for such of the factorials as are not included in the range of the Pearson!’ tables. Thus we have, by Sterling’s theorem, apenas that r denotes any term in the series, and writing s=m—rT, 10 ‘Tables for Statisticians and Biometrieians,”” ¢ edited by Karl Pear- fon, Cambridge, 1914. 148 THE AMERICAN NATURALIST [Vou. LI (P T | mrtg + s)etet8 ; Using Forsyth’s approximation, which is extremely ac- curate, one gets Etri ientrar katoi N hes pir | Esa) QH gm) +R) J * The gamma terms in (v) will, of course, be caleulated by some one of the well-known approximations (e. g., Sterling’s, Pearson’s, Forsyth’s) or by interpolation from a table of factorials (Pearl''). 4. The proposal which I wish to make for the expres- sion of a Mendelian result is that the expectation be expressed as the quartile limits for each class frequency in a second sample of the same size as the observed sample. In using such an expression it must be clearly understood that it does not measure the goodness of fit of the distribution as a whole, because it takes no account of correlations in errors. What it does give, in supple- ` ment to the x? test, is the limits of probability of each class frequency, taken by itself. The ordinary expression for a probable error (e. g., P. E. mean = + .67449(c/Vn)) gives the quartile limits (i. e., the limits within which one half the frequency oc- curs) on the assumption that the distribution is Gaussian, since in such a distribution of unit area the quartile limits are .6744898 . . . times the standard deviation on either side of the mean. But in our persent work we are making no assumption that the error distribution is Gaussian. Consequently we must determine the quartiles directly from the distribution. In cases where the number of terms is not too great the ordinates may be calculated from (iv) or (vi) and summed to find the quartile. In many cases, however, this would be practically too tedious CeCi 11 Pearl, R., ‘‘ Interpolation as a Means of Approximation to the Gamma Function for High Values of n,’’ Science, N. S., Vol. XLI, pp. 506-507, 1915; ‘‘On the Degree of Exactness of the Gamma Function Necessary in Curve Fitting,’’ Ibid., Vol. XLII, pp. 833-834, 1915. No. 603] MENDELIAN CLASS FREQUENCY 149 an operation, and we may resort to an approximate method. The simplest one is to take .67449¢ on either side of the median, which is approximately determined by remembering that the median lies between the mean and the mode and approximately twice as far from the -= mode as from the mean. The criterion of whether this — method of fixing the quartile limits may be safely applied will be found in the value of the skewness, Sk. In prac- tical work this approximate method will give sufficiently accurate results unless the skewness is very large, say > 0.6. 7 We have by definition Sk = Q1& (viii) Hence having calculated the values of mean, mode and « by (i), (ii) and (iii) we can readily obtain (viii), since d= mean — mode. We may now pass to the consideration of some numer- ical examples, by means of which certain facts can be better brought out than by further theoretical discussion. 5. As a first and simple example we may take some data, recently published by F. L. Platt,!? on the results of mating Blue Andalusian fowls. On account of the fre- queney with which the Blue Andalusian case is cited as a paradigm in Mendelism, coupled with the great dearth in the literature of exact statistics of actual matings of this breed of poultry, it seems especially worth while to dis- cuss these statistics furnished by Mr. Platt, on the au- thority of Mr. W. J. Coates, a breeder of Andalusians. Table I gives the data, and in the last line, the Men- delian expectation expressed in the form suggested in this paper. The occurrence of the ‘‘dark reds,’’ which Mr. Coates informs us had a pattern like a Red Game, is a phenom- enon not mentioned in textbook accounts of Mendelian in- heritance in the Blue Andalusian. In the present con- 12 Platt, F. L., ‘‘ Western Notes and Comment,’’ Reliable Poultry Journal, Vol. XXIII, p. 665, 1916. 150 THE AMERICAN NATURALIST [Vou. LI TABLE I SHOWING THE RESULT oF Martine BLUE X BLUE IN THE BLUE ANDALUSIAN or POULTRY (Coarres’ DATA) Oftspring Mating White Blue Black oo | W. O ERE EE E E EN EEES 4 10 3 1 MESS ESA REAM CANT BOs sg SANA EE 4 5 2 0 Cri 3 3 0 3 BP Di 0 12 1 0 fi OPRESL AER NEES DEL Meare ene CGM a PI kon TRS ee eo 3 3 1 0 SUA GDNGEVER Co. ce Cre Ot eet | 14 | 33 7 4 Total observed in Dre a white, (2) | me blue, (3) pigmented not blue............ 14 | 33 11 -If the true ratio is 1 white : 2 fears 1 pig- mented not it is an even chance, considering each class by itself, that the _ frequencies in a sample of this size fall between. ie a ss SS 11.5 and | 25.8 and 11.6 and 17.8 32.9 17.8 nection, however, we can not pursue that point, but will group together, as in the penultimate line of the table, the blacks and reds as ‘‘pigmented, not blue,’’ and assume that the three classes should occur in a 1:2:1 ratio. Do the actual results bear out this assumption, having regard to the errors of sampling? Examining the last two lines of the table, it is clear that each observed class, taken by itself, is by no means an im- possible approximation to what would be demanded by a 1:2:1 ratio. The blues and the ‘‘pigmented not blues’’ fall outside the range for which the probability is 4 but only slightly outside. It would be practically an even bet, if Blue Andalusians really follow a law of 1:2:1 segregation when bred together, that any particular sam- ple of 58 offspring would show in each particular class as great a deviation as the present sample.’ Now we may consider in detail the mode of calculating the figures in the last line of Table I. 13 Always on the assumption, of course, that it is legitimate to lump the blacks. and reds together. There is room for scepticism on that point, but we are here only concerned with the case as an illustration of method. No. 603] MENDELIAN CLASS FREQUENCY 151 We have, by hypothesis, and from the statistics E =D8. A distribution of 58 in a 1:2:1 ratio is 14.5:29:14.5. Assume a first sample of 58 to show exactly this distribu- tion 14.5 :29:14.5, what will be the mean frequency of one of the end classes, say white, expected in a second sample of 58? We have ac A45 p= 5S. = 205, g = .75, whence, by (i), Mean = 14.9833, and, by (ii), Mode = 14, a=. .9833. By the approximate method we get Median = 14.656 approximately. The standard deviation from (iii), is o = 4.6364, and, by (viii), SR aye? A Actual tests with curves of a degree of skewness no greater than this show that the approximate method gives the quartile limits with sufficient accuracy for practical purposes. We have for the approximate quartile limits, 67449 X 4.6364 = 3.1272. This value, added to and sub- tracted from the median 14.656, gives the results set down in the last line of Table I. Exactly the same procedure, with different numbers, is followed in the case of the blue column. 6. Let us now consider a more completely worked out illustration. Some time ago Mr. Alexander Weinstein, of Columbia University, consulted the writer in regard to a 152 THE AMERICAN NATURALIST [Vou. LI problem which had arisen in connection with his Men- delian breeding experiments. A certain type of mating gave the following class frequencies: (2) 6363 + 579 + 3638 +- 1208 + 128 + 115 + 350 + 6 = 12387 ... (a). Another group of matings gave a total of 9,017 off- spring, of which 30 fell in the v class, this being the only class in regard to which a comparison is to be made. On certain theoretical grounds the percentage frequency in this æ class in the second sample would be expected to be 0.582 times the percentage frequency of this same class in the first sample. The question is whether the actually ob- served frequency of 30 in this second sample is such as could reasonably be expected to occur if the theoretical assumption actually were the fact. It will be seen at once that, owing to the very small ab- solute frequency of this x class in both samples, ordinary probable error methods will be of no avail. Approaching the problem by the method here proposed, we have, as basic values for the computations, n= 12387 m= 9017 pe 41d, ¢ = 12272 p = p/n = .009284 g=q/n = .990716. Whence we have for the mean in the second sample of 9017 by (i) Mean = 84.428137 and by (iii) o = 12.020652. By (ii) the mode—83, whence d—1.428137 and 1.428137 sk = 12.020652 = 0.118807. No. 603] MENDELIAN CLASS FREQUENCY 153 Working directly from the moments of the hypergeo- metrical series and, in effect, replacing that series with a true curve, we find- Mode = 83.222141, d= 1.205996, and Sk = 0.100327 F .008696. TABLE II Meio wiied THE SUCCESSIVE ORDINATES OF THE HYPERGEOMETRICAL SERIES R THE SECOND SAMPLE r G Sum r C, Sum r cx. . Sum 35| .000000 | .000000 | 73 | .022620 | .182838 | 111 | .003215 | .983125 36} .000001 | .000001 | 74 | .024355 | .207193 | 112 | .002740 | .9858 37 | .000001 | .000002 | 75 | .025996 | .233189 | 113 | .002325 | .988190 _ 88] .000002 | .000004 |. 76 | .027539 | .260728 | 114 |. .001964 | .990154 39 0 77 | .028943 | .289671 | 115 | .001651 40 | .000005 | .000012 | 78 319854 | 116 | .001382 | .993187 41 000020 | 79 | .031236 | .351090 | 117 | .001152 42| .000013 80 | .032085 | .383175 | 118 43 000053 | 81 | .032715 | .415890 | 119-| .000791 | .996087 44| .000031 82 | .03311 44 120 1 | .996738 eo 000130 033285 | .482291 | 121 | .00053. 99727 46| .000068 | .000198 033220 | .515511 | 122 | .000436 | .997707 47 297 | 85 54844 23 | .000354 | .998061 48 | .000142 86 | .032419.| .580859 | 124 | .000287 | .998348 49 000639 | 87 | .031706 | .612565 | 125 | .000231 | .998579 50| .000279 8 | 88 643370 | 126 | .000186 | .998765 51 001301 | 89 | .029738 | .673108 | 127 | .000149 | .998914 52| .000520 | .001821 | 90 | .028526 | .701634 | 128 | .000119 3 | .000696 | .002517 | 91 | .027193 | .72882 29 999127 t| .000919 | .003436 | 92 | .025763 | .754590 | 130 | .000075 | .999202 >| .001199 63 93 | .024262 | .778852 | 131 | .000059 | .999261 ìà | 001545 | .006180 2271 15 132 | .000046 57 | .001967 | .008147 | 95 | .021136 | .822699 | 133 | .000036. 58| .002476 96 | .019556 2255 | 134 999371 59| .003082 | .013705 | 97 | .017991 | .860246 | 135 | .000022 60| .003793 | .017498 | 98 16459 | .876705 | 136 | .000017 | .999410 61| .004617 99 | .014974 | .891679 | 137 | .000013 99423 62 .005561 | .027676 | 100 522 38 | .000010 | .999433 63 2 305 | 101 | .012195 | .917424 | 139 999441 64| .007821 | .042126 | 102 | .0 928341 | 140 | .000006 | .999447 65 051261 | 103 | .009722 | .938063 ! 141 | .000005 | .999452 66 | .010569 830 | 104 | .008614 | .946677 | 142 | .000004 67 | .012109 939 | 105 | .0075 .954270 | 143 | .000003 | . 68| .013743 | .087682 .006 144 | .000002 | .999461 69| .015455 | .103137 | 107 | .005813 | .966743 | 145 | .000001 | - 2 40} .017223 | .120360 | 108 | .005049 | .971792 | 146 | .000001 | .999463 71| .019025 | .139385 | 109 |. .976156 | 147 | .000001 | .999464 12| .020833 | .160218 | 110 | .003754 | .979910 The two sets of values are evidently sufficiently near together to be used interchangeably for most practical 154 THE AMERICAN NATURALIST [Vou. LI purposes, a sort of result which is familiar to any one who has had any considerable experience with the method of moments. To return now to the series we find, using 10-place logarithms in the intermediate computations and the Forsyth approximation, log Co = 72.8493814 — 100. We have now to calculate the successive terms of the series. If this were done for the whole range it would in- volve a literally colossal amount of labor. Fortunately this is not necessary. We need only take that part of the range which includes appreciable frequencies. By a few trials we find that this part of the range begins with r= 36. In Table II are given the frequencies for the several terms in the series between r— 36 and r—147 in- clusive, the total area being taken as unity. To reduce these frequencies to the actual numbers for the second sample we have only to multiply in every case by 9017. We have calculated C;, by (vii) and used the Forsyth Co. From this table we easily deduce Median = 83.5331, Lower quartile = 75.6104 (ix) Upper quartile —91.8218. Now, remembering that if the same law holds for the second Mendelian distribution as for the first we should expect the x class in that distribution to be 0.582 times the value of the same class in the first distribution, we have Expected mean value of v class in second distribution = 49.14 Expected modal value of x class in second distribution = 48 Expected lower quartile value in second distribution — 44.01 Expected upper quartile value in second distribution — 53.44 No. 603] MENDELIAN CLASS FREQUENCY 1565. The actual experimental value obtained was 30, which is far below the lower quartile. From Table II we find, remembering again that on a priori grounds the experi- mental frequencies are reduced by the factor 0.582, that if the two distributions were really samples of the same population, obeying the same Mendelian laws, it would be expected that the æ class would show a frequency as low as or lower than 30, only 18 times in 10,000 trials of samples of 9017. Or, in other words, the odds against so low a value as 30 are about 556 to 1. These are about the same odds as those associated with the occurrence of a deviation 4.63 times the probable error (cf. Pearl and Miner"). We may, therefore, conclude with great certainty that the value of 30 is significantly smaller than would be ex- pected to occur in the x class on the basis of chance (de- viation due to random sampling) if the two distributions were really samples of the same population. Let us now go back and approach the problem de novo by the approximate method suggested in section 4. We ave d—= mean — mode = 1.4281, a ns = .4760, ‘84.4281 — .4760 =—83.9521—median (approx.), 12.0207 X .67449= 8.1078, 83.9521 — 8.1078 —75.8443— Lower quartile, 83.9521 + 8.1078 = 92.0599 = Upper quartile. Comparing these values, in the obtaining of which all the tremendously tedious and time-consuming arithmetic involved in calculating Table II was avoided, with those shown in (ix) makes it quite evident that for all practical statistical purposes the approximate method would have given sufficiently accurate results. 14 Pearl, R., and Miner, J. R., ‘‘A Table for Estimating the Probable Significance of Statistical Constants,’’ Me. Agr. Expt. Stat. Ann. Rept. for 1914, pp. 85-88. 156 THE AMERICAN NATURALIST [Vou. LI SuMMARY In this paper is presented a method of calculating and expressing the errors, due to random sampling, of a Men- delian class frequency. The method consists essentially in expressing each expected Mendelian frequency as the probable quartile limits for that class frequency in a sup- posed second sample of the same size as the observed sample drawn from the same population. These quartile limits are determined from the ordinates of a hypergeo- metrical series. Various simplifications of method are suggested and illustrated. The method is suggested as a supplement to, not as a substitute for, the x? test for the goodness of fit in Mendelian distributions. OBSERVATIONS ON THE ECOLOGY OF THE PROTOZOA By LEON AUGUSTUS HAUSMAN CORNELL UNIVERSITY A great deal of excellent work has been done in census taking among the Protozoa, and numerous catalogues of species with descriptions and accompanying plates have been issued for various states or portions thereof. This present communication, however, is a slight contribution toward census taking of a different sort, in that it aims to set forth some facts regarding the different types of Pro- tozoan habitats and the species usually associated with each. It is well known that certain common forms, used as type species in the laboratory (such as Ameba proteus and Paramecium caudatum), can be found only in certain fairly definite environments. The ecology of such forms is well known, but the bionomics of the majority of the Protozoa is still virtually a res ignota. The study of en- vironmental units and Protozoan communities will con- tribute to a more intimate knowledge of these elusive forms. Although little is definitely known concerning the ecol- ogy of the Protozoa, yet I think that we are in position to say that Protozoan inhabitants vary with their varying en- vironments. A record of the inhabitants of a marsh pool will not include the same species normally found in clear - running streams, nor will cold waters yield the same forms as warm. But we can go still farther, I think, and say that the various portions of any given environmental unit, even though they differ ever so slightly from one another, will each have its own characteristic group of organisms. The factors which are accountable for the variation in Species or numbers of individuals of any species in such instances ° may be hardly discoverable, but they are none the less potent. Thus the entrance point of a tiny thread 158 THE AMERICAN NATURALIST [Von LI of clear spring water in the midst of an impure pool may furnish an abundance of Monas and Cercomonas, and because of the presence of these, also the larger carniv- orous forms such as Dileptus or Onychodromus. Or the growth of myriads of bacteria at another point may induce the increase in the numbers of Vorticelle, Amebe, or Holophrye. A complete study of the multitudinous minute variations in the environments of Protozoa, and their complex relationships with the associated species is still far away. But I believe that we are now ready to undertake a preliminary survey of a number of typical en- vironments and to ascertain what genera and species may be normally expected to occur in each. Of the many factors responsible for the localization of species, the following are perhaps first in importance: light, character of food, temperature of water, chemical content of water, presence or absence of enemies. Light exerts a powerful influence in determining the distribution of species within a given body of water. Thus one will find certain species in sunlit areas and certain others in shaded ones. Not all of the Protozoa respond to ordinary light stimuli, as has been shown experimentally by Jennings. He has found that comparatively few of the ciliate infusoria react to light stimuli. The flagellates, on the other hand, show a definite reaction, congregating in that region where the light is ordinarily strongest. The reactions of Euglena viridis and Cryptomonas ovata are the forms which have been most carefully studied in this regard. Jennings! further showed that Paramecia collect in those regions of the water which contain a trace of any weak acid, and this I have noted to be true of certain other ciliates. This may be a dominant influence in localizing forms in pools and streams in marshes where organic acids are usually present. The presence of supernormal amounts of oxygen, carbon dioxide, calcium carbonate, iron, and other dissolved substances may also affect the distribution of species. 1‘*Contributions to the Study of the Behavior of Lower Organisms,’’ H. S. Jennings, Washington, 1904. No. 603] ECOLOGY OF THE PROTOZOA 159 Not only may the various stimuli arising from the dif- ferent elements of the environment assort species into communities, but the effects of these stimuli may not always be the same. A given stimulus may at one time produce a certain type of reaction, and at another time a somewhat different type. This fact, also, experimentally verified by Jennings, led him to suppose that something analogous to, if not identical with true varying ‘‘ physio- logical states’’ obtains in the Protozoan body. If this be so then the problem of Protozoan ecology is still further complicated, as far as any assignment of species to definite fixed local environments is concerned. For collecting Protozoa and water samples the writer has found the following useful: a small silk plankton net (with draw string bottom) about six inches in diameter and ten inches deep; a glass or metal pipette, fully a foot in length, operated by a compression bulb at one extrem- ity, for sucking samples of light sediments from the bot- toms of pools and streams; several small glass dipping tubes; a large table spoon; a thermometer, and a plentiful supply of variously sized water tight jars with screw tops. With such an outfit as this the material described in the following paper was secured. The samples were examined immediately after being brought into the lab- oratory, in order that the proportions of species might be accurately recorded. From each sample ten slides were searched. The sig- EE of the terms few, numerous, rare, etc., is as fol- ows: Over ten individuals of one species per slide ....... abundant Average of 6206s aroiu ke ben ee EC a numerous Average Of 2-600 is ans sis se sa Vas eid several Ave OF ME SEATS asan e ee Seen few Average Ol Tell hin B o nc sv cies cnenices rare The samples were taken from the uppermost layer of silt on the bottoms of pools, streams, etc., the utmost care being taken not to disturb the sediment below or to roil the water. Upwards of fifty samples were secured repre- senting five distinct environmental types. These five types described are: ` 160 THE AMERICAN NATURALIST [Vou. LI No. 1. Characteristic marsh pools. No. 2. Clear cold lakes and streams with no plant life. No. 3. Clear springs and streams with plants abundant. No. 4. Small, clear pools containing organic sediment in decomposition and fed by pure, cold rills. o. 5. Ditches and pools choked with alge, water warm. All of the material was taken during the months of late spring, summer, and early fall, while the aquatic vegeta- tion, and the semiaquatics about the water margins, were in vigorous growth. ENVIRONMENTAL Tyrer No. I: Marsu Poors Characteristic frog-inhabited marsh pools, containing much decaying vegetable matter, supporting rank growths of typical marsh plants about the margins, covered with lily pads and filled with Ceratophyllum, Myriophyllum, Sagittaria, duckweed, ete. Water warm, and emitting characteristic swampy odor. Predominant forms: Ameba limaz, several Difflugia globulosa, numerous Ameba proteus, several Euglypha alveolata, numerous mæba radiosa, several Oikomonas sp., numerous Arcella vulgaris, numerous Peridinium cinctum, abundant Carchesium polypinum, numerous Stylonychia mytilus, numerous Codonocladium umbellatum, numer- Stylonychia pustulata, numerous ous ‘ Coleps hirtus, abundant Synura uvella, numerous Coleps sp., abundant Trinema acinus, numerous Difflugia acuminata, numerous Vorticella microstoma, numerous Difflugia corona, numerous Vorticella nutans, numerou Other forms present in less numbers (few or rare) : Actinophrys sol Euplotes carinata Arcella dentata Euplotes charon Arcella discoides Halteria grandinella Anisonema obliqua Heterophrys myriapoda Astasia Be Holophrya sp. Biomyza vagans Lacrymaria olor Blephoriewa lateritia Lionotus wrzesmosky Centropyxis aculeata Lionotus sp. Colpoda sp. Loxophyllum Cothurnia maratima Paramecium Saidia Dactylosphærium radiosum Pleuronema sp. Dinomonas voraz Prorodon griseus No. 603] ECOLOGY OF THE PROTOZOA 161 Stentor ceruleus Urocentrum turbo Trichodina pediculus (on Hydra Volvox globator fusca) Vorticella sp. Blepharisma lateritia (Fig. 1) is usually described as showing pink or even reddish hues, and occurring rarely colorless. My experience has been to find the greater num- ber of individuals colorless and a very few showing even a faint trace of pink. In reproducing the colors of Pro- tozoa in plates there seems to be a tendency to represent them more vividly than they occur in nature. It is inter- esting to place a small chart of spectrum colors on the table near the base of the microscope and compare them with the hues of those species of Protozoa usually repre- sented as brightly colored. It has been my experience to find that the (depicted) decided colors of some of the Pro- tozoa show themselves to be only the faintest tints. Synura uvella (Fig. 2) is not entirely colonial in its habit. I have found it sometimes singly, and often in pairs. The normal number of individuals in a colony is from 10 to 20. As many as 35 in one group have been re- corded. The flagella are invisible unless iodine or some good stain has been used. The Coleps sp. (Fig. 3) which I found so frequently in this one group of samples I have never found since. It was appreciably smaller than Coleps hirtus (Fig. 4), being about 30 microns in length, whereas hirtus is 50. Hirtus is a species exhibiting a notable constancy in dimensions, and this I found true also of all the numerous individuals of Coleps sp. which I measured. No transitions in size from one to the other could be found. Coleps sp. was not the result of fission on the part of hirtus, for in all the material my search revealed not a single individual un- dergoing division. In appearance and activities the one form was the exact counterpart of the other, with a soli- tary exception: the movements of Coleps sp. were at all times much more rapid than those of hirtus. Lacrymaria olor (Fig. 5) described as a very variable Species, is variable in size almost entirely. Its form is quite constant, and offers a virtually certain criterion for identification. : 162 THE AMERICAN NATURALIST [Vor. LI Fig.2 Synura uvella Ea Fig.9 Cero fa ve se ae Fig.10 Cercomonas termo,normal and. abnormal forms No. 603] ECOLOGY OF THE PROTOZOA 163 PLATE II T Monas irregularis Monas Dallingeri Monas fluida Fig.11 Fig.l? Frontonia sp,2 Fig.18 Frontonia sp.l 164 THE AMERICAN NATURALIST (Vor LI The body of Loxophyllum (Fig. 6) is surprisingly elas- tic. For this reason the creature undergoes rapid varia- tions in length. Even within the confinement of a viscous gelatin solution which checks the motions of all of the other ciliates its strength is such as to enable it to per- form various evolutions over the slide. In staining it fre- quently dies in a contorted and contracted condition. Before staining, it should be narcotized with a weak ethyl alcohol or killed with .01 per cent. osmic acid. In Astasia (Fig. 7) the smaller secondary flagellum is sometimes either unusually filamentous, and therefore in- visible, or lacking altogether. In many individuals no amount of manipulation of light directions and intensities or stains will reveal the smaller flagellum. Urocentrum turbo (Fig. 8) can always be recognized by the frenzied and jerky rotation which accompanies its ex- tremely rapid movement through the water. It usually comes to rest either close beside or buried among masses of alge or other convenient material. In this resting con- dition it is liable to be overlooked. At such times its only evidence of activity is a scarcely perceptible twitching. This hiding in algal masses has all the appearance of being a deliberate attempt at concealment. Suddenly without \warning its crazied rotation through the water is begun :again. When swimming with retarded velocity through a ‘gelatin solution, functional anterior and posterior por- tions of the body are recognizable. ENVIRONMENTAL Type No. II: CLEAR Cotp WATERS ckInG PLANT GROWTHS Sediment, composed of quartz and shale sands free from „organic silt and supporting no plant life save a few Dia- -toms, in clear, pure, cold lakes and streams; temperature .of water, cir. 56° Predominant forms: . Astasia sp., abundant Nostolenus orbicularis, several . Holophrya sp., numerous Nostolenus sp., several iHolostichia vernalis, numerous No. 603] ECOLOGY OF THE PROTOZOA 165 Other forms present in less numbers (few or rare) : Cercomonas termo Euplotes sp. Cercomonas crassicauda Lionotus sp. Chlamydomonas sp. Paramecium bursaria Colpidium sp. Stylonychia mytilus The nucleus of Cercomonas crassicauda (Fig 9) is located near the functional anterior end of the body, and is not easily seen. The anterior flagellum is likewise diffi- cult to distinguish, even upon staining, being exceedingly filamentous. It seems to have little to do with locomotion, and I believe that it is used principally as a sort of an- tenna. The propulsion of the creature is apparently accomplished by the lashings of the stout posterior fla- gellum, no matter which end of the organism is directed forward. Cercomonas termo (Fig. 10), from 5 to 15 microns, is extremely variable in shape in the adult form. Its com- monest and therefore characteristic appearance is roughly heart shaped, the flagellum arising from the broader ex- ~ tremity of the body. The young are not so variable, but although constant in general form, confusingly resemble the monads, in particular Monas irregularis. Their mi- nuteness (.5 to 2 microns) makes them difficult to identify. They should be stained with a very weak solution of iodine or acetic methyl green, since the stronger stains, used for the larger Protozoa, cause them to lose their character- istic outline and to disintegrate in a short time. ENVIRONMENTAL Type No. IIT: Crear Frowrne Water WITH ÅBUNDANYT PLANT LIFE Clear springs and streams supporting vigorous growths of such alge as Spirogyra, Drapernaldia, etc., and water cresses; bottoms covered with Diatoms (often chiefly Meridion circulare), Desmids, and Oscillatoria. Tem- perature of water, cir. 66° F. Predominant forms: Ameba proteus, abundant Chilomonas paramecium, abundant Ameba proteus, flagellospores (%), Colpidium sp., several abundant Colpoda inflata, several Chilodon cucullus, abundant Difflugia constricta, numerous 166 THE AMERICAN NATURALIST [Vou. LI Difflugia globulosa, numerous Monas irregularis, abundant Diffiugia lobostoma, numerous Monas Dallingeri (?), abundant Holophrya sp., several Nostolenus orbicularis, numerous Holostichia vernalis, several Oxytrichia pellionella, cages Monas fluida, abundant Prorodon teres, numerou Other forms present in less numbers (few or a : Arcella artocrea Euglena minima Aspidisca costata Frontonia s Astasia contorta Holophrya sp. Cercomonas crassicauda Lionotus oa Chilodon sp. Masti Chlamydomonas sp. Giyimi pee Colpidium striatum Oxytrichia bifaria Dile Pe Enchelys sp. Parameéci caudatum Epiclintes radiosa Stylonyenta mytilus Euglena viridis Stylonychia pustulata The identification of the Monas species (Fig. 11) is very difficult. Members of this genus resemble some of the members of the genus Bodo. Because of the hyaline char- acter of the body its outline is not easy to distinguish. I have found that the most successful method of treatment before identification is attempted, is staining with a strong aqueous solution of iodine and potassium iodide. This oth kills and stains. Examination must be made almost immediately since the organisms begin to lose their char- acteristic form in a short time. , Within the genus Holophrya have been provisionally laced a number of forms very closely allied in general characters. For a considerable number of these a separa- tion into species has not yet been made. The pseudopodia of the Mastigamceebe (Fig. 12) do not invariably disappear with the appearance of the flagellum. They often remain, though much diminished in size, and even exhibit appreciable movement. Such movement, however, does not appear to aid in locomotion. Enchelys (Fig. 13) is very like Holophrya (Fig. 14) and only under the most favorable conditions of staining and lighting can the side opening buccal orifice be defi- nitely located. The proboscis-like projection at the an- terior extremity of the body is subject to considerable modification, and can not be relied upon as a distinguish- No. 603] ECOLOGY OF THE PROTOZOA 167 ing feature, but the position of the buccal cavity is virtu- ally constant. Euglena deses (Fig. 15) is recognizable by its enor- mously elongated body, for the propulsion of which its small flagellum seems inadequate. It frequently assumes an almost ameeboid form, and not infrequently contracts to such an extent as to become nearly globular. Being variable in size, its form, when swimming by means of its flagellum, offers the readiest means of identification. One particular locality, given especial notice, deserves detailed mention. This was a spring pool in a boggy up- land meadow, almost hidden among large tufted clumps of luxuriant sedge-like grasses. The depth of the pool was about four inches, and its area about six square feet. Although exposed to the rays of the sun during the greater part of the day, yet the temperature of the water was kept low by a constant trickle of cold water from seepage springs in the glacial clays which underlay several acres of the region. The iron content of the water was unusu- ally high, though the pool was erystal clear. Of higher aquatic plants there were none; Desmids were numerous; Diatoms were sufficient in numbers to form a brown film over the entire bottom. The pool contained seven large black-nosed dace, and other adjacent pools harbored sev- eral more. Snails, Limnea and Physa crawled every- where. With the exception of half a dozen large Hydro- philide there were no aquatic insects, and but very few Entomostraca. The ai forms were Monas fluida and irregu- laris, occurring not alone in the bottom film but every- where throughout the water and in great numbers. One species of Holophrya and Cercomonas crassicauda (but few individuals of each) completed the census of the forms. ExNnvmonmENTAL Type No. IV: Crear Smari Poots with ABUNDANT DecomPosIiNG ORGANIC SEDIMENT Small clear pools in depressions in rocks containing de- caying leaf material and a small quantity of such algæ as 168 THE AMERICAN NATURALIST [Vou. LI Spirogyra, Zygnemea, Mougeotia, ete., Desmids and Dia- toms abundant. Water kept pure and cool by influx from small rills or by seepage from adjacent stream. Tem- perature of water, cir. 60° Predominant forms: Chilodon cucullus, numerou Holostichia — numerous Chilomonas eee stunt Monas sp., abunda Coleps hirtus, numero aramecium bursaria, numerous Diffiugia globulosa, numerous Trachelocerca olor, numerous Holophrya sp., numerous Vorticella asad severa Other forms present in less numbers (few or rare) : meba proteus Euglena viridis Anisonema obliqua Euglypha alveolata Arcella vulgaris Euplotes sp. Aspidisca costata Frontonia sp. Chilodon vorax H Siok dais papilio Chlamydomonas sp. Loxophyllum sp. Colpoda cucullus Nostolenus pie tens Cyphoderia ampulla Oxytrichia bifaria Dallasia frontinia Oxytrichia pellionella Diffiugia acuminata Paramecium caudatum Diffugia pyriformis Pleuronema sp Dileptus gigas Prorodon armatus Dileptus monilatus Stylonychia pustulata Euglena deses Stylonychia putrina Dileptus gigas (Fig. 16) is surely the king of beasts among the ciliate Protozoa. It is entirely carnivorous and | its appetite is apparently insatiable. The prey is stung by the well developed trichocysts which Dileptus bears upon its long ‘‘neck’’ and if too large to be swept into the bugeal cavity by the cilia is forced in by the writhings of ‘neck.’? The creature varies greatly in size, but is normally about 450 microns in length. Individuals have been reported measuring 800 microns! ENVIRONMENTAL Type No. V: Poots CHokep WITH ALG; Water Warm Ditches and pools choked with heavy, luxuriant masses of alge (Spirogyra, Ulothrix, Zygnemea, etc.) and ex- posed to the sun during the entire length of the day. Cy- clops, Canthrocamptus, Gammarus, Daphnia, Cypris, Simocephalus, etc., abundant. In lesser numbers: aquatic No. 603] ECOLOGY OF THE PROTOZOA 169 insects (especially the Hydrophilide), Limnea, Physa, Planorbis, etc. Temperature of water, cir. 70° F. Predominant forms: Cercomonas termo, abundant Monas irregularis, very abundant . Chlamydomonas sp., numerous Peridinium cinctum, numerous Diffiugia globulosa, numerous Synura uvella, numerou Euglena viridis, very abundant Trepomonas = tty very yaad $n Monas fluida, very abundant Other forms present in less numbers (few or rare) : Actinophrys sol Colpoda campyla Arcella sae (?). Cyphoderia ampulla Arcella mitrata Frontonia sp. Arcella easa Vorticella nutans The species which I identified as Arcella discoides (Fig. 17) was much smaller than usual. Although there is con- siderable variation in size among the members of Arcella and particularly among Arcella discoides and vulgaris, yet I have never seen recorded individuals so small. The average diameter of twenty-five individuals from one sample was 70 microns. The Frontonia species (Figs. 18 and 19) are very fully equipped with trichocysts. When irritated with acetic acid and then stained with iodine or methyl green they show beautifully. For use in the laboratory in the demon- stration of the trichocysts of the ciliata they offer the best possible material. The species are not very common, how- ever, and I know of no culture medium. Trepomonas agilis (Fig. 20) varies greatly in size. Its apparent variation in shape ean be explained, I think, by the fact that it swims sometimes with one aspect of the body presented to the observer and sometimes with an- other. The curious irregularity of its body would there- fore allow it to show a number of different outlines. Often it swims in one position for a long period, and again it twirls rapidly about going through the whole gamut of its apparent changes of form in a few seconds. In one typical roadside ditch overhung with grasses.and weeds and literally filled to overflowing with rich masses of Spirogyra I found an almost pure culture of Euglena 170 THE AMERICAN NATURALIST [Vor. LI viridis. The only other forms present were but very few Colpoda campyla and Cyphoderia ampulla. Several samples were taken from leaves and grass frozen together, lying beneath three inches of solidly com- pacted. snow in a small oak grove not far from a ravine through which flowed a perennial stream. This material was allowed to stand in its own snow in a cotton-tamponed jar until the snow had melted and the resulting water had attained room temperature, cir. 72° F. The Ameba proteus were more abundant than I had ever seen them before except in artificial cultures. Not infrequently as many as ten individuals could be counted at once in the field of the 10 mm. objective. The other predominant forms were: Holostichia vernalis, abundant. Opalina ranarum, several Monas irregularis, abundant. Paramæcium caudatum, several Other forms present in less numbers (few or rare) : Astasia lagenula j Paramæcium trichium Oikomonas sp, Platyrichotus opisthobolus CATALOGUE OF SPECIES NOTED WITH MEASUREMENTS IN Microns In each case length refers to the antero- Jito axis, except with globular or subglobular forms when it refers to the diameter of the body. All dimensions are given in microns (= 1/1000 mm.). It is interesting to recall, in connection with the sizes of these organisms, that the diameter of the sl eh human hair is 3 100 microns. ae a Species Length Tinea? AA sol ; : 60 Ameba LORS “C Vimax 50—60 Amæba ' proteus 125 (variable) Amæba : _ radiosa 50-100 (variable) Anisonema obliqua Arcella artocrea ` 160 Arcella i dentata Ye OB, Arcella discoides 115 Arcella | mitrata 100-150 Arcella ` vulgaris 55 (variable) Aspidisca costata BD fess No. 603] = contorta lagenula s vagans lateritia polypinum aculeata crassicauda termo cucullus vorax sp. paramæcium sp. umbellatum hirtus sp. striatum campyla cucullus inflata maratima pyriformis igas giga monilatus vorax viridis . alveolata carinata charon grandinella myriapoda sp. ECOLOGY OF THE PROTOZOA 171 30 (variable) 20 75 50 (variable) 170 60 (bell only) 125-150 25 85 75 25 (variable, body only) 150 | 100-300 150-250 20-50 90-120 250-350 450 (variable) 18 (variable) 50 50 50-200 30 (variable) (variable) 172 Holostichia Hyalosphenia rymaria Oxytrichia Oxytrichia Paramecium Paramecium Paramecium acus Platyrichotus Pleuronema Prorodon Prorodon Prorodon Stentor tylonychia Stylonychia Stylonychia Synura Trachelocerca vernalis papilio wrzesmosky sp. sp. sp. (repetans?) Dallingeri fluida irregularis orbicularis trichium sp. opisthobolus sp. my pustulata THE AMERICAN NATURALIST and 5 0 100-400 100-300 80-150 80-150 50 (variable) 3 70-130 10 [Vou. LI THE ROLE OF ISOLATION IN THE FORMATION OF A NARROWLY LOCALIZED RACE OF DEER-MICE (PEROMYSCUS).* DR, F. B. SUMNER SCRIPPS Institution, La JOLLA, CALIF. No one who has critically examined large numbers of specimens, belonging to such a widely distributed and diversified genus as Peromyscus, can fail to be impressed with two facts. First, the differences upon which the so-called ‘‘subspecies’’ are based are real and obvious ones. But, secondly, the actual subspecies which are recognized and named are necessarily highly artificial groups. On the one hand, each subspecies intergrades with others to such an extent that the assignment of a given specimen to one or the other group is often quite arbitrary. And on the other hand, even these ‘‘sub- species’’ themselves are far from being elementary. They are composite groups, comprising, in many cases, a num- ber—perhaps a great number—of distinguishable local types. The word distinguishable is here used in a quali- fied sense. It is likely that the distinctions would com- monly be obvious just in proportion as the collections were made at points which were remote from one another. Indeed, it has been said by one who has monographed this genus of mice’ that ‘‘classification becomes . . . like dividing the spectrum and depends largely upon the standards set, for, theoretically at least, the possibilities of subdivision are unlimited (p. 17).’’ None the less, it is generally believed that where well- marked physical or other barriers are interposed between two groups of individuals, this continuous intergradation *Read before Ecological Society of America, San Diego meeting, August, 1916, 1 Osgood, ‘‘ Revision of the Mice of the American Genus, to sua cual North American Fauna, No. 28, Washington, 1909 173 174 THE AMERICAN NATURALIST [Vou. LI of racial characters may be largely interrupted. It is the object of this paper to discuss a ease of this sort which I have had the opportunity of studying during the past year. The subspecies Peromyscus maniculatus rubidus, ac- cording to Osgood,? who first described it, occupies a strip of varying width on the ‘‘coast of California and Oregon from San Francisco Bay to the mouth of the Columbia River.’’ In discussing certain local variations shown by this subspecies throughout its range, the same writer states that ‘‘six specimens from the Outer Peninsula, near Samoa, Humboldt Bay, are decidedly paler than others from the neighboring redwoods. They evidently repre- sent an incipient and very local subspecies, and well illus- trate the plasticity of the group to which they belong.’’ Osgood further remarks that ‘‘a careful study of this variation and the local conditions doubtless would prove instructive’’ (p. 66). During the latter part of May, 1916, I trapped on two consecutive nights in the neighborhood from which Os- good obtained his six ‘‘aberrant’’ specimens of rudibus.® About one hundred live-traps were set on each occasion. Twenty-eight specimens were taken, of which twenty-one were later available for skinning and for careful meas- urement. These last were all in either mature or adoles- cent pelage, and were about evenly divided in respect to sex. The distinctness of this race from the rubidus of the redwood forests on the mainland was evident from a casual inspection of the living mice. A more careful com- parison of freshly killed specimens from the two locali- ties, and later of their prepared skins, justifies the fol- lowing generalizations. These impressions were formed independently by several other persons to whom I showed ‘he specimens, and were confirmed by more careful ex- _ 2 Loe. cit., p. 65. 3 The trapping was done between one and two miles northwest of the village of Samoa. Besides these Peromyscus, the only other animal caught was a single specimen of Microtus. No. 603] THE ROLE OF ISOLATION 175 amination and measurement. (1) The Samoa lot, as a whole, were paler than the redwood lot; (2) the tails of the former were shorter, and (3) the ears were longer. To consider first the coat color, the mean difference be- tween the two series of skins is evident at a glance. Like- wise, it is plain that the palest Samoa specimen is paler than the palest Eureka (redwood) specimen, and that the darkest among the former is paler than the darkest among the latter. It must be admitted, however, that the two series overlap rather broadly,* the darker skins. of the Samoa stock being as dark as or darker than the paler ones of the Eureka stock. An attempt to express the color of a mammal’s pelage in terms of any set of ‘‘standard’’ colors is beset with great difficulties. Instead of a uniformly tinted, plane surface, we have to do with a mixture of variously colored hairs, further diversified by minute shadows and reflec- tions. I have, nevertheless, endeavored, in a rough way, to ‘‘match’’ the colors of these two races with those of Ridgeway’s ‘‘Color Standards and Color Nomencela- ture.” In the Samoa race, the general tone of the lateral regions of the body lies between the ‘‘tawny olive” and ‘‘Saccardo’s umber,” that of the dorsal darker stripe being not far from ‘‘sepia.’? In the Eureka mice, the lateral regions range from ‘‘Saccardo’s umber’’ to ‘‘sepia,’’ the dorsal stripe being of a depth somewhere between ‘‘sepia’’ and black. These comparisons will ‘at least enable the reader to judge of. the degree of differ- ence between the two racs.® As regards the tail, it was plain without measurement that the average length of this member was greater in the +I have at present for reference twenty-one skins of the Samoa lot and thirty skins of wild adults from the redwoods. Ten of the latter indi- viduals were trapped and skinned at about the same time as the former, so that the factor of season may be disregarded. 5 Washington, 1912. Published by the author. 6 In my further studies of Peromyscus I plan to employ two revolving color-wheels, on one of which the skin itself will be rotated, on the other sectors of black, white and various primary colors. This apparatus is now being tested by Mr. H. H. Collins and myself. 176 THE AMERICAN NATURALIST [Vou. LI Eureka than in the Samoa race, though here again the difference related to averages and did not hold for all individual cases. A comparison of the mean figures for absolute tail length in two series of mice is not entirely justifiable, par- ticularly if the two lots of individuals differ somewhat in mean body size. But the relative tail lengths (ex- pressed as percentages of body-length) may be fairly compared, since there is good evidence that these ratios remain nearly constant after the first few months of life. The following table allows of a comparison between the two races, in respect to this character: Number of Mean Standard Cases (Percentage) Deviation Eureka (males) 83 104.39 + 0.387 | 4.95 Eureka (females) 53 103.60 + 0.54 | 5.85 sexes combined ) 21 97.48 +0.94 | 6.38 The differences between the Samoa lot ee com- bined) and the Eureka males and females are 6.91 per cent. and 6.12 per cent., respectively. These differences are about seven and six times their probable errors, re- spectively. Their significance may therefore be regarded as fairly certain, despite the small numbers comprised in the Samoa series. As regards foot-length, the two races do not differ significantly. But the ear, as already stated, is appre- ciably longer in the Samoa mice, this difference being perceptible, even without measurement. Here, as in the ease of tail-length, a simple comparison of gross averages for the two groups would be unjustifiable. But in the present instance, the conversion of the absolute values into percentages of body-length would be equally unjus- tifiable, since the growth of the ear is not at all propor- tionate to that of the body as a whole. We must there- fore resort to the method of ‘‘size groups,’’ i. e., we must divide each of our two lots of animals into small groups comprising individuals of nearly equal size, In the case at hand, we have fifteen groups, or othen No. 603] THE ROLE OF ISOLATION 177 pairs of groups, within which a comparison of average ear-length is possible. In twelve cases the mean figure is greater for the Samoa mice, in two cases it is greater for the Eureka mice, while in one case the two figures do not differ appreciably. The probabilities against such a pre- ponderance being due to chance are of course high. The mean difference in ear-length between the two lots, com- puted according to a method described by me in an earlier paper,’ is 0.87 mm. Those who have made careful meas- urements of mice will regard such a difference in the length of this appendage as far from trivial. Let me now say something as to the environmental con- ditions under which these two races of rubidus live. Those which I have designated as the ‘‘ Eureka”’ or ‘‘red- wood’’ race were trapped by me during two different years, within a distance of two miles from the southern limits of the city of Eureka, California. The region is one covered in large part by redwood forest, most of which is of second growth, although there are some small areas that have never been logged. The predominant tree is the redwood (Sequoia sempervirens), but several other conifers are common, the most abundant of these being the Sitka spruce (Picea sitchensis), Douglas fir (Pseu- dotsuga taxifolia), and lowland fir (Abies grandis). The red alder (Alnus rubra), cascara (Rhamnus purshiana), waxberry (Myrica californica), red elderberry (Sam- bucus racemosa), and a willow (Salix hookeriana) appear to be the chief non-coniferous trees of this district.6 The “wild lilac’’ (Ceanothus thyrsiflorus) is likewise com- mon in some of the more open areas, often reaching the proportions of a small tree. Except in recently cleared tracts, the region is one of dense underbrush, the shrubbery and vines forming, in fact, a veritable jungle which is frequently hard to pene- T Journal of Experimental Zoology, Vol. 18, April, 1915, particularly, pp. 341 et seq. : 8 For the determination of many of the plants referred to in this paper I am indebted to Professor H. M. Hall, of the University of California, and to Mr. J. P. Tracy, of Eureka. 178 THE AMERICAN NATURALIST [Von LI trate. Here we meet with the thimble-berry (Rubus par- viflorus var. velutinus), the salmon-berry (Rubus spec- tabilis var. menziesii), huckleberry (Vaccinium ovatum), red bilberry (V. parvifolium), salal (Gaultheria shallon), and in the more open areas the blackberry (Rubus viti- folia): Two ferns (Aspidium munitum and Pteris aqui- lina) are extremely abundant, the latter in particular forming dense growths higher than a man’s head. Inthe . op Lp Wy Ce “uy, : p % m o Uy y samog *, yy by, ~ Wie go a ee Vi if 7 Ae ; HUMBOLDT ae ie Eas YW, ce tte es ma X Vi, a ns BS ne ) Wy ie Hy Oss > Fie. 1. Map of the vicinity of Humboldt Bay, California, kd upon J. N. Lentell’s map of Humboldt County. The three principal trapping sta- tions are designated by the letter T. Area occupied by redwood forests is indicated by oblique shading. NS more open areas a tall annual of the evening primrose family (Epilobium angustifolium) constitutes an im- portant element in the vegetation. One coming from the more arid parts of California can No. 603] THE ROLE OF ISOLATION 179 not fail to be impressed by the prevailing humidity of both soil and atmosphere in this region. In the dense shade of the great redwoods the ground is damp, even during the summer months, and the fallen logs are coy- ered with mosses and fungi. When we cross Humboldt Bay to the narrow peninsulas separating this body of water from the ocean (Fig. 1), we enter a quite different environment. No redwoods are found, the woods, where present, are open, and the ground is prevailingly dry and sandy. In the wooded area, ex- tending down the axis of the northern peninsula, the pre- dominant tree is a small pine (Pinus contorta), though the waxberry and willow (Salix hookeriana) are likewise abundant, and small specimens of the Sitka spruce are fairly common. Among the more frequent shrubs are the huckleberry (V. ovatum), the twinberry (Lonicera involucrata) and silk tassel bush (Garrya elliptica). The ground is largely covered by two plants of trailing habit, the bearberry (Arctostaphylos uva-ursi) and the beach strawberry (Fragaria chilensis). On its ocean side, the peninsula is bordered by a wide strip of shifting sand. Here the process of dune forma- tion may be witnessed to perfection, the dunes often reaching a height of forty or fifty feet. In places the en- croachments of the sand upon the hard-pressed vegeta- tion are evidently rapid, solid ramparts of willows and spruces being steadily engulfed by an advancing wall, frequently as high as the trees themselves. Nevertheless, even on the open sands of the dunes, certain trailing plants maintain a precarious foothold. Among the commonest of these are to be mentioned the yellow sand verbena (Abronia latifolia), the beach strawberry (F. chilensis), beach pea (Lathyrus littoralis), and two species of Fran- seria (F. chamissonis and F. bipinnatifida), while the suc- culent Mesembryanthemum aequilaterale is occasionally met with. Despite the nearness to the ocean and the high atmos- pheric humidity, the peninsula region seems dry in com- parison with the redwood forests. This is due in part to 180 THE AMERICAN NATURALIST [Vou. LI the loose, sandy character of the soil—where, indeed, any real soil exists—and to the comparative lack of shelter from the prevailing westerly winds. Evaporation here is doubtless more rapid than in the comparatively stagnant air of the forests. To my surprise, the footprints of mice and other small mammals were abundant, even on the shifting sands, in the areas of sparsest vegetation. Since these tracks, for the most part, were effaced every day by the wind, the animals must have been present in large numbers. In- deed, it was in or close to the dune region that I trapped most of the twenty-eight Peromyscus. It seems more than possible, therefore, that the predominantly paler shade of the mice dwelling here may be due to the same causes which are operative in producing the yet paler hues of many of the desert rodents. | What the effective factors are can not yet be stated with certainty in either case. Protective coloration is of course an obvious explanation, but it is one of doubtful applicability in the case of animals which are almost wholly nocturnal in their habits. For this and other reasons it seems more likely that the pale coloration of these mice stands in some more direct relation to the humidity of their immediate surroundings. That it is not, however, a strictly ‘‘somatic’’ phenomenon, called forth anew in each generation, I have already shown for the desert race, P. m. sonoriensis.® Whether or not the peculiar color of the pelage in the Samoa race is likewise hereditary I have endeavored to test experimentally. Seven living females and a number of males were brought to La Jolla in June, 1916. Unfor- tunately, it was not possible to obtain more than two broods of young, comprising three individuals, one male and two females. These animals were carefully examined at the age of five months, in comparison with over forty individuals, derived from the redwood stock, which were 9 AMERICAN Naruratist, Nov., 1915. I have since reared this race in Berkeley as far as the third (in one instance the fourth) cage-born genera- tion, without any certain modification in color. No. 603] THE ROLE OF ISOLATION 181 mainly of the same age or older, and likewise reared from birth at La Jolla. Not a single individual of the latter stock was as pale as either of the two females of Samoa parentage. The male of the Samoa race was, however, of about the average shade of the redwood descendants. As stated above, some of the wild parents, trapped on the peninsula, were likewise as dark as many of the redwood series. No certain conclusions can, of course, be based upon these three individuals. But the condition of the two females certainly lends support to the belief that the peculiar coat color of the Samoa race, however it was ac- quired, has become fixed germinally. Reference to the map shows that the northern penin- sula of Humboldt Bay is largely isolated, so far as land- living rodents are concerned. In addition to the ocean and the bay, a marshy tract extends from the latter to the Mad River, which, in turn, interposes a further barrier on the north, and nearly converts the peninsula into an island. Beyond the mouth of Mad River, this same type of sand-dune formation extends uninterruptedly to the mouth of Little River, about six miles to the north, where it ends abruptly and the shore line becomes precipitous. Now this northward extension of the sand-dune region is not isolated by any physical barrier from the redwood forest, which here comes near to the coast. It occurred to me, therefore, to attempt the collection of Peromyscus from a point somewhere within this region. The locality chosen was close to the ocean, about two miles south of Little River and four to five miles north of Mad River. Here the conditions were found to be closely similar to those on the exposed side of the northern peninsula of Humboldt Bay. The dunes were on the whole lower, how- ever, and some minor differences were noted in the flora. The belt of shifting sand here ranges from five or six hundred feet to perhaps a fourth of a mile, giving place on the landward side to a narrow meadow or marshy area, sueceeded by a high, steep, wooded ridge. About ninety traps were set on two consecutive nights, 182 THE AMERICAN NATURALIST [Vot LI yielding in all forty-eight Peromyscus, all belonging to the subspecies rubidus. Many of these were still in juvenile pelage and such individuals were kept and al- lowed to mature in captivity. A hasty comparison of the living Little River animals (as I shall call them) with those from the Samoa and Eureka trapping grounds made it plain that, in respect to color, they belonged with the latter group rather than the former. Careful comparisons of series of dead mice and of skins were made later and the bodies were sub- jected to the customary measurements. Owing to numer- ous deaths, however, only twenty-eight individuals were available for these purposes. This more critical examination confirmed my earlier ‘belief that the Little River mice agreed pretty closely, in average color, with the redwood stock, but that they dif- fered widely from those taken on the peninsula. It ‘seemed probable, however, that the mean shade was ‘slightly lighter than that of the former animals, making ‘them, to this ‘extent, intermediate. One conclusion then seemed plain. The peninsula race, exposed to certain modifying conditions, was enabled to differentiate from the mainland stock, owing to the almost insuperable barriers to migration. The Little River stock, exposed to practically the same conditions, have not formed a distinguishable race, because the rate of dif- ferentiation has been far exceeded by the rate of diffusion, or intermingling with the great body of more typical “rubidus,” dwelling in the redwood forests which extend back from the coast. We might seem to have, therefore, a particularly clear cut example of the effectiveness of isolation in the formation of a local race. Now, I am not yet prepared to admit that such con- clusions would be groundless. But here, as so often hap- pens, a further study of the data has shown that the prob- lem is more complex than was at first suspected. It is true that the mice of the more northern sand dunes have not formed a distinct race as regards color. But it is none the less certain that they differ from those of the No. 603] THE ROLE OF ISOLATION 183 Eureka region in regard to both the length of the tail and that of the ear. In respect to the former character, they agree pretty closely with the Samoa race, the difference from the redwood stock being statistically even more cer- tain in this case. To still further complicate the situation, we find that the ear, instead of being longer, is shorter than that of the redwood mice by about half a millimeter, and thus averages about one and one half millimeters shorter than in the peninsula race. Here, too, the dif- ferences are even more certain statistically than those which distinguish the Eureka and Samoa series. The numbers are small, of course, only twenty-eight of the Little River mice having been available for meas- urement. But as regards tail length, the difference be- tween the averages is seven to nine!’ times its probable error, so that the likelihood of its being due to random sampling is very small. Have we, then, here merely another example of incon- clusive data, which might best have been left unpublished? I do not think so. The mere existence of these local dif- ferences in color and in the size of parts deserves careful description, whatever interpretation we may place upon them. Moreover, I am disposed to a that the case of coat color is not entirely comparable with that of the length of the appendages. In another article’! I have given reasons for thinking that some of the differences in the former may have arisen in nature as more or less direct effects of environmental conditions. On the other hand, I have shown that such an explanation would be of very difficult application as regards some of the measurable differences in the parts of the body, even though the latter are known to be readily influenced by various experi- mental agencies. Now the evidence at hand is sufficient to show that any environmentally produced modifications of coat color are 10 Depending on whether the comparison is made with the Eureka males or females, the sexes being combined in the case of the Little ot tet group. 11 AMERICAN NATURALIST, Nov., 1915. 184 THE AMERICAN NATURALIST [Vou. LI at best rather gradual. Rubidus remains rubidus and sonoriensis remains sonoriensis, after several generations of captivity in changed climates. But even the first cage- born generation of each of my subspecies is found to be highly modified by confinement, in respect to the mean length of: certain of the appendages. That this somatic plasticity would be accompanied by a high degree of germinal instability, as regards these parts, could not, of course, be predicted in advance. But the frequent ap- pearance of local differences of type renders it probable that this is true. Whether or not these local peculiarities are due in some indirect way to environmental factors, or whether they are due to ‘‘spontaneous’’ mutation, need not concern us here. The main point to bear in mind is the probability that the pelage color is somewhat more stable in these mice than are the bodily proportions, despite the fact that it is the former, rather than the latter, which gives the clearest evidence of a definite cor- relation with known factors of the environment. For the reason just stated, it is possible that the dif- ferentiation of a new color race might require fairly rigid isolation; whereas local differences in some of the measurable parts might arise in the presence of no other barrier than the naturally slow rate of diffusion of a non- migratory animal. As was remarked earlier, we have reason to suppose that representative collections from an indefinite number of localities would reveal the exist- ence of statistically certain differences between the mice of many of these localities. In most cases, it would prob- ably be unjustifiable to assign these series to distinct races, or other definite taxonomic groups, since it is likely that — perfect intergradation would be found between most of them, and that the degree of difference would be largely a function of the distance apart of their respective habi- tats. These last remarks are, of course, largely conjectural. Part of the author’s present program consists in a care- ful study of local differences of the sort here discussed. No. 603] - THE ROLE OF ISOLATION 185 It is hoped that this will render possible more definite answers to some of these difficult questions. It seems to be held by certain zoologists that any dis- cernible difference between two local types, if at all con- stant, ought to be in some way recognized in the nomen- clature. Indeed, I have been advised to name fhis modi- fied race of rubidus from the northern peninsula of Hum- boldt Bay. Such a practise, if carried out consistently, would lead either to an endless multiplication of sub- species, or else to the introduction of quadrinomial names. Hither procedure would, I think, be deplorable. The actual needs of the situation can commonly be met, I believe, by stating the locality from which a given speci- men or collection was taken. The bestowing of formal names creates the false impression of a multitude of well- defined entities which do not, in reality, exist. Moreover, it is my firm conviction that nomenclature should have for its object the recognition of resemblances as well as the recognition of differences. The first of these func- tions is all too frequently overlooked. SHORTER ARTICLES AND DISCUSSION THE MIGRATION OF FISHES Unber the head of ‘‘The Migrations of Fish,’’ Professor Alex- ander Meek has given a voluminous account of what is known of the movements and the distribution of the various families of fishes. The work is illustrated with drawings and photographs of many species, showing not only their forms and their move- ments, but often the stages of development and the structure of fins and scales. Especially valuable is a series of maps showing the geographical distribution of interesting groups. The word migration is taken in its largest sense, including not merely move- ments of individuals or of masses, but the larger problems of distribution, extending often over geological periods. It is plain that distribution is intimately related to migration and migration to currents. As the problems of fish conservation depend directly on the facts of migration and distribution, especial attention is given to the development and movements of food fishes: and naturally to those of the North Atlantic. After a general discussion of the continental and oceanic changes which have taken place since Eocene times, these having a direct bearing on modern conditions of fish-distribution, Pro- fessor Meek takes up the various groups of fishes, beginning with the lowest, treating of the habits, movements and distribution of each group in turn. The excellent account of the lampreys and hag-fishes shows a certain omission. While the lampreys fasten themselves to other river fishes, sturgeons, catfishes and the like, rasping great holes with their teeth, the hag-fishes attack the throats of large sea- fishes, entering the muscular system and almost destroying it before the fish concerned finally dies. Around Monterey Bay, various flounders and rock-fishes (Sebastichthys) are thus at- tacked and drift about as living hulks while the hag-fish (Polisto- trema) devours their muscular tissues The interesting parallelism in habits and distribution of the 1*¢The Migrations of Fish,’’ by Alexander Meek, M.Sc., professor of zoology, Armstrong College in the University of Durham, and director of the Dove Marine Laboratory, Cullerecoats. Edward Arnold, London, Long- mans, Green & Co., New York. Price $4.50. 186 No. 603] SHORTER ARTICLES AND DISCUSSION 187 sturgeon and the lamprey is noted by Professor Meek and in both eases the facts now observed are of long standing. Both had ap- parently ‘‘refuge-regions’’ during glacial times. We do not, how- ever, see the reason for the suggestion that the green sturgeon of California (now almost extinct) ‘‘may have been derived from the Atlantic during the post-glacial disturbance.’’ The migrations and breeding habits of the herring are treated with special fullness, proportionate to the economic value of the species. For in the north, as Björnson informs us, wherever a herring school touches the coast a town springs up, like drift- wood on the beach. In the account of the trout, very good as a whole, we may note that the genus Salmo is represented by different forms, originally derived from the Pacific coast, in the Great Basin of Utah, and also in the headwaters of the Colorado, Rio Grande, Arkansas and Platte, as well as the Columbia and Missouri. It is probable that the freshwater irideus (Rainbow trout) and the sea-run gairdneri (steelhead) are not really different, but both are quite separate from the cutthroat trout (Salmo clarki, wrongly identi- fied at first by American authors with the Kamchatkan Salmo mykiss) from the Tahoe trout (Salmo henshawi) and from the several local forms which have sprung from these or which have preceded their advent. The last seems to be the case with the silver trout of Lake Tahoe (Salmo regalis). The suggestion of Professor Meek that the European salmon (Salmo salar) of the Miocene was divided into a North Atlantic and a Mediterranean form is interesting. The latter developed as a ‘‘trout’’ dividing into sea trout (Salmo trutta) and burn (or brook) trout (Salmo fario). But these two are as yet not really differentiated, corre- sponding in a way to the rainbows and steelheads of the Pacific coast. Professor Meek says: There is thus good reason for believing that the sea trout and the common trout may be the same, the one retaining the migratory habit and the other confining itself to fresh water. Our own experience with the species lends probability to this view. That ‘‘the salmon preceded the trouts’’ in time is also probable, but the western species of trout must have been derived from the trout of Europe and Asia. The Pacific salmon must be older and more primitive than the Atlantic salmon, for the six species differ from one another, more than any trout or even the Atlantic salmon itself differs from any other black-spotted trout whatever. 188 THE AMERICAN NATURALIST [Vou. LI Professor Meek hardly does justice to the spawning habits of the red or blueback salmon. It runs up rivers to varying dis- tances—-from one mile to 1,500 miles (Lake Labarge on the Yukon). But it never enters a stream which does not flow from a lake and it spawns always in the small streams at the head of the lake. At Boca de Quadra in Alaska, the small stroni is barely. a mile long. It comes from a clear like, perhaps five miles long. Into this stream and lake the salmon crowd by the thousands. The Yukon is not a good red salmon stream, because the nearest tribu- tary lake, Labarge, is about 1,500 miles from the sea. Yet red salmon enter the river and reach the lake. In streams without lakes as the Skagway, red salmon are never seen. The King salmon (Oncorhynchus tschawytscha) also runs for great dis- tances, but it is absolutely indifferent to the presence of lakes. It is probable that the red salmon spend their first winter in the lake and some never leave it, remain landlocked and dwarf until spawning time (usually four years). One of the most difficult of problems is to understand the in- stinct of the red salmon. Every individual of this and of each of the other species of Pacific salmon (Oncorhynchus) dies after spawning. How does the spawning fish, stupid in most regards, know when it enters a river that there is a lake before it? How does it come to avoid all lakeless tributaries as it goes up, finally reaching the lake’s head and the brooks that feed it? And why do the other salmon species totally lack this instinct? There are other problems, yet unsettled, regarding the supposed homing instincts of salmon. The majority (but not all) seem to return to spawn to the parent stream which they left as fingerlings. Why not all? And why any? The recognition of the age of salmon and trout by the adjust- ment of the rings on the scales, as recently worked out by Dr. C. H. Gilbert and others, received full attention from Professor Meek. The scales of the salmon are marked by concentric rings of growth, and these are more widely separated in the summer, the Peas time of the salmon when the individual grows most rapi Profit Meek devotes much space to the singular breeding habits of the eel, which spawns in the.sea, entering rivers to feed. But many individuals, in our Mississippi Valley never descend to the sea. A very large eel, once taken by the present reviewer, above the Cumberland Falls in Kentucky, 2,000 miles from the No. 603] SHORTER ARTICLES AND DISCUSSION 189 sea, must either have never spawned or cast its spawn into the river. The larve of eels as well as of some other soft-rayed species are quite pellucid, without pigment cells and with ‘‘a roomy space between the skin and the muscles, distended by a watery fluid.’’ Many of these larve, in their transformation to the con- dition of young fishes become much reduced in size, though in- creasing in weight, by the obliteration of these interspaces. Professor Meek’s studies pass through the whole long series of fish-families. For want of space, we may not follow them further in these pages. We must give the work, as a whole, very high praise as carefully, intelligently and scientifically done, and as constituting a reference book of great value. The author has well covered the range of the periodicals which treat of the dis- tribution and habits of fishes. He seems, however, to have over- looked the most extensive recent work of a similar range, ‘‘ Jor- dan’s Guide to the Study of Fishes,’’ published in New York in 1905. DAVID STARR JORDAN NEW LIGHT ON BLENDING AND MENDELIAN INHERITANCE UnperR the above heading, Dr. Castle reviews a paper by Yuzo Hoshino on the inheritance of the flowering time in peas., an rice. Since reading this review, Prof. Hoshino kindly sent us his paper, and we have ourselves examined it with care to see whether indeed it necessitates Dr. Castle’s rather sweeping con- clusions, namely, that certain genes are themselves modified by crossbreeding, one of the conclusions of Hoshino himself, and that selection within a pure line, within a genotypically pure population is effective. It is well known that Dr. Castle counts among the few last geneticians, who still believe that the genes themselves are modi- fiable by selection. Hitherto in nearly all his writings on the subject Dr. Castle claimed, that unit characters vary, and may be modified by selection, a statement which can not very well be opposed, given the loose way in which the obsolete term unit character is usually applied. But it was clear, that Dr. Castle really believed the genes themselves to be capable of variability in potency, quality and value, and we think it of the utmost im- portance that in the review under discussion he has stated the 190 THE AMERICAN NATURALIST [ Vou. LI question in these words. Thus the issue between Dr. Castle on one side, and Johannsen and us on the other narrows, and there need be no more difficulty as to the exact meaning of the term unit character. As to the effectiveness of selection in genotypi- eally homogeneous material, all the evidence so far adduced shows that selection in such material is absolutely ineffective. It is evident that selection in a population is usually effective, but this only shows that in ordinary populations, even in so-called pure strains of animals, there is a good deal of genotypic varia- tion, or in other words impurity. The fact, for instance, that Dr. Castle’s selection in hooded rats was effective, shows that his material was not originally pure for all the genes. In all those instances where the guar- antee for genotypic purity of the material was reasonably good, selection has, until now, proved ineffective. We need only point to the fifty years of selection in wheats by the de Vilmorin fam- ily, and to the numerous selection-experiments with clones of Paramecium by Jennings and others. As to the so-called instances of the effectiveness of selection on the genes themselves in alleged genotypically homogeneous animal material, we repeat that the only way to show such an effect in material which offers no sure proof of purity would be to change a strain of severely inbred animals by selection to a point removed from the range of the ordinary modification in the material, continuing the inbreeding, and then, by contraselec- tion, to bring the character under consideration back to its start- ing-point. Since we wrote down this challenge to the believers in the variability of genes, one such a series of selection-experi- ments has been performed, namely, on flies, and in this series it has been proved to be impossible to get the material back to its original quality. According to Dr. Castle, Hoshino’s Table 6 shows the effect of selection within a pure line. In the cases taken from Hoshino’s paper, in which the progeny of an early-flowering and a late- flowering individual of the same ‘‘pure line’’ can be compared (in the original table there is one more case in which the earliest parent gives the latest progeny) there are more instances in which an early parent gives an earlier progeny than a late parent, than cases in which an earlier parent gives a later progeny. But if we examine the figures more closely, we observe that the mean deviation of offspring from parents in the case in which the earlier parent gives the earlier progeny is 0.52 day, No. 603] SHORTER ARTICLES AND DISCUSSION 191 whereas the mean deviation of offspring from parent in cases in which the earlier parent gives the later progeny is larger, 0.65 day. The only conclusion from Hoshino’s Table 6 is the one he makes himself, namely, that the variation is insignificant. We are sure that no unbiased person would conclude from the negative facts in the table in question that the variation in these pure lines was genotypic, or that selection in these groups has had an appreciable effect. On page 332 Dr. Castle writes: If I have correctly interpreted Hoshino’s observations, flowering time in peas is clearly a Mendelian unit character, entirely devoid of domi- nance, so that a strictly intermediate hybrid form is the commonest end product of a single cross between early and late varieties. Indeed, if Hoshino’s work on the inheritance of flowering-time of peas were the only, or the first, or the most comprehensive of its kind, we could see reasons for such a belief. But Hoshino only crossed two varieties differing in time of flowering. But in peas there do not exist only one late and one early variety, but several thousands, each with its own time of flowering. It would not be difficult to give a list of ten names of pea-varieties of which in every preceding one all the plants would be in flower before one of the next opened its first flower. Therefore crossing experiments involving two varieties can never be sufficient basis from which to conclude that flowering-time in peas is one thing or the other. Tschernack, in his well-known experiments with flowering-time of peas (1911, Mendel’s Festchrift), cited by Hoshino, made eight different variety-crosses. Whereas in Hoshino’s work the two varieties crossed happened to be of such a constitution, that in the resulting F, generation there did not occur plants which commenced flowering at an earlier time than the earliest parent, or at a later date than the latest parent, in Tschernack’s work such cases were met with. In Tschernack’s experiment No. 81 (1906) there were in F, found plants flowering seven days earlier than the early parent; in experiment No. 82 (1916) even plants beginning flowering nine days earlier than the early parent. In experiment No. 81 (1906) there were also found plants starting to flower four days after the latest parent, and i in experiment 38 (1902) there were plants, which did not begin to flower before the late parent had been in flower for a week. It is perfectly clear, that a sort of blending may be the result of a difference between the parents in a number of genes, in- fluencing the quality under observation in different directions. 192 THE AMERICAN NATURALIST (Vor. LI On page 333 Dr. Castle writes: In typical blending inheritance the determiners of contrasted parental conditions apparently blend into a determiner of intermediate char- acter, the gametes formed by an F, individual being practically as uni- form in character as those of either parent. Blending is illustrated in the inheritance of ordinary size-differences in birds and animals. No one who knows the work of Punnett and Bailey (cited in Hoshino’s paper) on chickens, in which they found not only in- dividuals in F, as small as the smallest parent and as large as the largest, but even individuals lighter than the lightest parent and heavier than the heaviest, could maintain that ordinary size- inheritance in birds is blending. The gametes formed by the Hamburg X Sebright hybrids, or by our Leghorn fighting bantam certainly were not as uniform as those of any of the four parental strains! We are perfectly in accord with Castle when he reasons that if once we admit a contamination of genes and qualitative changes in genes, we do not need to assume that flowering-time in peas is influenced by two genes, in the cases studied by Hoshino. In such a case the difference in one gene would suffice. Indeed, we would go one step farther than Castle and declare, that, on the assumption of qualitative changes in genes, we need not assume a genotypic difference between the parent varieties at all. Where we differ from Dr. Castle is in the fact that we do not believe in qualitative variation of genes. Surely more than ten genes must influence the beginning of flowering in the pea, else there could not be so many varieties differing in the time of flowering. All the genes which influence stature, shape of flowering axis, color, must necessarily influence the onset of flowering. And we need not look for coupling between color factors and flowering-time factors, because the factors influencing color influence the metabolism of the whole plant, and thus the period at which it starts flowering. If we compare Hoshino’s paper with Tschernack’s extensive ‘experiments on the subject, we find nothing in it, which would e us assume contamination of genes by crossbreeding, or any qualitative variability of genes. A. C. HAGEDOORN, A. L. HAGEDOORN NAGASAKI, December 19, 1916 THE AMERICAN NATURALIST Vou. LI. April, 1917 No. 604 THE SOURCES OF ANATOMICAL LITERATURE PROFESSOR ROY L. MOODIE DEPARTMENT OF ANATOMY, UNIVERSITY OF ILLINOIS, CHICAGO Tue study of anatomical literature has not received the attention that has been given the writings of men in other lines of intellectual endeavor. When we compare, for instance, our knowledge of the literature of anatomy, and the men who have made this literature, with the work that has been done on the history of poetry and the poets, or fiction, or the history of nations, we see how greatly the development of anatomical knowledge and literature has been neglected. Locy! especially has shown us how this field of study may be used as a field of research in early human documents relating to anatomy. ‘The subject has been further developed by Stieda, Holl,? Sudhoff,? Fors- ter, McMurrich, Roth,* Téply? and others who have con- tributed sundry studies along these lines. That there has been no great amount of attention paid to the subject is probably due to the fact that the subject matter of anat- 1 Journal of Morphology, Vol. 22, pp. 945-988, 1911. 2 Archiv fiir Anatomie und Physiologie, Anat. Abth., Jahrgang, 1905, . 96. 3 Karl Sudhoff is editor of the Archiv fiir die Geschichte der Medizin, to which he is an active contributor (especially to be noted are his original con- tributions concerning medieval anatomical knowledge) and he is also the editor of Pagel’s ‘‘Einfiihrung in die Geschichte der Medizin.’’ 4 Archiv fiir Anatomie und Physiologie, Anat. Abth., 1904, pp. 372-384. 5 Medical Library and Historical Journal, Vol. 4, pp. 338-350, 1906. — fiir Anatomie und Physiologie, Anat. Abth., 1905, p. 79; 1906, if T Anatomische Hefte, Bd. 25, Erste Abth., pp. 351-398, 1904. 193 194 THE. AMERICAN NATURALIST [ Vor. LI omy has been and is of more interest than the form in which it is presented. There have been a number of classical studies in the history of anatomy, the latest and best of which is that by Toply.’ Other and earlier studies to be mentioned are the productions of Lauth,® Haller,” Portal, LeClere and Manget,!* Tarin,'** James Douglas,!? Goelickius and many other early attempts at bringing together the re- sults-of anatomical study. More recently the work of Daremberg in France, Carus,'* Wieger,’® Weindler,!® and Hopf? in Germany; Osawa'® in Japan, Bardeen’® in America, and Chievitz?° in Denmark are to be especially mentioned. Besides the results contained in the above-mentioned works there is much information to be gleaned’ from the numerous histories of medicine and especially from the 8 Geschichte der Anatomie, in ‘‘Handbuch der Geschichte der Medizin,’’ begriindet von Th. Puschmann, — von Max Neuberger und Julius Pagel, Bd. II, pp. 155-326, 1 9 Thomas Lauth, 1815, ‘‘ Histoire i Taalma’ Strassburg. Up to the time of Thomas Bartholin, 1671 10 Albrecht von Haller, 17 74-177 7, ‘‘ Bibliotheca Anatomica, ’’ Tomes I-II. 11 Antoine Baron Portal, 1770-1773, ‘‘ Histoire de la anatomie et de la chirurgie,’’ Paris, Tomes I-VI. , 12 Daniel LeClere and Jacob Manget, ei ‘í Bibliotheca Anatomica. ’’ 124 Pierre Tarin, 1753, ‘‘Dictionaire Anatomique.’’ 120 James Douglas, 1715, ‘‘ Bibliographiw aiai specimen, seu cata- logus omnium pene auctorum, qui rem anatomicum professo vel obiter sah fi illustrarun 18 Ch. Paxdithare. 1870, ‘‘Histoire des sciences médicales comprenant DEER la Sergei la médicine, la chirurgie et les doctrines de patho- , logie rale. 2t 14 J, ‘Victor oe 137%, ‘í Geschichte der Zoologie, tis auf Joh, Mueller und Charl. Darwin.’ 15 Friedrich lagi 1885, ‘‘ Geschichte der Medizin und ihrer Lehran- stalten in Strassburg vom Jahre 1497 bis zum Jahre 1872.’’ 16 Fritz Weindler, 1908, ‘‘Geschichte der gynaekologisch-anatomischen Abbildung.’’ 17 Ludwig Hopf, 1904, ‘‘Die Anfänge der Anatomie bei den alten Kul- turvélkern.’’? Abhandl. zur Geschichte der Medizin, Heft IX, Breslau. 18 Gakutaro Osawa, 1895, ‘‘ Zur Geschichte der Anatomie in Japan.’’ Vor- trag gehalten in der Naturforscher-Ges. zu Freiburg i.B. am 20, Nov. 1895. A.A.B. 11, N. 16/17, pp. 489-504, 2 Abb. 19 Charles R. Bardeen, ragi Sineme in America,’’ Bulletin of the University of Wisconsin, No. 20 I. H. Chievitz, 1904, epee historie,’’ Copenhagen. No. 604] SOURCES OF ANATOMICAL LITERATURE 195 biographical dictionaries of Panckoucke,2* Gurlt und Hirsch,?* Pagel,? and the various biographical encyclo- pedias. . Much valuable biographical data of many biolo- gists is to be found in the ‘‘ Nouveau Larousse Illustré,”’ as well as in other general encyclopedias. Even a hasty survey of the geographical development of anatomical literature will suffice to show that the con- tinents of Europe and North America are the chief ones to be considered. Asia, Africa, Australia and South America each come in for some slight claim to attention, as will be evident from the discussion of the geographical distribution of anatomists given below. It can not be said that all of the men considered in mak- ing up the list referred to below have contributed new ideas to anatomy. Many have not. None of the Romans were men of original ideas, at least so far as a knowledge of anatomy is concerned. Celsus is the only Roman whose knowledge of anatomical subjects demands any sort of respect and his knowledge, as given in the ‘‘De Medicina,’’ is not acquired first hand although he is to be greatly respected for producing a medical classic. Stieda?** has shown, however, that the Romans were not entirely devoid of anatomical knowledge, though this knowledge was often erroneous. The ‘‘Donaria’’ de- scribed and figured by Stieda are supposedly offerings to the deity in connection with the suppliant’s plea for health. The part offered, in the form of a model of a leg, foot, breast, viscera or head, indicates the region in which the suppliant suffered and from which distress he wished to be relieved. The objects are of marble and bear the date of about the first century B.c. They are, for the most 21C. L. F. Panckoucke, 1820, ‘‘Dictionaire des Sciences Médicales-Biog- raphie Médicale,’’ 7 volumes. 22 <í Biographisches eea Martor agondn Aerzte aller Zeiten und Völker, ”” 1884-88, 6 volum 23 a E Taikon hervorragender Aerzte des XIX. Jahrhun- derts,’’ 190 230 ag Stieda, 1901, ‘‘Anatomisch-Archiiologische Studien,’’ II. Anatomisches über alt-italische Weihgeschenke (Donaria), Anatomische Hefte, Bd. XVI, pp. .1-84, Taf. I-IV. 196 THE AMERICAN NATURALIST [ Von. LI part, crudely done and can not be taken as indicating any attempt to illustrate anatomy. Often the internal parts shown bear some resemblance to the human structures. Occasionally the liver of a mammal is incorporated in the same piece with the human heart and lungs. The viscera are very crude and can not be taken as indicating any de- gree of positive knowledge concerning the parts shown. None of the Arabians produced original ideas or liter- ature concerning anatomy. Abdollatif (1162-1231) is the only one of the Arabians who departed in the slightest degree from the writings of Galen and Hippocrates. While in Egypt he was studying some human bones in a cemetery when he ascertained that the jaw is formed of one piece; that the sacrum, though sometimes composed of several, is most generally of one. On the basis of these observations he criticized the writings of Galen and thus showed himself to be a man of original ideas. Flores** in his ‘‘History of Medicine in Mexico’’ has listed the following teachers of anatomy in the University of Mexico: Febles, Benitez, Garcia, Cheyne, Peña, Garcia Cabezon, Rendon, Escobedo, Villar, Jecker, M. Andrade, Muñoz, Villagran, Duran, F. Ortega, Chacon, Montes de Oca, Velasco, San Juan, and Cordero é Icaza, but, so far as I have been able to ascertain, none 7 these men have been productive. There is no Egyptian literature of anatomy, and ap- parently there was no definite knowledge of anatomical structure. The practise of embalming had attained at one period great perfection in Egypt and this may have resulted in a certain degree of anatomical knowledge, but it was so overclouded: by religious fanaticism and super- stition that it amounted to little. Thus MceKay*** says: “The Egyptians probably knew next to nothing about anatomy, as their religion forbade dissection, and the em- balmers probably learnt little. After the Paraschistes 24 Francisco A. Flores, 1888, ‘‘ Historia de la Medicina en Mexico desde la epoca de los Indios hasta la Presente,’’ Tomo I- 24a W, J. S. McKay, 1901, ‘‘ The History of Ancient Gynaecology,’’ New York, p. 6. No. 604] SOURCES OF ANATOMICAL LITERATURE 197 had made the preliminary abdominal incision, the Tari- cheutae were accustomed to pass their hands through the incision into the body and remove the heart and kidneys and digestive organs. If they were accustomed to remove the uterus and ovaries, they must have gained some knowledge of the organs, but we have no authority for saying that the uterus was really removed. The custom that the Egyptians followed, that of making models of the parts?* that had been healed and then hang- ing them in the temples, may have been useful for clin- ical instruction.”’ Among the Hebrew peoples of ancient times sacrifices (Genesis xv. 9-10) were common, and the appearance of the viscera of the animals sacrificed were probably fa- miliar. This may have resulted in a degree of anatom- ical knowledge. There are, apparently, no definite state- ments concerning anatomical matters in the Bible, the in- formation there given being of a purely popular charac- ter. In the Babylonian Talmud, however, there are a number of references” to subjects of anatomical interest. The number of bones in the skeleton is estimated at 248 or 252, and one of these, the bone Luz, which was supposed to be situated somewhere between the base of the skull and the coccyx, was regarded as the indestructible nucleus from which the body is to be raised from the dead at the resurrection.22 The Talmud also displays some knowl- edge of the esophagus, larynx, trachea, and the mem- branes of the brain. The pancreas and other internal organs are briefly referred to. Among the ancient peoples who thrived in and around Mesopotamia there is, apparently, no anatomical liter- ature. What little of anatomy may have been known was acquired through religious observances, such as auguries and sacrifices. Stieda? especially has studied the indi- 24b These were ‘‘Donaria’’ and the custom was doubtless derived from the Romans. 25 Julius Preuss, ‘‘ Biblisch-Talmudische Medizin,’’ Berlin. 26 F, H. Garrison, N. Y. Med. Jour., 1910, Vol. 92, pp. 149-151. 27 Ludwig Stieda, Anatomische Hefte, Bd. 15, pp. 673-720, Taf. 57. 198 THE AMERICAN NATURALIST [ Von. LI cations of anatomical knowledge among these ancient peo- ples as this knowledge has been preserved in their sculp- tures. It is a matter of great interest that he has inter- preted a terra-cotta object from Babylon, to which an age of from 2000-3000 s.c. may be assigned, as a model of a sheep’s liver, supposedly used in connection with sooth- saying or with sacrifices. This interpretation is sustained by the description of two other similar objects of a later date, one in alabaster from Piacenza, and a bronze liver from Settina. One can recognize on the visceral surface of these objects the processus papillaris, the processus caudatus, and the vesica fellea; all of which are very clearly represented. The lower surface of the object from Babylon is divided into squares and studded with inscriptions, supposedly of a prophetic nature. In addition to these very definite anatomical models many plastic representations exhibit some knowledge of the superficial musculature of the extremities. The larger subcutaneous veins, such as the cephalic, basilic, and saphenous, are often clearly shown. From an anthro- pological standpoint it is noteworthy that various racial types are indicated in some of the representations of the head, so that we can not say that these peoples were en- tirely devoid of anatomical knowledge and we are forced to admit their keen powers of observation. Such anatomy as was taught to students in the medical schools of China was highly erroneous and fantiful. Al- though the study of medicine has a very ancient history in China, as ancient as the history of its civilization, going back to more than 3000 years B.c., anatomy was not studied at all in any laboratory form. They taught, for instance, that there are 365 bones in the human body; that the small intestines were attached to the heart; that they were tra- versed by the products of digestion; that the larynx opens into the heart; the spinal cord into the testicles, that the lung has eight lobes; the liver seven, that the kidneys, sus- pended the vertebral column, have the form of an egg and possess the subtle principle of generating the spermatic No. 604] SOURCES OF ANATOMICAL LITERATURE 199 fluid, primarily elaborated by the brain, condensed in the testes and from there conveyed to the spermatic duct. Osteology was somewhat better known, although the skull, pelvis, forearm and leg were regarded as being formed of one bone each, or at times eight bones were assigned to the head in the male and six in the female. Something was known of the tendons and ligaments. The spleen and heart were regarded as the organs of reason. Ancient Chinese medical literature consists of a large number of works, none of which are of any scientific im- portance. There is no modern Chinese anatomical lit- erature. What little of anatomy was known in ancient India is contained in the writings of Atreya, a physician who wrote a good description of the bones of the human body, and who is said to have taught in the Taxila University during the sixth century B.c.; as well as in the writings of the surgeon Susruta, of a somewhat later date. “As the writings of these men have been interpreted by Charaka, ancient Hast Indian anatomy regarded the body of man as possessing seven skins, seven elements, 300 bones, 24 nerves, 3 fluids, 107 joints, or 68 movable joints and 142 immovable ones, 900 ligaments, 90 tendons, 40 principal blood vessels with 700 branches, and 500 muscles. The blood vessels and nerves®® radiated out from the um- bilicus as a center. Nothing was known of the courses of these structures within the body. There is a later publication of about a.n. 800 entitled **Amarakosha,’’ which discusses somewhat the nature of the human body but there is no later treatise which might be termed anatomical, and there is no modern East In- dian anatomical literature. While engaged in a biographical study of the men who have contributed to the advancement of our knowledge of vertebrate anatomy, the writer has been attracted by a number of interesting facts relating to the sources of anatomical literature, which, he believes, are not generally 28 Haeser, ‘‘ Lehrbuch der Geschichte der Medizin,’’ Bd. I, p. 18-20. 200 THE AMERICAN NATURALIST [Vou. LI recognized. The following preliminary study is an at- tempt to arrive at some conclusions as to the origin and development of our modern anatomical literature. It is presented here in anticipation of further studies along these lines. The names of the men included in the above-mentioned study are of those anatomists, or contributors to ana- tomical literature, who are no longer living, and who have contributed in any way to the anatomy of the vertebrates, whether by practical or theoretical studies. The men of all countries and all times have been listed, so far as it has been possible to ascertain them. Doubtless many have been overlooked because their records are in rela- tively obscure places. The subjects represented in the present study are: human anatomy, artistic anatomy, anthropology, com- parative anatomy, embryology, histology, zoology, verte- brate paleontology and subjects of general interest which bear theoretically on the morphology and evolution of the vertebrates, such as Mendel’s work at Briinn, and the work of Darwin, Weismann, Charles Bonnet and Lamarck. It may be contended that these subjects con- stitute biology rather than anatomy, but biology is cer- tainly the more inclusive term. We may, to be sure, speak of the anatomy of the bacteria and in a sense the bac- teriologist is an anatomist, but for my present purpose the names of those men who have contributed to our knowledge of the morphology of the vertebrates will suf- fice. One of the guides, which has been useful in select- ing the names suited for the list, is that of the anatomical terminology. If there are anatomical structures (such, for instance, as Hesselbach’s triangle) named for the man, he is included, though this is by no means the only guide. Those men whose writings are of a strictly taxo- nomic nature are not included, unless important. theo- retical results have arisen from such writings, such as has been the case with the taxonomic work of Lamarck, Lin- nus, Cuvier and others. No.604] SOURCES OF ANATOMICAL LITERATURE 201 It is hoped later to make a comprehensive survey of the development of anatomy. Over one thousand names have been compiled, and will ultimately be studied. Eight hundred of these have already been partially examined and will shortly be published. In the development of our anatomical literature there has been involved a whole host of men in a number of dif- ferent lines of activity, which it will be interesting to dis- cuss. Many of the men included in the list were not professional anatomists but since they contributed to the advancement of anatomical knowledge we may regard them as contributors and they are hence deserving of consideration. As would be expected, the following survey of the anatomists indicates, in general, an intellectual develop- ment in each country at the time when other conditions, physical, social, religious and political, favored the growth of mental work among the nation, although this statement finds certain contradictions, as in the case of Michael Ser- vetus, Abano, Fallopio, Malpighi, Vesalius, and many more of the early students especially, whose work was ac- complished under adverse conditions. The dates of the majority of the anatomists who find a place in the present study belong to a period when intellectual endeavor had attained a firm place in each country. Among the Greeks for instance, there is no one who attained eminence in anatomy later than the third century a.D’. No modern Greek anatomist is included in the list. There are representative anatomists of twenty-seven nationalities, though many of the countries are not widely separated geographically. The following geographical distribution indicates nothing previously unknown, but is _ presented here simply as an interesting survey. ‘There are doubtless many more men of all these nations who are deserving of mention. The list contains 1 Japanese (Mitsukuri, 1858-1909, an embryologist, who is here re- garded as the only one of this nation who has occupied a high place in anatomical work); possibly also the zoolo- 202 THE AMERICAN NATURALIST [ Von. LI gist Nishikawa and the anatomist Taguchi should be mentioned; 1 Armenian (Aleana Mosali, who in the thir- teenth century wrote a treatise on the anatomy and dis- eases of the eye, chiefly compiled from Arabian, Chaldean, Jewish, Greek and other sources) ; 1 Hungarian, 1 Polish, 1 South American (Florentino Ameghino, 1854-1911, a student of vertebrate paleontology, is the only represen- tative of the large South American continent in the list. Ameghino’s attainments in vertebrate paleontology en- title him to a high place among the anatomists of the world); 1 Turk (Schanzi Zadeh Mehemmed Ataullah, a Turkish physician, who after completing his studies in Italy, published, in 1820, a work on human anatomy, in folio, illustrated with 56 copper plates); 2 East Indians (Atreya, a physician, who was a teacher in the Taxila University in the sixth century B.c. He wrote an osteol- ogy, which was later edited by one of his students, Cha- raka. Susruta, an East Indian surgeon also deserves mention), 2 Bulgarians, 3 Flemish (of whom the greatest was Vesalius), 4 Romans (none of them men of original- ity), 5 Russians, 5 Belgians, 7 Irish, 7 Swedish, 7 Span- ish, 7 Bohemians, 9 Arabians (Abdallatif, Albucasis, Avi- cenna and others), 11 Scottish, 12 Danish, 16 Swiss, 17 Austrians, 22 Greek, 36 American, 40 Dutch, 77 English, 78 Italian, 127 French and 240 Germans, making a total of seven hundred and thirty-six. The citizenship of many of the men studied has been hard to determine on ac- count of the migration of teachers from country to coun- try, which in the seventeenth, eighteenth and nineteenth centuries has been very common; but the above is a fair representation of the proper distribution of the men who have developed anatomy. More than twenty-five professions are represented by the men who have been given a place in the list. In at- tempting to decide the position of a man in the following scheme it is not always easy, on account of the varied in- -erests of some of them, to place them properly. Should Albrecht von Haller, for instance, be regarded as an anato- No. 604] SOURCES OF ANATOMICAL LITERATURE 203 mist, philosopher, poet, physiologist, botanist, or as an ad- ministrator, since he:attained some eminence in all of these lines? Should Emanuel Swedenborg be classed as a philosopher, anatomist, geologist, civil engineer or theo- logian? In the present scheme Albrecht von Haller is arbitrarily regarded as an anatomist, although a very large share of his work was physiological. Swedenborg is regarded as a philosopher, for as such he is usually classed, although his anatomical writings were of a high type. Descartes is likewise regarded as a philosopher, although he might with justice be called a mathematician or anatomist. The subdivisions of histology and embry- ology are necessary since a few men specialized strictly ‘n these branches of anatomical work, and they are known for their contributions to these subjects; such for in- stance as Balfour’s noted studies in embryology and Corti’s in histology. The following list will show in a general way the numer- ical distribution of the men in various professions: 1 jur- ist, (Johannes Peyligk, who in 1499 published in Leipzig his ‘‘Philosophie Naturalis,’’? which contains ten figures of separate organs of the body), 1 statistician, (Francis Galton), 1 beadle or exciseman (Leeuwenhoeck,”® who for thirty-nine years worked as a subordinate customs officer or beadle at a salary equal to $125 per year. In spite of this meager income he contributed 375 papers to the Royal Society of London and 17 to the Academy of Science in Paris, besides making all of his microscopes), 1 pope (In- nocent XII, who, working under the direction of Lancisi (1654-1720), is said to have been one of the first to ob- serve, under the microscope, the circulation of blood in the capillaries); 1 prior, 1 journalist, 1 theologian (Cas- par Bartholin, the founder of a professorial dynasty in the University of Copenhagen whose members taught in 29 It should be noted that the following account is that of Sir Benjamin Ward Richardson, ‘‘ Disciples of Ausculapius,’’ Vol. 1, p. 111. Garrison, how- ever, says in his ‘‘Introduction to the History of Medicine,’’ p. 185, that Leeuwenhoeck was ‘‘an inheritor of well-to-do brewers (and) led an easy- going life.’’ 204 THE AMERICAN NATURALIST [ Vor. LI the university for one hundred and thirty-five years) ; 1 lawyer (Michel Alberti who contributed nearly 300 separate works on several phases of human knowledge) ; 1 bibliographer (Mangetus, ‘‘ Bibliotheca Anatomica’’), 1 clergyman (Wm. Buckland), 2 monks (of whom one was Gregor Mendel and the other Michael Servetus, a Spanish monk, the discoverer of the pulmonary circu- lation, which he published in 1553, seventy-five years before the appearance of Harvey’s great work on the motion of the heart and the circulation of the blood); 3 physicists (Helmholtz and others), 5 poets (Mark Akenside, who wrote his inaugural dissertation on the fetus; Goethe, who is widely known for his papers in comparative anatomy, and for his homology of the inter- maxillary bone of men and mammals; Lucius Francois Anderlini, a surgeon of Saint-Angelo, in the duchy of Urbino, who published in 1739 a poetical work ‘‘The Anatomist in Parnassus, or a Compendium of the Parts of the Human Body arranged in Verse’’; Scipion Abeille, a military ‘surgeon in Flanders who wrote, in 1689, a poetical anatomy on the parts of the head and neck; Al- brecht von Haller was a poet of note, and many other men, interested in anatomical subjects, have been poetically in- clined), 8 artists (Albrecht Dürer is well known for a work on human proportions which is of value from an anthropological standpoint; Michelangelo, working with Realdo Colombo (1494-1559), became deeply versed in human anatomy; Leonardo da Vinci, about 1510, com- pleted a wonderful series of anatomical sketches, based on his own dissections) ; 5 ophthalmologists, 5 anthropolo- gists (Blumenbach (1752-1840) was the founder of this science); 5 comparative anatomists, 6 embryologists, 7 pathologists, 9 histologists, 8 botanists, 14 paleontolo- gists, 17 philosophers (Aristotle, Descartes, Swedenborg, ete.), 24 physiologists, 52 zoologists, 69 surgeons, 175 physicians, and 255 professional anatomists who devoted most of their attention to the teaching of anatomy. It has not been possible to determine accurately the No.604] SOURCES OF ANATOMICAL LITERATURE 205 profession of many of the men included in the list, and for this reason many whose names are in the list are not classified here. For instance, Petrus d’Abano, who pub- lished in 1496 the first illustrations of the abdominal mus- cles, was either a physician or a professional philosopher, and probably the former, since Locy says the illustrations seem to have been based on a dissection. And there is a story concerning the large fees charged by Abano, which indicates that his profession may have been medicine, although his intellectual interests were chiefly philo- sophical. Bartholomeus Anglicus was probably a physi- cian.or a publisher. At any rate he published in 1485 one of the first printed illustrations (a wood cut) of any anatomical interest. Many surgeons have contributed to anatomy, and were really at the same time teachers of anatomy, such as Nicolas Ivanovitch Pirogoff, who wrote an enormous cross-section anatomy in five volumes, pub- lished in 1852; notwithstanding which he is classed in the list as a surgeon, since his anatomical teaching appears to have been incidental to his surgery. The social status of the men who have developed ana- tomical knowledge has been difficult to determine because of scant biographical data. There is sufficient, however, to indicate that contributors to anatomical knowledge have been recruited from a wide range of social condi- tions. Some of them, and often the brightest, have lived in poverty. Others have been representatives of a much higher social class. It may safely be said that the study of biological matters has attracted attention of no special class, but that interest has been scattered. It may be said that the great majority of men who have developed anatomical knowledge have been men of moderate attain- ments, belonging to an average rank in the social scale. The above statements must be modified by the conditions under which the men lived and the age in which they lived. During the early centuries of the Christian Era living conditions in general were not so wholesome® as they have since become. 80 Hirsch, August, ‘‘Handbook of Geographical and Historical Pathol- ogy,’’ 3 vols. 206 THE AMERICAN NATURALIST [ Von. LI Precocity and productiveness have gone hand in hand among the few anatomists who have exhibited these in- teresting traits; yet it is only fair to state that pro- ductivity has not been dependent on precocity. Such men as Bichat, Balfour, Haller, Vesalius, Johannes Mueller, Bernard Siegfried Albinus, Pollard, Sir Charles Bell, are rather unusual examples of precocity. It may be of interest to give a few detailed accounts of some of these men. During the short period of seven years, beginning his career at the age of twenty-three, which Marie Francois Xavier Bichat (known as the father of histology) devoted to his scientific studies, he came to be recognized as one of the foremost biologists of all time. He exhibited unusual talents for prolonged and intense application to the pursuit of his favorite science. Be- sides editing the surgical writings of his teacher, Pierre Joseph Desault, in three volumes, he is himself the author ` of three separate works, any one of which would have secured him fame. Bichat’s claim to recognition as a great biologist lies in his division, in 1800, of the tissues of the body into twenty-one non-microscopic varieties. Francis Maitland Balfour ended his brief career at the same age as did Bichat, thirty-one; but during the few years he devoted to his favorite study of embryology he laid a secure foundation for lasting fame. Especially in his monograph on the development of the elasmobranch fishes and in his comparative embryology, he exhibited a broad grasp of the subject which has seldom been equalled in the same field of learning. Andreas Vesalius, the great Flemish anatomist, de- scended from a family of learned physicians, began his study of anatomy at the age of fourteen with Dubois in Paris, and at the age of twenty-two was called to Padua to give public demonstrations in anatomy. His large work on human anatomy, ‘‘De corporis humani Fabrica,’’ which earned him the title of the founder of modern sys- tematic anatomy, was published when he was thirty. Al- though he lived for twenty years after its appearance he No. 604] SOURCES OF ANATOMICAL LITERATURE 207 did little or nothing to develop anatomy save to issue suc- cessive editions of his ‘‘Fabrica.’’ Three years before his death there appeared from the press at Madrid his edition of Fallopio’s anatomy. Albrecht von Haller, Swiss anatomist, physiologist, poet, botanist and administrator, deserves to be regarded as the most precocious and one of the most productive of all the men who have contributed to the advancement of anatomy. At the early age of eight he is said to have compiled a biographical index of over 2,000 eminent men and women. This prodigious activity he continued for the next sixty years, and it is stated that he conducted a monthly scientific journal to which he himself contributed 12,000 articles on nearly every phase of human knowl- edge. Nor were his contributions superficial, for Sir William Turner says that his anatomical descriptions and his beautiful and accurate figures were the most val- uable which had appeared up to that time (1746-51). A list of his medical writings alone fills eleven octavo pages of closely printed type. Late in life he returned to Berne from Göttingen, where from 1736-1753 he had held the position as professor of anatomy, physiology, surgery and botany, to engage in his native land in municipal ad- ministration. : Johannes Mueller, who, in the first half of the last cen- tury, became famed as an anatomist, zoologist, and physi- ologist, became, at the age of twenty-five, professor ex- traordinary of physiology at the University of Bonn. He began an early career of prodigious activity, which he continued for thirty-three years. Avicenna at the age of seventeen was regarded as an excellent physician. At twenty-one he was the author of several treatises. He was called ‘‘The Prince of Arabian physicians’’ by his contemporaries. ` Bernhard Siegfried Albinus (1697-1770), for fifty years a teacher at the University of Leyden, was called to _ the University at the age of twenty-one, from Paris, whither he had gone on the advice of his father, to study 208 THE AMERICAN NATURALIST [ Von. LI medicine and especially anatomy with Winslow and Sénac. He had hoped to spend some years in Paris, but after six months, on the retirement of Rau from the professorship of medicine, anatomy and surgery at Leyden, Albinus was called, at the suggestion of Boerhaave, to take charge of the anatomy. Shortly after reaching Holland the Uni- versity of Leyden gave Albinus his doctorate of medicine without either examination or thesis. His inaugural ad- dress ‘‘Oratio inauguralis de anatome comparata’’ clearly showed the master mind. At Leyden, Albinus gave a new . direction to the study of anatomy which had lain dormant since the appearance of Vesalius’s ‘‘De Corporis Humani Fabrica” (1543). He brought to greater perfection the art of anatomical illustrating, which had not pro- gressed since Vesalius, and especially in his ‘‘ Historia musculorum hominis, Leyden, 1734, in-4°’’, on which his fame as an anatomist rests. This magnificent work was twice reprinted and translated into French by Pierre Tarin in 1753. Albinus published also other valuable works and left a marked impression on his subject. It would appear, from the above study, that the sources of anatomical literature are to be found in the writings of the men who have developed the subject in the various countries mentioned. The literature of anatomy has now attained sufficient dignity to warrant the prepara- tion of a ‘‘Source Book,” which would be very useful. Africa, aside from the Grecian incursion in the early cen- turies of the Christian-era which resulted in the Alexan- drian school, has produced no men of attainments in anatomy. South America has one man to its credit. Mexico and China have none. The literature of the rest of the world has radiated out from those European coun- tries which have fostered our modern civilization. The outlook for an excellent type of anatomical literature in the future is better than it has ever been and the student who attempts to work in the field of the history of this literature will find himself among interesting and de- lightful surroundings. THE CASE OF TRICHOMONAS! DR. PHILIP HADLEY THE great group of flagellated protozoa has, within the past two decades, afforded a wealth of interest for those concerned with pathogenic protozoology; and only in slightly lesser degree for those concerned with taxonomic problems involving these highly interesting microorgan- isms. The field of trypanosome research has, in itself, afforded much new data on morphology and on compli- cated life histories; and has been the chief center of in- terest for many years. But there exists another group of the flagellated proto- zoa, represented by some of the commonest forms encoun- tered in the intestinal tract of man and the lower animals, whose frequency of occurrence, simplicity of organiza- tion and freedom from imputations of possessing patho- genic powers, have enabled them to go their way, for the most part unmolested by the protozoologist. If the proto- zoan would escape the inquiring gaze of the researcher he must be self-effacing; he must lead a quiet life of seclu- sion, free from those public manifestations of unrest and mob movement which are sure to bring him, sooner or later, before the bar of investigation, whereupon his whole life is laid bare. Trichomonas was such a quiet law-abiding protozoan before the trouble began, before he was detected in insti- gating internal revolutions which bid fair to annihilate the turkey-raising industry of the country. The cireum- stantial evidence which has been brought forward against him has served to reveal many aspects of the life history of Trichomonas with which we were not previously ac- quainted; to disclose his participation in activities for which he was previously regarded as scarcely capable, and to demonstrate the existence of certain family resem- blances to some of his companions in mischief who have long been recognized as trouble-makers in the cell or- ganizations of many animals. j 1 Contribution No. 231 from the Agricultural Experiment Station of the Rhode Island State College, Kingston. 210 THE AMERICAN NATURALIST [ Von. LI Trichomonas is found living in the intestinal contents of nearly all animals and has, since its discovery by Donné in 1837, appeared under many different names. Itisa small organism, built on an oval or pear-shaped plan, and meas- uring in the adult trophozoite stage, about 10» in length by 5 to 6» in breadth. The youngest free-swimming stages are much smaller, about 5» in length; and some- times trophozoites are encountered that measure 12 or 13». Although usually of an elongate oval or pear shape, the morphology of the trophozoites is highly variable, and triangular or crescentic forms are frequently encountered, especially among the young. The anterior end is usually blunt, while the posterior end is frequently drawn out into a point. If one adds to the salt solution in which these flagellates are being examined a little albumen or glycerin, to lessen the rapid swimming of the organism, some of the details of structure can be made out. The body plasm shows a greenish tint, and the nucleus, which is situated an- teriorly, appears pinkish. In fresh preparations, one of the most obvious features is the axostyle, a short bristle- like structure which projects outward somewhere in the posterior quarter of the body, and which is seen, upon careful focusing, to extend into the body of the flagellate, running anteriorly to terminate somewhere in the vicinity of the nucleus (Figs. 1, 2). Inside the body the axostyle appears homogeneous in structure and bandlike. - Next to the axostyle, the most obvious feature is the vibratory or undulatory membrane which extends like a curved fin down the dorsal side of the flagellate body (Fig. 1). It is shallow at the beginning and at the end, but midway of its length it may have a depth of 2 to 3». Over this membrane may be seen to travel at 3 to 4p in- tervals, waves of motion from the anterior toward the posterior end of the body. If one follows closely the. „course of this membrane, it is found to have its origin in a granule, or in a group of small granules, located in front of the nucleus at the most anterior part of the animal, and known as the blepharoplast-complex. The granules No. 604] THE CASE OF TRICHOMONAS 211 are very small, measuring not more than 0.5 to 1.02, and stain deeply with the chromatin stains. Their function is at the present time only a matter of speculation. Tracing the dorsal membrane posteriorly, it is found to extend to the extreme end of the body, where it narrows and is con- tinued in the form of a terminal flagellum (‘‘Schlepp- geissel’’) which has a length ordinarily about dome to that of the body. _ From the anterior end of the flagellate extend cu more flagella (Fig. 1). These may be even longer than the body of the flagellate itself, and beat downward, as indicated by the arrow in Fig. 1. It frequently appears as if two of these three flagella were united in a common stalk at their base, so that they beat together, while the third flagellum beats independently. The origin of these three flagella is difficult to make out, but in many cases they appear to arise from one of the granules of the blepharoplast-complex, and usually not from the granule- which is the origin of the undulatory membrane. The only other structures which can be seen well in fresh preparations are the mouth or cytostome and the food vacuole. The cytostome is a horn-shaped open- ing which extends into the body on the ventral side, and just behind the nucleus (Fig. 2). It may be bordered by cilia. The beat of the anterior flagella isin such a direc- tion that currents of fluid containing the bacteria which serve as the chief food for the flagellates, are driven into the mouth opening. Posterior to the nucleus, usually in about the middle of the cell body lies an oval space, the food vacuole (Fig. 2). It may sometimes be represented by a group of smaller vacuoles which coalesce to form a single cavity. In these vacuoles are usually present bac- teria and cocci undergoing digestion. The structures mentioned above can be seen well in unstained organisms, but there are others which appear to advantage only upon Staining. In preparations stained by the Heidenhain iron-hematoxylin method (wet process) the most note- worthy of the remaining structures is the chromatic line. This is a heavily-staining band which extends like the are 212 THE AMERICAN NATURALIST [ Vou. LI of a circle from the blepharoplast to the point where the undulating membrane terminates. It thus follows closely, in the body plasm, the trend of the membrane, and is re- garded as representing a kind of supporting structure. The chromatin line is heavier in its mid part and tapers at each end. As will be pointed out later, when in the process of spore formation, the trophozoites round off, the chromatin line becomes bent into a hoop, so that its extremities come very near to meeting (Fig. 4). . Another structure which appears with distinctness in stained preparations is the line of chromatic blocks (Figs. 3,4). These peculiar bodies appear as a single or double row, or as a somewhat irregular line, of deeply staining granules extending from the region of the blepharoplast backward through the plasm to end somewhere in the posterior quarter of the cell. The anterior portion is likely to be thicker and sometimes may partially obscure the nucleus. The curve followed by the line of blocks is about parallel to that of the chromatic line and the two are seldom far distant from one another. It is interesting to observe in connection with all of these struetures that, in their arrangement, they produce in the flagellate organism a more or less perfect bilateral symmetry. The normal swimming position of the tropho- zoite is with the undulatory membrane above. Directly below this extends the chromatic line and below the chro- matic line is the ‘‘line of blocks.’? The cytostome is in the midline and somewhat ventral. The blepharoplast is in the midline except in some of the stages of division. The food vacuoles occupy a variable position, but are usu- ally grouped near the middle of the posterior body and caudad of the chromatic line. Sometimes it appears as if the line of blocks and the axostyle passed through the food vacuoles. The axostyle projects from the cell body in the midline although not necessarily at the most poste- rior part of the body. This symmetry is easily seen when the organisms are observed swimming freely in a favor- able medium. Owing to the fact that the dorso-ventral diameter is greater than the transverse diameter, most 214 THE AMERICAN NATURALIST [ Vou. LI of the flagellates when stained on the slide present a lateral aspect as shown in Fig. 2, since they fall over on to their side in the drying out of the film. But the appearance of the flagellate as described above does not endure for very long, simply because the tropho- zoite stage itself does not endure. The development of the trophozoite marks the period of youth, and when the organism has sufficiently fed on bacteria and cocci, and obtained a sufficient amount of reserve food, it passes on either into division or into a form of autogamous repro- duction by which the flagellate population is increased at a rapid rate. In the case of division, the process seems to be for the most part longitudinal. The first indication of it is to be seen in the blepharoplast-complex and in the nucleus. From each new blepharoplast there appears to grow out a new chromatic line, extending more or less parallel to the old line. From these new lines the new undulating mem- branes appear to arise. The writer has not been able to observe the division-stages of the flagella, although stages have been seen in which new flagella are present in con- nection with each new blepharoplast. Neither has it been possible to follow the changes in the axostyle. As to the chromatic blocks, these also seem to disappear and are probably formed anew in the daughter cells. But reproduction by division, though occurring com- monly in the intestinal content, is probably not the chief method of reproduction. At all times, though at some times more markedly than at others, the flagellates enter into a course of autogamous reproduction in which several daughter cells are formed out of a single mother cell. This interesting process can be followed in considerable detail by means of suitably stained smear preparations. The first step in this process is the ‘‘rounding-off’’ of the previously elongate or crescentic trophozoite after it has reached maturity. If the body-form was erescentic there occurs a filling-out of the concave surface so that at first a full oval shape is produced; later the organism be- comes spherical. This rounding-off process, which is No. 604] THE CASE OF TRICHOMONAS 215 usually accompanied by some increase in size, is marked by important changes in the structures alluded to above. These may be considered in some detail first with ref- erence to the external features. Perhaps the most noteworthy change, aside from the assumption of a spherical shape, involves the chromatic line. This gives the appearance of lengthening until it forms a hoop almost completely encircling the organism (Fig. 4). Itis common to see the ends of the line occupy- ing positions less than 45 degrees apart as measured on the circumference of the spherical flagellate. At the same time the flagella have been lost and the undulatory mem- brane has decreased in size, though it follows approxi- mately that part of the circumference corresponding to the chromatic line. Of course its functioning has been proportionately reduced and although its undulatory mo- tion may continue, this movement fails to cause progres- sive movement of the flagellate, but brings about a slow rotation of the organism in the same position. Occa- sionally this movement may be assisted by a single an-, terior flagellum or a remnant of one which remains after the others have disappeared. In the final stage all trace of the cytostome is lost, and in fresh preparations the or- ganism appears as a ball of fairly homogeneous fluid, sur- rounded by a granular cytoplasm containing the nucleus (Fig. 5). But more interesting are the changes that have been occurring in the internal structures, as revealed by stained preparations. The alterations in the food vacuole are possibly the most significant. In the trophozoite stage the vacuole was made up of one or more spaces represent- ing probably not more than one eighth to one tenth of the organism (Fig.2). As the rounding-off process proceeds, the vacuole increases in size until it oceupies the greater part of the ventral portion of the flagellate (Fig. 4). It begins to crowd the cytoplasm against the dorsal wall, and in this area lies the nucleus, which, as a result of pressure, becomes somewhat flattened. At the same time the ‘‘line of blocks’? and the axostyle, which gives the 216 THE AMERICAN NATURALIST [ Von. LI appearance of passing through the food vacuole, begin to degenerate and eventually both disappear. The chromatic line endures for a longer period, however, and remnants of it may be seen for some time after the ‘‘line of blocks,’’ axostyle and undulatory membrane have vanished. The blepharoplast also can be detected as long as the remnants of the chromatic line are visible (Figs. 6, 7). This in- “crease in the size of the food vacuole seems to be due, partly at least, to the taking-in of fluid, since while this process is occurring the flagellate is increasing in size and becoming more plastic in the constitution of its proto- plasm. The food vacuole has now increased in size to represent the greater part of the flagellate cell and is surrounded by a crescentic ring or layer of cytoplasm seemingly much reduced in amount (Fig. 5). From this time on the most important changes concern the nucleus. This is now flat- tened or sometimes flask-shaped, and soon divides into two equal portions which travel through the region of cytoplasm to take positions at opposite sides of the ball of reserve substance (food vacuole). Here each experi- ences a further division resulting in the production of four daughter nuclei (Figs. 7,8). These apparently may divide again until either eight or sixteen daughter nuclei are formed occupying positions about the periphery of the cell. Frequently smaller portions of nuclear sub- stance are to be seen in the cytoplasm following the first nuclear division and it is probable that these represent reduction bodies (Fig. 8), although the writer has not observed them in the course of formation. About the daughter nuclei there seems to gather by slow degrees a layer of cytoplasm and eventually they break out of their . peripheral ring of maternal cytoplasm to enter the ball of reserve substance occupying the center of the cell (Fig. 10). This is gradually consumed by the young organisms which slowly take on an elongated shape. During this time the cyst wall which had formed about the mother cell has been weakening and finally the young organisms break out of the mother, cell and appear as the youngest No. 604] THE CASE OF TRICHOMONAS 217 trophozoites measuring from 4 to 5» in length and about 3 in breadth, equipped with anterior flagella at least, and possessing a relatively largé nucleus and minute blepharoplast. The other organelles characteristic of the mature trophozoites appear to develop by degrees as the trophozoite increases in size. These, then, are the two chief methods of reproduction. Ordinarily the course is very simple, but from a study of both fresh and stained material it is clear that several complicating factors may enter. For instance there is evidence that conjugation may occur, not only between two individuals but perhaps between three or four. This process is aided by the extrusion of a viscid membrane by those organisms that have rounded-off. This naturally helps to cause the individuals to adhere together. After conjugation this viscid membrane appears to harden into a protective cyst wall. Usually the size of the single cyst is about 10 to 12, but in the ‘‘fused’’ or conjugated forms the diameter may reach 20 to 30» as seen in fresh preparations. It is also clear that the ‘‘double’’ and ‘*triple’’ cysts sometimes seen may represent a division of the original cyst, whereupon each daughter cyst con- tinues independently the production of daughter cells by the usual method, described above. Reproducing by the methods described above, Tricho- monas ordinarily lives in the intestinal tract and causes no recognizable injury to the host. It has never been re- garded as other than a harmless commensal. Recent | studies? have demonstrated, however, that, upon occasion, this flagellate may depart from its usual mode of life, may penetrate the tissues of its host and cause fatal lesions, not only in the walls of the intestinal tract, but in the liver as well. It is especially this.assumption of a pathogenic rôle, this sudden adaptation to a new manner of life, to- gether with the morphological changes that accompany it, that constitute perhaps the most interesting phase of the life of Trichomonas. First, how does it happen that the flagellate gets started on its tissue-despoiling career? 2 Rhode Island Agricultural Experiment Station, Bul. 166, 1916. 218 THE AMERICAN NATURALIST [Vou. LI What is the first stimulus that creates out of a commonly law-abiding protozoan, an invader that has no equal among protozoan forms in the rapidity and completeness with which it carries on its ravages in the -intestinal tissues? This is a difficult question, and one which can not be answered with any degree of finality at the present time. The facts of the matter are these: The manifestation of the disease, as it appears for instance in the so-called blackhead of turkeys, is invariably preceded by a diarrheal condition in which the flagellates appear in increasing numbers as the course of the disease advances. Finally they appear, not only in the liquid cecal content, but in the very depths of the cecal tubules or crypts; and finally in the tissues behind the epithelial wall. From this posi- tion, by a process of autogamous reproduction, the in- vasion of the mucosa, submucosa, muscularis mucose and even the muscular layers, goes on rapidly; and even- tually the whole cecal wall is crowded with the parasites. Secondary bacterial infections may intervene and the re- sults are almost invariably fatal. The question now arises: Are these countless flagellates, present in the liquid cecal contents at the beginning of the attack, the cause or the result of the diarrheal condition? Clinical evidence, which can not now be presented in detail, seems to indi- cate that the latter circumstance is the actuality: that the diarrhea is the primary condition and the increase in the number of parasites the secondary. To explain the ‘first cause’’ of the disease, then, one must explain the cause of the diarrheal condition; and this, of course, is likely to prove in itself, a complex problem, but seems to lead back to certain circumstances related to the nature of the food materials and their assimilation, lying outside the province of the present paper. For a long time it was not clear how, after their rapid multiplication in the intestinal content, the parasites were able to penetrate the epithelial wall and reach the sub- epithelial tissues. Recent studies* have shown the rôle 3 Rhode Island Agricultural Experiment Station, Bul. 168, November, 1916. No. 604] THE CASE OF TRICHOMONAS 219 played by the goblet or chalice cells of the crypts of Lieberkiihn in this respect. Trichomonas, after congre- gating in vast numbers in the fundi of the crypts, with a consequent bulging of their walls forces its way into the goblet cells. It is not deterred by the nucleus or the cell wall at the basement end, but throws the former out of place and breaks through the latter to assume a position beneath the epithelium of the crypt. The wall having been ruptured and an avenue created to the deeper tissues, other flagellates follow by the same path until many are present between the epithelium and the basement mem- brane. But Trichomonas does not halt here. It is now filled with the spirit of the invasion and quickly pushes through the basement membrane into the loose connective tissue of the mucosa. This tissue is speedily overrun by the advancing hosts, the barrier of the muscularis mu- cose is passed and the entire submucosa exposed to the ravages of the parasite. It is here that we recognize Trichomonas in a new rôle. Having experienced its first taste of blood its whole nature is changed; it becomes another animal, raging through the tissues and impeded by no protective action that the host organism is able to muster to the defense. Here then we must recognize Trichomonas as a cell parasite, an or- ganism that has the power to actively invade living cells and to bring about their destruction. One may remark that the type of cell invaded is highly specialized type, and one that, by its nature, is more or less open to in- vasion. But the fact remains that host cells are invaded, and actively invaded; and in this circumstance we can detect, in the behavior of Trichomonas, a foreshadowing of those cell-invading activities regarded as character- istic of the sporozoa. But of course the host-organism must put up some de- fense, and sometimes a very vigorous defense is offered, chiefly by means of its batteries of endothelial and other phagocytic cells. These come out in numbers to meet the invaders and as a result many of the flagellates are en- gulfed, either by single endothelial cells or in giant cells. 220 THE AMERICAN NATURALIST [ Vou. LI But the curious part of this circumstance is that the en- gulfing of the parasites seems to be of slight avail in re- tarding the invasion; and, in many instances without ap- preciable detrimental effect upon the parasites engulfed. From observations on the staining reactions and on the morphological features of the ingested parasites there is good evidence that Trichomonas is not disintegrated by the process; and much less killed outright. It shows a marked resistance to the plasm of the endothelial cells, within which it frequently appears that development may proceed, and from which a new generation of flagellates may break out to continue the course of infection. This would imply that the parasites, once engulfed, are able to make use of the plasm of the endothelial cell as food. And some evidence actually seems to support the view that the parasites fare better in the endothelial cells than they do without. In any region of invaded tissue the ma- jority of the organisms are present within the engulfing cells. If these views should prove valid it must be ad- mitted that a curious situation is produced: the parasites, to survive, must be ingested by the defensive cells, while these phagocytic agents in carrying out their normal de- fensive function, are favoring the growth and activity of the invaders. Of course the residence of Trichomonas within the endothelial cells is purely a passive cell-para- sitism, although the penetration of the goblet cells is an act of active cell-parasitism. But when we regard both together, the matter is of considerable interest in its bear- ing upon the origin of the sporozoa, cell parasites most exclusively. From such elementary invasive power and from such primitive toleration of unfavorable host-cell influences as we see in Trichomonas, it is easy to imagine . how the most effective stages of sporozoan parasitism may have evolved. It is a beginning of that marked adaptability of form and of physiological organization which lies at the base of all pure parasitism as it occurs in the higher orders of the protozoa. Another noteworthy feature in the life of Trichomonas, and one which again serves to connect the organism with No. 604] THE CASE OF TRICHOMONAS 221. - the accepted type of sporozoan parasitism deals with the manner of obtaining its food. It has already been pointed out that when the trophozoites are developing in the in- testinal contents they ingest large numbers of bacteria; whether, at this time, osmosis plays any part in cell nutri- tion is a question. When Trichomonas has entered the deeper tissues, however, the situation is different, since there are ordinarily few bacteria in these regions. Here it seems that nutrition by osmosis must play an important role in supporting the life of the rapidly multiplying or- ganisms. It thus appears that Trichomonas is sufficiently adaptive to new conditions of existence in the tissues to substifute an osmotic method of nutrition for the in- gestive. This nutrition by osmosis it will be at once recognized is one of the characteristic features of the sporozoa, and here again is to be seen a link connecting these two protozoan types. But there is another point of interest involved in this change in the manner of nutrition when Trichomonas enters upon its tissue despoiling career, and this concerns the influence of the manner of nutrition upon some of the morphological features of the parasites. In regarding thé appearance of the flagellates pre- ceding their invasion of the tissues, and after they have gained a foothold in the submucosa, a marked difference is to be observed. This has already been mentioned and may be so great as to deceive one into the belief that the parasites which are found in the intact crypts and which penetrate the epithelial wall, are not identical in nature with the organisms occurring in the deeper tissues. It is this difference which has led some writers to believe that we are dealing with two different protozoan forms. The difference lies primarily in the following circumstance: In the cecal content the flagellates are represented by two forms, the motile trophozoite and the encysted organism. In the case of the latter, one can usually observe clearly the large ball of reserve-substance, and the relatively large daughter nuclei. When developing in the tissues, on the other hand, although the motile forms can be recognized 222 THE AMERICAN NATURALIST [ Vou. LI without difficulty and although the sporulating forms, characterized by the presence of the daughter nuclei, are also observable, both of these are relatively uncommon, and the stage which shows the well rounded ball of re- serve-substance (‘‘Reservestoffsballen’’) is seldom met with. How can these phenomena be explained? The writer has introduced this point in connection with the discussion of the methods of nutrition of the flagel- lates, simply because it seems possible that the morpho- logical differences alluded to above are conditioned by the nature of the food supply. The writer has already traced the changes which the food vacuole of Trichomonas undergoes during the process of encystment. It was shown that there is a direct transformation from the food vacuole of the trophozoite, laden with bacteria and cocci, to the ball of reserve-substance which eventually crowds out the nucleus, chromatic line and line of blocks from the inner part of the cell and may possibly absorb the axo- style. Finally it comes to lie as a mass of varying size with respect to the cell, in the center, or slightly to the ventral side of the organism. Its staining qualities sug- gest a glycogen-like substance, and its density appears to vary with the stage of digestion of the food substances which are to serve the young daughter cells. As stated above, this well defined reserve-substance mass is seldom observed, at least well developed, in the flagellates located deep in the tissues where the evidence favors a view of nutrition by osmosis. The question is therefore raised: Can it not be that the marked difference between the appearance of the flagellates in the tissues and in the cecal content is dependent directly upon the nature of the store of reserve food; and thus indirectly upon the manner of nutrition. This view is in agreement with the general observation that protozoa that subsist by osmosis seldom manifest either food vacuoles or definitely segregated bodies of reserve food substance. In just what way the presence or absence of a ball of -reserve-substance would explain all the differences ob- served in the parasites in and out of the tissues it is diffi- No. 604] THE CASE OF TRICHOMONAS 223 cult to say. That its absence would determine a more homogeneous cytoplasm at all stages of growth is, of course, obvious, but its effect upon the cell structures such as chromatic line, blocks and axostyle, and upon the relative size and distribution of the daughter nuclei is still not clear. It can scarcely be wondered at, however, that such a radical change in the manner of nutrition of a parasite would be accompanied by alterations of some de- velopmental significance. Upon superficial observation it appears that, in a para- sitism of this sort, when the organisms are driving ever deeper into the tissues, one of the essential features of complete parasitic activity is absent, namely, the ability to escape from the tissues and to secure a position by virtue of which the parasite can insure the possibility of reaching other hosts. Without this possibility provided for, no parasitism can be called complete. Although, in the case before us, many of the parasites are so buried in the tissues, a study of the trend of the infective process as a whole has revealed a means by which the organisms return to the cecal content after their invasive career has ended. This is by spreading downward and inward through the reticular tissue of the cores of the villi and pushing the epithelium off of the villus tips. Behind the epithelial wall at these points the parasites congregate in vast numbers until finally the epithelium breaks and liberates the flagellates into the cecal contents. That this process of escape from the tissues takes place only over certain areas of the intestinal wall is apparent; but the fact that it occurs at all is sufficient evidence to indicate that Trichomonas is not wholly lacking in this essential element of successful parasitism. And finally we find in the case of Trichomonas one more lesson, and this is one for the etiologist, this being of course any one who concerns himself seriously with disease etiology. This important person, confronted with a disease of unknown cause, busily sets about to discover the germ; and having found the germ, he as busily en- gages himself in ascertaining means and measures \ 224 THE AMERICAN NATURALIST [ Vou. LI whereby the germ may be avoided by all susceptible folk. We are warned to avoid the places where the germ lurks, to boil our drinking water and to put cotton in our noses; and of course this has been of immense value in prevent- ing infection in the case of many communicable diseases. But this conception of escaping the germ, a procedure still by force of habit widely applied, unfortunately does not work out successfully in all cases, simply because we have at last found that the germ is not always escapable. It may be right with us day and night; and whether we succumb to an eventual invasion depends not upon our side-stepping the organism, but upon our maintaining cer- tain of the body defenses at the proper level of efficient working. The case of Trichomonas in its proper host is an instance. For twenty years (under other names) it has been consistently avoided and wholesomely feared by in- telligent turkey raisers. Five hundred regulations more or less have been directed against it; and now we find that it is always there and always will be there. To keep it in an amicable state, to deter it from making destructive excursions into the tissues, all that is required is to main- tain a normal and hygienic condition of the intestinal tract, whatever this may mean; this alone appears to be sufficient. Thus, although no other intestinal protozoan is able to exert, in a brief time, a greater destructive activity than Trichomonas when properly aroused, still we are far from justified in placing its name upon the blacklist of unqualifiedly pathogenic types which are, by both heredity and training, trouble-makers. On the other hand we can not continue to place this flagellate in that sainthood of parasites, the ‘‘ harmless commensals,’’ since, upon occasion, it may be far from harmless. Trichomonas must now be registered as a facultative parasite, which offers a wealth of interesting subject-matter for research covering several fields of biological study. LINKAGE IN MAIZE: ALEURONE AND CHLORO- PHYLL FACTORS! E. W. LINDSTROM CORNELL UNIVERSITY Genetic linkages or correlations are beginning to con- firm the modern chromosome conception of heredity. As the Mendelian analysis of a species reaches a point where the known genetic factors exceed the number of chromo- some pairs, certain group relations between the factors should become evident. Comparatively few genetic link- ages have been observed in plants, however, probably be- cause the number of Mendelian factors that have been determined is relatively small compared with the number of chromosomes in most species. Physiological or morphological correlations, on the other hand, are far more common. But in the present state of knowledge they are not classified as genetic and consequently can not be used as material for determining the relationship between any series of heritable factors. An intensive, Mendelian study of maize is gradually revealing genetic correlations. Although more than thirty definite Mendelian factors have now been determined in this species, linkages are limited in number because of the relatively large number of chromosomes (at least nine pairs). As early as 1906, Webber noted a general correlation in maize between color in the aleurone layer and in the stamens, glumes and silks. At that time, the genetic con- stitution of color in the kernels and other parts of the plant was unknown. Consequently the correlation was not analyzed on a factorial basis. In 1911, Emerson described an apparent linkage be- tween color of cob, pericarp, husks, silks, and anthers in 1Given before the Botanical Society of America at the annual meeting held in New York City, December 28, 1916. Paper No. 58. Department of Plant Breeding, Cornell University, Ithaca, N. Y. 225 226 THE AMERICAN NATURALIST [ Von. LI maize. Further evidence (unpublished), however, has led Professor Emerson to prefer a simpler explanation for this phenomenon than linkage. Collins and Kempton (1911) and Collins (1912) have reported a genetic correlation between aleurone color and endosperm texture. In this case one of the pairs of aleurone factors is apparently linked with one pair of endosperm factors (horny and waxy). The data indicate that there is a little less than twenty-five per cent. cross- ing over. In a later paper, Collins (1916) describes five character pairs in maize that show apparent genetic correlation. But since these characters exhibit ‘‘blending’’ inheritance and have not been analyzed from any factorial stand- point, it becomes impossible to use these correlations in determining the group relations between the genetic fac- tors concerned. Collins also notes a large number of physiological correlations in this article. LINKAGE BETWEEN ALEURONE AND CHLOROPHYLL FACTORS A definite linkage has been found in maize between one of the five pairs of aleurone factors (Aa, Cc, Rr, Pp, Ii)? and one pair of chlorophyll factors of which there are at least seven, as the writer has determined. Manifestly this is not a physiological correlation, for it is difficult to conceive of an anthocyanic pigment, limited to a single layer of cells in the grain, being caused by the same physiological factor that produces a plastid color, like chlorophyll, so widely distributed in the plant. Breed- ing evidence demonstrates that the correlation is genetic, as will be seen later. The aleurone factor ene is the R factor, which together with C and A is needed to produce any color in the aleurone cells of the corn grain (Emerson, 1917). To determine linkages with any of the five aleurone factors, it is obviously desirable to use material that is homo- 2 Reported in a paper given before the Botanical Society of America at the annual meeting in New York City, December 28, 1916, by Professor R. A. Emerson. The paper is now in manscript sing under the title, ‘‘A Fifth Pair of Factors for Aleurone Color in Maize. No. 604] LINKAGE IN MAIZE 227 zygous for as many of the non-linked factors as possible. Fortunately, the plants used in this experiment were homozygous for A, P and i, as will be shown later. This left only C and R, giving either 3:1 or 9:7 ratios of purple to colorless grains. The aleurone segregation on the ear proved to be exceptionally distinct because the factors A and P were homozygous. The other factor involved in the linkage is concerned with chlorophyll development in the mature plant. It is one of at least seven factors necessary for the production of full, normal green color in maize, and it has been termed the G factor. Its allelomorph g produces a distinct yellow or golden color in the leaves and stalk of the mature corn plant. This color is comparable with that of the familiar golden-leaved shrubs. It has been described and its in- heritance discussed by Emerson (1912) and Miles (1915). Suffice it to say that it is a simple recessive to normal green. During the summer of 1914, a green plant heterozygous for R and G was pollinated by a golden plant that lacked aleurone color. The cross can best be described by the following factors: GgRrCCAAPPw _, ggrrCcAAPPu 3472 (11) e A > ~ Proof for the correctness of these formule will be given later. _ A selfed ear of the female parent bore a 3:1 ratio (271:88) of purple to colorless grains, showing that only one aleurone factor was heterozygous. The male parent was likewise selfed and showed no aleurone color. On the ear of this cross there were 67 purple and 55 colorless grains, approximating a 1:1 ratio. The purple seeds on this ear were planted separately from the colorless seeds. From the field counts of this planting, which are given in Table I, it is seen that the 1:1:1:1 ratio of independent inheritance is noticeably modified. The observed numbers in the four classes give a ga- 228 THE AMERICAN NATURALIST [ Vou. LI metic ratio of 3.9:1, or 20.4 per cent. of crossovers. It is evident that the actual results agree very closely with the theoretical expectancy on a 4:1 basis of linkage. Indeed, the goodness of fit (Elderton, 1901 and Harris, 1912) is so perfect that «?—.317, giving a very high value for P. TABLE I $ SHOWING THE F, DISTRIBUTION FROM THE CROSS GgRrCC _ggrrCe 3472 (11) ^` 3468 (10)'` From Purple Seeds From Coiorless Seeds | Green oe | Golden (Rg) | Green (rG) ‘cok Golden esd WA WORE A a | pD T Observed, corrected for aleurone ratio? 33, Lah S4 6517; 300 Theoretical, on a 4:1 gametic ratio.. 29.2 lo ee 7.3 29.2 From these results it may be said that, at gametogen- esis, the factors of the female parent are so linked that the gametes RG and rg are produced about four times as often as the crossover gametes Rg and rG. Thus far, no _ evidence is available for demonstrating the linkage of Rg and rG (‘‘repulsion’’ between R and G Additional evidence on the linkage was afforded when the female parent of the cross and four other plants, in which the R and G factors were heterozygous, were selfed. Each of these plants bore ears with 3: 1 ratios in aleurone color (total, 1052 purple: 339 colorless). The results ap- pear in Table IT. Obviously this F, distribution does not resemble a 9:3:3:1 ratio of independent inheritance. The agree- ment between the actual results and the theoretical on a 4:1 gametic ratio, however, is close, P being .6733, which is considered a good fit. It may be concluded, therefore, that the factors in these five F, plants were so linked that R and G occurred together in one chromosome, while r and g were located in the homologous chromosome and 8 This correction is applied to equalize the proportion of the plants from the purple and the colorless seeds. The proportion should be 1: 1 in this ease, but variation in the percentage of germination or in the number of purple and colorless seeds actually planted, often disturbs it. Such a cor- rection is of course legitimate No. 604] LINKAGE IN MAIZE 229 that crossing over occurred about twenty per cent. of the time. TABLE II SHOWING THE F, DISTRIBUTION WHEN F, PLANTS HETEROZYGOUS FOR R AND WERE SELFED | From Purple Seeds F Colorless Seeds Selfed Plants | i EET | RG it? ee rG | rg OHS A ONDE Bie PO Pong 12. | 9 3347 (2) parent of above | 3 3 8 | 6 3954 (1) heer wens. 0 gi toe 3954 (3) Bite ocho 2 3 3954 (11). E E hts hota Sat Be Oe | 8 | 3 2 6 Total..... | 106 10 26 41 Correction for E Peta Miota EY a 248 | | 28 Theoretical, 4:1 basis........... fice ae b A208 Gh 16 16.5- | -: 29.2 When plants that were heterozygous for G and for ‘both R and C were selfed, the ratio between the four classes in the next generation appeared rather unusual. This was found to be due to the influence of the 9:7 ratio in aleurone color of each of the four plants that were selfed (total, 477 purple: 387 colorless). The data from such plants are arranged in Table III, which follows: _ TABLE III SHowine F, DISTRIBUTION WHEN PLANTS HETEROZYGOUS FOR G, R AND C WERE SELFED From Purple Seeds From C Seeds Plants Selfed Green Golden Green Golden noa hs a 32 3 il 4 3347 (3) 79 6 5l 31 3347 (4) 96 6 47 40 3477 (2) 38 nA 19 11 Total 245 18 128 86 Theoretical 4:1 basis 236 32 122 87 _ There is considerable deviation from the theoretical ex- pectancy in this case and the value for P is small (P= .0808). Nevertheless, these results, taken in con- * The correction is applied because at planting time no attempt was made ~ plant three purple grains to one colorless grain. 230 THE AMERICAN NATURALIST [ Vou. LI junction with those in Tables I and II, accord with the idea of linkage between R and G on a 4:1 basis, especially since they deviate so widely from the 27:9:21:7 ratio of independent inheritance. Another source of evidence on the genetic interrela- tions of the R and G factors was noted in a back cross in which the J factor for aleurone color was involved. No aleurone tests have been made, but evidence from related plants in pedigree cultures makes it seem reasonable that the following factors are concerned: GgRrecAAPPIi s 9gttCcAAPPu 021 (1) SHE 0) In this case, the ears of both parents showed no aleurone color. The ear from the cross gave a distinct segregation of purple and of colorless grains (36 purple:195 color- less), approximating the theoretical 1:7 ratio. Assuming the factors as given above, this aleurone ratio is reason- ably close to the expected proportion, the numbers being relatively small. As a check upon this aleurone ratio, it might be mentioned that six of the F, plans from purple seed were selfed. Each one showed a 9:7 ratio of purple ‘to colorless grains on the ear (total, 1441 purple: 1088 colorless). Also three F, plants from colorless seed were selfed and the ears showed no aleurone color. The field counts of the plants from purple and from colorless F, grains are classified in the following table: TABLE IV SHOWING THE F, DISTRIBUTION FROM THE CROSS GgRrecAAPPIi _, ggrrCcA APPii 3021 (1) ~~ +3018 (10) From Purple Seeds | From Colorless Seeds | | Green | Golden | Green | Golden ao / | Observed... | 12.0 3. | 22.0 Correctio n for kienrone ratio .. E 2." tl | 10.4 29.0 Theoretical, 4:1 ba | 4.5 fe Ft 18.0 21.4 Here again, while the value for P is relatively small (P=.1171), it is noted that in general the observed re- No. 604] LINKAGE IN MAIZE 231 sults correspond with the theoretical when a linkage on a 4:1 basis is assumed. Considering only the classes from purple seeds (first two columns in Table IV), the agree- ment is perfect. These classes are not disturbed by the aleurone situation and consequently ought to show the gametic ratio directly. DETERMINATION OF THE ALEURONE FACTOR CONCERNED IN THE LINKAGE It has been shown from four independent sources that the G factor is linked with one of the five aleurone factors, which has been termed the R factor. It now becomes nec- essary to prove that it really is the R factor that is in- volved. From the aleurone ratios observed in this ex- periment, às well as from evidence from related plants, it is certain that the inhibiting factor J is lacking, with the exception of the last cross described (Table IV). It is also obvious that the P factor, which produces purple color when C, R and A are present, must be homozygous (PP) in both parents of the first cross since the color in the colored grains is a deep purple with no trace of red. This leaves only C, R and A to be considered. In order to facilitate the presentation of the proof that the R factor is the one concerned, Table V has been pre- pared. In this table, the zygotic formule of the F, plants of the first back cross (see Table I) are listed and the plants that were tested are noted in the last column. The factors are merely assumed as here given. TABLE V SHOWING THE ZYGOTIC FoRMUL2 OF THE F, PLANTS OF THE Cross 3472 (11) X 3468 (10) Parental Formule: GgRrCCAA X ggrrCcAA Fı Zygotes F, Aleurone Color | F; Plant Type | Fi Pana aer mee a GgRrCCAA ...... Purple | Green (1) (3) (11) (81) (33) GgRrCeAA....... t A (34) ggRrCCAA....... * Golden geRrCeAA........ t t! (32) GgrrCCAA........ Colorless Green 44 GgrrCcAA........ ciel sie a ce ENE erbg Geen | fan) (22) (29) (48) (45) 232 THE AMERICAN NATURALIST [ Vou. LI The fourth column shows the plant numbers of the F, plants, which were self-fertilized. In addition, plants (31), (82) and (34) were crossed with certain aleurone testers of known constitution provided by Professor R. A. Emerson. All the grains on the ears of plants (21), (22), (29), (43), (45) and (44) were colorless. The aleurone counts on the ears from purple seed are arranegd in Table VI. TABLE VI SHOWING THE ALEURONE CouNTS ON SELFED Ears OF F, PLANTS FROM THE Cross 3472 (11) X 3468 (10) 3:1 Ears 9:7 Ears Ped. 3954 Plant No. —— ae é Purple Colorless | Ped. 3954 Plant No. | Purple | Colorless (1) 230 83 (32 we pee 68h ise. ss 303 | 88 (34) sae krmo oe eee 248 80 | | OE eee OR ay 180 45 | | Se 224. | 79 | | Potabeiassiyi 1,185 | 375 Totais | 238 | 228 Theoretical....... 1,17 390 |Theoretical....... 262 | 204 These ratios indicate that plants (1), (3), (11), (31) and (33) are heterozygous for one aleurone factor only. Plants (32) and (34) also contain that factor but appar- ~ ently they have in addition another aleurone factor, which likewise is heterozygous because the aleurone ratios, al- though they deviate somewhat from the 9:7 proportion, can not be classified as 3:1 or 27:37 ratios. Only tests with plants of known aleurone formule will determine whether it is the C, R, or A factor that is linked with G. Such tests have been made by using the aleurone testers described by Emerson (1917). These aleurone testers possess colorless grains in which all the aleurone factors except P are homozygous. For example, the R tester has the formula rrCC A Ait, the C tester the formula RRecA Ati, and the A tester the formula RRCCaati. In Table VII are presented the data involving the various tests for the aleurone factors. Plants (31), (32) and (34) possess all the three aleurone factors in question No. 604] LINKAGE IN MAIZE 233 (R, C, and A), as can be seen from their records in Table VI. Plant (41) is a golden type from colorless seed. TABLE VII SHOWING THE TESTS FOR THE PRESENCE OF ALEURONE FACTORS Crosses with Aleurone Testers Aleurone Color of F; Grains L A tester (RRCCaa) X 3954 (32) ............. 450 purple 2. C tester (RRccAA) X 3954 (32) F.. AR. rs T2 “z 69 colorless 3. E tester (rrCCAA) X 3954 (IZER ER TEKAN " 4. 3954 (34) X C tester (RRccAA) ......:...... B0Gar**a EAT VETS es 5. 3954 (31) X E tester (rrCCAA) -iiai p e ss BAG a 6. 3954 (41) (rrCcAA) X 3954 COIF ocijeni r : 101 e: The data in Table VII prove several things, namely: 1. That the factors A, C and R must be present in all the plants tested, because color was produced in some of the F, grains of each cross. 2. That it is the R factor that is linked with G. This is demonstrated in the fifth cross. Plant (31) was a green plant from purple seed bearing an ear with a 3:1 aleurone ratio, showing that it was heterozygous for the one aleu- rone factor that isinvolvedin the linkage. Thetestclearly | proves it to be the R factor, the aleurone ratio from the fifth cross approximating the theoretical 1:1 proportion. 3. That the assumptions made in Table V as to the- genotypic formulæ of the F, plants are correct. This statement follows from a series of deductions. First, one aleurone factor of the three concerned must be homo- zygous (dominant) in all the plants, because the aleurone ratios observed permit only two heterozygous factors at the most (9:7 ratios). Since the C factor is heterozy- gous in plants (32) and (34), and the R factor heterozy- gous in plant (31), while Æ is homozygous in plant (32), it is obvious that only the A factor could be homozygous (dominant) in all the F, plants. Second, one of the re- maining factors (C or R) must be homozygous (domi- nant) in some plants and heterozygous in others to ac- count for the 3:1 and the 9:7 ratios, respectively. The second and fourth crosses in Table VII indicate that C is heterozygous in plants (32) and (34), whereas the fifth 234 THE AMERICAN NATURALIST [ Vou. LI cross, together with the 3:1 ratio of the selfed ear, shows that C is homozygous in plant (31). Third, the remain- ing factor, R, should occur only in a heterozygous or ina homozygous recessive condition to account for the 1:1 aleurone ratio on the F, ear of the original back cross. The fifth cross in Table VII proves that R is heterozygous in plant (31), and the sixth cross shows that in plant (41) this same factor is homozygous recessive. From this series of interrelations it is seen that the hypothesis is verified in all cases and that it is the R factor for aleurone that is linked with G. On THE QUESTION oF CROSSING OVER IN THE MALE AND FEMALE An interesting observation regarding the question of crossing over in plants can be derived from some of the data presented. It has been shown that crossing over occurs in gametogenesis of the female (Table I). Does it take place in the formation of the male gametes as well? In certain animals, crossing over seems to be limited to one sex. It occurs only in the female of Drosophila (Mor- gan, 1915) and only in the male of the silkworm (Tanaka, 1914). Castle (1916) has noted that the phenomenon occurs in both sexes of the rat. Among plants, the studies with sweet peas and Primula indicate that crossing over is not restricted to one sex. Perhaps it is to be expected that, in the case of most plants, where the pistillate and staminate parts are borne on the same individual, there should be no difference in the genetic behavior of the two reproductive systems in this respect. Nevertheless, it is interesting to note the condition in the monecious corn plant. From the data in Table II, it is possible to demonstrate that crossing over is found in both sexes of maize. In order to do this, the observed frequencies can be com- pared with the theoretical expectation when crossing over occurs in both sexes and when it takes place only in the female. Such a comparison is arranged in Table VIII, which follows: ` No. 604] LINKAGE IN MAIZE 235 TABLE VIII A COMPARISON OF THE RESULTS FROM TABLE II WITH THE THEORETICAL EXPECTATION WHEN CROSSING OVER OCCURS IN BOTH SEXES D WHEN IT OCCURS ONLY IN THE FEMALE | | g no | ty | E [Coen | with Observed Observed (from Webie try. 2 (125 112 |18 | 298 | Theoretical, when crossing over oc- | | curs in both sexes 120.8 | 16.5 | 16.5 | 29.2 | .6733 Theoretical, when crossing over oc- | _ curs only i in the female 128;0| 9.0 |. 9.0. | 37.0 | .0067 Clearly; the first theoretical expectation fits the case adequately, for with such a high value for P it is almost certain that the deviations of the observed results from the theoretical are due to errors of random sampling only. Consequently one is justified in saying that crossing over occurs in both male and female. This is especially true when the fit in the second case is so poor. In fact, the great difference between the two values for P makes it seem reasonable that the intensity of the linkage is equal in both male and female, although the high value for P in the first case suggests that directly. ÅDDITIONAL LINKAGES BETWEEN CHLOROPHYLL Factors AND ALEURONE Preliminary tests indicate that the same chlorophyll factor, G, that is linked with R, is also concerned in a linkage with one of the other chlorophyll factors termed L. The latter has been found to be one of three factors, two of which have already been described by Miles (1915), involved in the production of chlorophyll in the seedling stage of maize. — Factors G and L seem to be linked although the data from three back crosses, in which the numbers are small, exhibit some variation in the percentage of crossing over. Details of this linkage will appear in a later paper, deal- ing with the inheritance of the three seedling chlorophyll factors. Apparently, then, the factor pairs Rr, Gg, and Ll con- 236 THE. AMERICAN NATURALIST [ Von. LI stitute one factorial group in maize. It is to be expected that Rr and Ll should bear a definite relationship to one another. This has not yet been fully determined, although there are some indications of such a linkage, for aleurone color and chlorophyll development appear to_be genet- ically related in different manner from that noted pre- viously. When purple seed of certain ears are planted sepa- rately from the colorless ones, the former give a distinct segregation of green and white seedlings, while the latter give rise not only to the green and white, but also to a constant proportion of yellow seedlings. The writer has determined that these yellow seedlings depend upon a definite genetic factor. Over two thousand seedlings have now been grown and not one yellow seedling has resulted from the purple grains. Discussion of this linkage will also be reserved for the later paper, or until the aleurone factor concerned has been identified. This may be the R factor showing its theoretical relationship to the L factor. - UMMARY 1. Linkage between the R aleurone factor and the G- factor for chlorophyll development shows approximately 20 per cent. crossovers. , 2, Crossing over takes place in both male and female gametogenesis of the monecious maize plant. 3. Preliminary tests indicate that G is also linked with L, a seedling, chlorophyll factor. Consequently the factor pairs Rr, Gg and Ll constitute one factorial group in maize. To Professor R. A. Emerson, of Cornell University, who so generously has shared his material for this inves- tigation, the writer is deeply indebted and desires to ex- press his sincere gratitude. LITERATURE CITED Castle, W. E. 1916. Further Studies of Piebald Rats and Selection, with Observa- tions on Gametic Coupling. Part III of Carnegie Inst. Wash, Pub. No. 241, pp. 175-180. No. 604] ‘ LINKAGE IN MAIZE 237 Collins, G. N. & Kempton, J. 1911. Inheritance of aoe Endosperm in Hybrids of Chinese Maize. In IV Conf. Internat. acaba Paris. Compt. Rend. et Raps., 1911, pp. 347-357 Colin G. N. 1912. Gametie Coupling as a Cause of Correlations. AMER. NAT., 46: 569-590 Collins, G. N. 1916. gri Characters in Maize Predial Journ. Agr. Re- rch, 6: 435-453, Plates 55-63. Elderton, W.P 1901. Tables for Testing the Goodness of Fit of Theory and Observa- tion. Biometrika, 1: 155-163. Emerson, R. A. 1911. Genetic Correlation and Spurious Allelomorphism in Maize. In Nebr. A xpt. Sta. 24th Ann. Rept., 1910, pp. 59-90. 1912. The Pilian of Certain Forms of Chlorophyl! Reduction in aves. In Nebr. Agr. Expt. Sta. 25th Ann. Rept., pp. 89-105. 1917. A Fifth Pair of Factors for Aleurone Color in Maize. In manuscrip Harris, J. A. 1912. A Simple Test of the Goodness of Fit of Mendelian Ratios. R MER. NAT., 46: 741-745. Miles, F. C. 1915. A ATONE and Cytologie cal Study of Sear: Types of Albinism 3 . Genetics, 4: 193-214. Plate VII ie T 3 Sturtevant, wer H., Muller, H. J., and pien Cc. B. 5. e Mechanism of Mendelian Heredity, pp. 1-262. incites Hy 1914. Further she on the Reduplication in Silkworms. Journ. Col. Agr. Tohoku Imp. Univ. Sapporo, 6 16. Webber, H. J. 1906. Correlation of Characters in Sa Breeding. Amer, Breeders’ Assoc. Proc., 2: 73-83, Pl. SHORTER ARTICLES AND DISCUSSION THE APPLICATION OF CORRELATION FORMULA TO THE PROBLEM OF VARIETAL DIFFERENCES IN DISEASE RESISTANCE: DATA FROM THE VERMONT EXPERIMENTS WITH POTATOES THE ultimate practical object of any study of disease resistance in a series of varieties is the selection for future cultivation of the few which are least susceptible. In practise a relatively large series of varieties or strains is taken into cultivation for pre- liminary study. The size of the cultures of the individual strains must, on a given area, be inversely proportional to their number. Since the individual cultures are necessarily small, it is impos- sible to assert from the results of a single test that the observed differences between the strains really represent varietal differ- ences in disease resistance. They may be due merely to inade- quately large cultures or to imperfectly controlled experimental conditions. It is therefore necessary to repeat the experiment another year or in a different locality in order to determine whether the observed differences are really persistent, and so characteristic of the strain, or whether they are due to transient conditions only. The problem is then purely and simply one of correlation. This is obviously true whether one chooses to avail himself of the advantages of the statistical formule or not. If the correlation between disease incidence in cultures of the series of varieties grown in different years, or places, be zero, the vari- eties show no permanent differentiation in disease resistance. If the correlation has a significant positive value it indicates at once that there are réally inherent varietal differences in disease re- sistance. The numerical magnitude of the correlation indicates something of the extent of this differentiation. If the correla- tion be low, the prospect of isolating varieties sensibly more re- sistant that the average will be slight. If the correlation be high, it should be relatively easy to secure highly resistant strains. Since the correlation method seems to have considerable value in the analysis of data of this'kind, I have thought it might be of service to geneticists and plant pathologists to illustrate it by the constants which I have found it necessary to deduce for another purpose from the published records of the series of experiments on disease resistance in varieties of potatoes carried on during 1 Stuart, W., ‘‘Disease Resistance in Potatoes.’’ Bull. Vt. Agr. Exp. Sta., 179, 1914. 238 No. 604] SHORTER ARTICLES AND DISCUSSIONS 239 the past several years at the Vermont Agricultural Experiment Station, In a recent bulletin Stuart? summarizes the data obtained dur- ing five years’ observations on percentage infection by early blight (Alternaria solani). In his Table I he gives the estimated percentage infection in a series of varieties during the period. Since during a portion of the experiment all the varieties were not considered, I have calculated the correlations in two groups. In one case N=149, in the other N=50. The smaller group comprises only varieties also included in the larger. The corre- lations, calculated by the usual product moment method? with- out grouping, appear in the accompanying table.® CORRELATION FOR VINE RESISTANCE TO EARLY BLIGHT IN Two YEARS Years Compared Series of 149 Varieties Series of 50 Varieties 1905-1906 ...... + .055 + .055 — .056 + .095 1905-1907 ...... + .438 + .045 + .420 + .079 1905-1908 ...... — .021 + .095 1906-1907 i .... + .042 + .055 + .226 + .091 1906-1908 ...... + .323 + .085 1907-1908 ...... —— + .082 + .095 Only 2 of the 9 constants are negative; these are insignificant in comparison with their probable errors. All the constants which have substantial values and are materially larger than their prob- able errors are positive in sign. The average of the two negative constants is —.038, of the seven positive coefficients + .227, and of all the (unweighted) values + .168. Thus there is clearly a measurable differentiation of the varieties in respect to suscep- tibility to Alternaria. The values are, however, exceedingly variable, ranging as they do from —.056 to + .438. The great variation in the actual constants I am inclined to attribute to (a) the difficulty of esti- mating the percentage of infection, (b) the unavoidable experi- mental errors associated with relatively small cultures, and (c) the wide variation in average percentage infection from year to year. Both (a) and (b) are factors which tend to render the actually recorded percentages somewhat erroneous as measures of the real susceptibility of the variety, and tend in consequence to dilute the strength of the correlation. With respect to the third 2 AMER. NAT., 44: 693-699, 1910. 3 The chief discrepancy between the results for the larger (N =149) and the smaller (N = 50) series of varieties is to be seen in the interrelationship for 1906 and 1907 where the two correlations are .042 + .055 and .226 + .091, Here the disagreement is apparent rather than real. The difference is -184 + .106, which can not be considered statistically trustworthy. 240 THE AMERICAN NATURALIST [ Von. LI factor, (c), it is obvious that if there be only a very slight aver- age percentage infection the test of disease resistance will not be a very critical one, whereas the average percentage can not be very high indeed unless conditions are so unfavorable that all varieties are affected. The percentage infection of early blight varies enor- -mously from year to year. Thus: Percentage Infection Year N = 149 N = 50 PD 5 eile lineal ah 3.3 2.4 PONG sce rials krai 85.0 83.2 ADDL Teorainn] 41.7 38.9 TOOG IE y yes — 10.8 With an incidence of 3 per cent. one year and of 85 per cent. the following season one can, in view of the considerations men- tioned above, hardly expect to obtain smooth values of the corre- lation coefficient. It is interesting to compare these results with those for other maladies of the potato. In the same publication Stuart gives the results of trials for resistance of tubers to scab. Unfortu- nately the experiments of the second year, 1907, included only 20 of the 65 varieties from the first year. Calculations may be based on the percentage of tubers which are free or nearly free from scab. This is much lower the second year. ~ 1906 1907 ORR oy Zee ck ee 64.21 f 28.20 ED l oe n 11.59 16.33 COPPEIBTION “soa ees 7 = 591 + .098 The probable error is high because of the fewness of the vari- eties retained in the second year’s test, but the correlation is of more than medium value and is relatively about 6 times as large as its probable error. Thus susceptibility to seab is probably to a very considerable extent a varietal character. The results for tuber rot tests are not available for successive years, but Stuart has given* the percentage of tuber rot in 89 varieties grown on sandy loam and on clay loam soil in 1905. For these I find Sandy Loam Clay Loam MOAN noenee eens 8.68 39.76 Se TE EE 11.28 6.85 653 + .0: 4 Stuart, W., ‘‘Disease Resistance in Potatoes,’’ Bull. Vt, Agr. Exp. Sta. 122, Tables VI-VII, 1906. 5 The value of r given as .707 + .045 in Science, N. S., 38: 402-403, 1918, is deduced from the 62 varieties for which laboratory cultures were avail- able, and from modified percentages. The values agree within the limits of their probable errors. No. 604] SHORTER ARTICLES AND DISCUSSIONS 241 The correlation of these data, somewhat smoothed by Jones,’ with a series of determinations of the percentage growth of the fungus on tubers in the laboratory has already been determined.’ For laboratory growth and loss on clay loam, r==.584 + .059. For laboratory growth and loss on sandy loam, r= .594 + .055. Taken as a whole these correlations indicate (a) that suscep- tibility to both early and late blight and to scab differs greatly from variety to variety, and (b) that, so far as the evidence goes, the varieties differ more in resistance to tuber injury than to foliage infection by early blight. It is not at all necessary that the correlations be drawn be- tween the amount of injury to the same organs of the plant or by the same disease. In many instances the so-called cross cor- relations yield valuable results. For examplé Stuart® discusses the question of the relationship between vine infection and tuber rot. The point may be sub- jected to a statistical test by correlating between the maximum percentage of foliage affected by late blight as given in his Table V for potatoes grown on sandy loam soil in 1905 and per- centage of rot as recorded in his Tables VI and VII. Unfortu- nately, the percentages are available for the vines for sandy loam soil only (Table V) while the figures for tuber rot are given for both sandy loam and clay loam soil. Both correlations may be worked out. I fin For percentage Idia infection on sandy loam soil and per cent. tuber rot on sandy loam soil N=131, r—.316 + .053. For percentage foliage infection on sandy loam soil and per cent. tuber rot on clay loam soil N= 80; re=.102 + 075. In both cases the correlations are positive, and hence ii evi- dence as they furnish indicates that the varieties which show the greatest infection of the leaves actually are the worst to rot.. That the correlation between injury to the tops and tuber rot is higher on the sandy loam soil is not at all surprising, since the same individual plants—not merely the same varieties—are in- 6 Jones, L. R., N. J. Giddings and B. F. Lutman, ‘‘ Investigations of the Potato Fungus Phytophthora infestans,’’ Bull. Vt. Agr. Exp. Sta., 168 74-81, 1912. i Jones and collaborators, loc. cit., and the Reviewer, Science, N. By 38: 402-413, 1913. 8 Bull. Vermont Agr. Exp. Sta., 122, p. 116. 242 THE AMERICAN NATURALIST [Vou. LI volved in the correlation. The problem is, however, a compli- cated one and much more extensive data are needed for a com- plete analysis. A problem of very great biological interest as well as of prac- tical importance is that of the specificity of disease resistance. Concretely : Do varieties differ in their susceptibility to a specifie disease only, or do they differ merely in susceptibility to disease in general ? A comprehensive and final answer will require far more data than are available and more stringent statistical analysis than ean be illustrated here. Some progress can be made by the method of correlation as follows. If susceptibility be purely specific there should be no correla- tion between the incidence of disease x in year (or culture) p and disease y in year (or culture) q, although there should be a correlation between the incidence of disease x or disease y in different years or cultures. If, on the other hand, differences in disease resistances from variety to variety are determined solely by general weakness or vigor of the stocks, one should expect the correlations between the incidence of different diseases in dif- ferent years or cultures to be (within the limits fixed by the _ errors of measurement and the probable errors of random sam- pling) as high as those between two series of determinations of incidence of one and the same parasite. Consider first the relationship between the percentage of foliage injury by early blight in 1905, 1906, and 1907 and the percentage of tuber rot in 1905. The correlations are: Per Cent. Tuber Rot on Per Cent. Tuber Rot on Sandy Loam, 1905 Clay Loam, 1905 Foliage Injury N= N = 89 Early Blight, 1905.... .167 + .057 .256 + .057 Early Blight, 1906.... .211 + .056 .249 + .067 Early Blight, 1907.... .291 + .054 440 + .058 Without exception the correlations are positive in sign. While numerically low, the most of them taken individually may be considered statistically significant in comparison with their prob- able errors. Thus it seems clear that the varieties with foliage most injured by early blight are also most subject to tuber rot, just as has been -shown to be the case in foliage and tuber infection by late blight. For foliage infection by early blight in 1905, 1906, and 1907 and foliage injury by late blight in 1905 I find: No.604] SHORTER ARTICLES AND DISCUSSIONS 243 Correlation, N = 131 Early Blight, 1905, and Late Blight, 1905.. — .066 + .059 Early Blight, 1906, and Late Blight, 1905. dg. 190 + .057 Early Blight, 1907, and Late Blight, 1905.. — .040 + .059 The results are not so consistent as those of the preceding table. The two negative coefficients are insignificant in com- parison with their probable errors, and the positive one is not large, either absolutely or relatively. Possibly the laxness of the correlation is in part due to the fact that the measurement of both characters is subject to a large possible error.’ For freedom of the tubers from scab with the incidence of other diseases every possible correlation has been determined. The coefficients are shown in the accompanying table. Note that in this case the correlation is between freedom from one disease and occurrence of another disease. Hence a negative coefficient has the same meaning as a positive one in the foregoing discus- sions. For PERCENTAGE OF TUBERS FREE OR NEARLY FREE FROM SCAB IN 1906 AND 907 AND INCIDENCE OF OTHER DISEASES es Compared with Freedom fr cab Correlation 1906 Correlation 1970 Per Pen rot on Sandy Loam Soil, 1905 .. .. — .136 + .077 —.857 + .182 (N = 74) (N = 20) Per Cent. rot on Clay Loam Soil, 1905 .... — .280 + .096 — .030 + .187 (N = 42) (N = 13) Top Injury by Late Blight, 1905 ........ :+ .035 + .084 — .224 + .151 (N = 65) (N= 18) Top Injury by Early Blight, 1905 ...... — 561+ .054 + .240 + .142 (N = 74) (N = 20) Top Injury by Early Blight, 1906 ...... — .119 + .077 — .352 + .132 (N == 74) (N ==20) Top Injury by Early Blight, 1907 ...... — .118 + .077 — .040 + .151 (N = 74) (N = 20) Because of the small number of varieties involved and the roughness of the measurements the correlations are low and irregular. In ten cases the negative sign indicates that the varieties which are most free from scab are also least susceptible to attacks by other diseases. In neither of the two cases of posi- tive correlation is the constant statistically significant in com- parison with its probable error. Thus altogether 23 of these cross correlations—that is correla- tions between injury to different organs by the same disease, or 9 In one case, pe: the correlation is between foliage injury by two dif- ferent organisms in the same year. What interrelationship is to be ex- pected in such case requires further consideration. 244 THE AMERICAN NATURALIST [ Vou. LI to the same organ by different diseases, or to different organs by different diseases—have been worked out. Only 4 of these— that is only about one ease out of six—are exceptions to the rule that varieties which show more than the average amount of injury by one disease will, on the whole, show more than the _ average injury by another disease. No one of these exceptional constants can be considered significant with regard to its prob- able error. Several of the 19 which indicate the rule may be looked upon as individually trustworthy. Thus notwithstanding the large variations in numerical magnitude incident to small series of data and rough measurement, the determinations taken collectively certainly furnish highly convincing evidence that to a considerable extent susceptibility to disease is general rather than specific. The fact that the series of correlation coefficients herd pre- sented justify much more definite conclusions than those who have considered the data without statistical analysis have drawn, is sufficient indication of the usefulness of the biometric method in the preliminary stage of disease-resistance experiments in which large numbers of strains are being tested, and in which the mass of data is highly confusing. The special cases illus- trated by no means exhaust the possibilities of the biometric formule now available. Had the data been more extensive, the analysis might have been carried much farther. Nothing that has been said in this paper in emphasis of the statistical method must be taken to imply that the most careful individual analysis is not desirable and essential. The two methods are not mutually exclusive, but supplemental. J. ARTHUR HARRIS THE DIFFERENT MEANINGS OF THE TERM ‘‘FACTOR”’ AS AFFECTING CLEARNESS IN GENETIC DISCUSSION? IN the analysis of alternative (or segregating) heredity, we find that certain potentialities, such as that of producing a cer- tain color in some part of the soma, appear to be inherited inde- pendently of certain other potentialities. We assume that the germ-plasm carries various corresponding genes, factors, or deter- miners, whose independence in gametogenesis determines the in- dependence of the somatic characters. Cytological study leads 1 Paper No. 39, "n of California, Citrus Experiment Station, Riverside, California No. 604] SHORTER ARTICLES AND DISCUSSIONS 245 to some very probable conclusions as to the time and method of segregation of these determiners; though it does not yet enable us to identify with absolute certainty the actual physical unit of segregation,. the cytological and genetic evidence indicates strongly that such a unit exists. In the current chromosome hypothesis, as developed especially by Morgan and his collaborators (Morgan, Sturtevant, Muller, and Bridges, 1915), the material unit of segregation is assumed to be a part of achromosome. Breaks in two homologous chromo- somes at meiosis, with consequent exchange of parts by the pair, presumably occur at certain definite points only. How close these points may be we can not say, but the general stability of Men- delian characters indicates that the number of points is limited. On these assumptions, the portion of a chromosome between two adjacent points of possible breaking is the ultimate physical unit of genetic segregation—essentially a locus as defined by Morgan (1915, p. 419).? It is now widely recognized that, in effect, a single real unit of segregation may influence very diverse characters of the soma, often in ways which can not be at all inferred one from another. Very possibly one physical unit of segregation may affect, say, fiower color and height in ways just as distinct physiologically as may two distinct units of segregation, although transmission is different in the two cases. In the latter case we say that two genetic factors are concerned; are we compelled in the former case to admit only one? As a matter of fact, as will be evident on further considera- tion, either course is possible, according to the definition of factor accepted. If by ‘‘factor’’ we mean a developmental potentiality, the de- limitation of a particular factor is largely a matter of convenience in analysis. On the other hand, if the term is used to designate a supposed actual physical unit of segregation, a factor has a definite objective extent. The former view is that of the presence-and-absence terminol- ogy, as it is generally understood at present. In this sense, a factor is not an element of the germ-plasm; it is rather a prop- erty or characteristic of the germ-plasm or of some element of the germ-plasm. The characters of an organism, as Gates (1914, 2 Though Morgan, Sturtevant, Muller and Bridges (1915, p. 155) suggest the possibility that the loci of linked factors may be so near together in a chromosome ‘‘that they never (or very rarely) cross over.’’ The definition. of locus is discussed below. 246 THE AMERICAN NATURALIST [ Vou. LI p. 269) has remarked, are ‘‘attributes,’’ no more to be separated from the organism than are the properties of a chemical com- pound from that compound. The factors of the presence-and- absence scheme, similarly, are inferred properties or attributes of the germ-plasm, by whose behavior we explain the alternative transmission of certain properties or attributes of the soma. Obviously an organism is not composed of ‘‘characters’’—and neither is its germ-plasm composed of ‘‘factors,’’ so long as the factors are those of the presence-and-absence scheme. Such factors are nothing but characters of the germ-plasm, and, like the characters of the soma, they are more or less conventionalized in description. We have no warrant for projecting these con- ventionalized descriptions back into the actual germ-plasm, and assuming the presence and absence there of strictly correspond- ing material units of segregation. The presence-and-absence scheme, when not encumbered with non-essential hypotheses, is a strictly neutral instrument of genetic analysis. If there is segregation in the formation of the germ- cells, it is merely a matter of definition to state that a factor is allelomorphic to its absence. That is, the assumption of segrega- tion is the only assumption — by this scheme, which is the logically simplest form of the ‘‘conceptual notation’’ (East, 1912) of genetics. Very special emphasis must be placed on the fact that the “absence”? is absence of a potentiality, without reference to the presence or localization in the germ-plasm of any other poten- tiality that may actually take its place. The allelomorphism of the presence-and-absence notation is a logical opposition; when it makes ‘‘A’’ and ‘‘no-A’’ allelomorphs, this involves no as- sumption as to what may be physically opposed, in the chromo- somes, to the physical basis of “A.” If an ‘‘absence’’ a of a given factor A is always or commonly. accompanied by an actual presence of a corresponding factor A’, and we wish to represent this fact, it is provided for by the terminology of linkage; we may use Aa’ and aA’, since the presence-and-absence ‘scheme makes no assumptions as to the structure of the germ-plasm. It is obviously simpler to write simply A and A’, or A and a, for the two factors, and conveni- ence may justify this practise; we should note, however, that in thus abandoning the presence-and-absence terminology we intro- duce a second assumption, that of actual factor-to-factor opposi- tion or allelomorphism. This assumption is, of course, in view of all the evidence, a highly probable one, and especially convenient No. 604] SHORTER ARTICLES AND DISCUSSIONS 247 in cases of apparent multiple allelomorphism (Morgan, Muller Sturtevant, and Bridges, 1915, chap. 7). In fact, this added as- sumption probably permits a more direct and therefore prac- tically simpler representation of the actual course of segregation. Whether the corresponding factor-to-factor notations now used for Drosophila (Castle, 1913; Morgan, Sturtevant, Muller, and Bridges, 1915, p. 233) are everywhere adequate and convenient is another question, as Emerson (1913) has shown. Cases of multiple allelomorphism involve no special difficulty in principle for the presence-and-absence scheme. They can of course be represented only by linkage formule, in which the ‘*presence’’ of one factor of the set is linked with the absence of the rest—but their very occurrence suggests that we might, as is suggested above, correctly enough represent single factorial dif- ferences in the same way. All this can affect only the conveni- ence of the notation, and not at all its logical applicability. If we adopt a factor-to-factor system of notation, it is natural to conceive of the opposed ‘‘factors’’ not as mere potentialities, but as physical units responsible for genetic potentialities. When we have taken this viewpoint, we have begun to ‘use the word factor in the second sense mentioned above; we are thinking of assumed physical units of segregation, not merely of observed peienticlitios of development. Moeaa (1915, p. 419), in dis- cussing ‘‘presence and absence,’’ uses factor in this sense, as do Morgan, Sturtevant, Muller, and Bridges (1915, pp. 220-222). No doubt what has been said above is an old story to experi- enced geneticists in general, in view of such discussions as those of East (1912) and Morgan (1915). The distinction is so funda-\ mental, however, and the double use of the term factor so in- creases the difficulties of the case, that consideration of the gen- eral problem from the present terminological viewpoint seems highly desirable. Perhaps greater precision in the use of several terms could be attained. Johannsen (1909, pp. 124-125), in defining the term Gen (gene), makes it perfectly plain that he means the material basis? or cause (the Anlage), of whatever sort, of a ‘‘unit char- acter,’’ defining a unit character as one dependent on a special kind of gene. He is evidently inclined to consider the gene as the material unit of segregation, holding that the sum of the genes constitutes the germ-plasm, and there is a widely prevalent 3 Or immaterial basis, if we must admit the theoretical possibility of the existence of immaterial ‘‘entelechies’’ associated in some more or less mys- tical way with the germ-plasm. ; 248 THE AMERICAN NATURALIST [ Vou. LI tendency among geneticists to use gene, and its synonyms factor and determiner, in this sense, which is the second of the two dis- cussed above. Evidently gene is not properly used in the first sense. A moment’s thought, however, will show the impractica- bility of confining factor to that sense; the meaning of this term shifts back and forth continually in common usage, and often remains indefinite. Further, when factor (or gene) is used in the second sense, we consider it coextensive with locus (Morgan, 1915, pp. 419-20; Goodspeed and Clausen, 1917, p. 32). A factor is a particular state or condition of a locus. Let us, then, define locus as the physical unit of segregation, almost certainly identified as a genetically indivisible portion of a chromosome. Genetically in- separable (‘‘completely linked’’) potentialities, then, belong to the same locus, and hence to the same factor or gene; ‘‘completely linked factors’’ are mainly‘ relegated to non-cytological discus- sion, and especially to use with the presence-and-absence ter- minology. No doubt Mendelian analysis considers in any case, only some of the most readily identifiable properties of the real units of segregation concerned, and this fact seems to deserve a large place in our genetic thinking. Especially is it important that the two meanings of factor and its synonyms should be clearly distinguished ; when these meanings are unconsciously inter- changed and confused, vagueness and misunderstanding are sure to result. The student of genetics may read, for example (East, 1912), If we forget ourselves and begin to speak of unit factors as particles, only a confusion follows similar to that caused by Nägeli, Spencer, and Weismann. Nothing is gained and even facts are obscured. On the other hand, he will find the factors of Drosophila located with mathematical exactness in diagrams of the chromo- somes, and often apparently or explicitly considered as material components of the chromosomes. In the interest of clear think- ing, especially in the case of beginners and casual students in the field of genetics, the explanation of this apparent contradiction deserves very special emphasis. As an example of the way in which this terminological con- flict may cloud an argument when the essential facts are clear to the writer, we may take the following case. Morgan (1913, ra n temporarily complete linkage of genes must be excepted, as in the usual case with the male of Drosophila. . The point is that the chromo- some naa pai well consider, in any case, that potentialities s awaye asso- ciated are manifestations of the same factor or gene. No.604] SHORTER ARTICLES AND DISCUSSIONS 249 p. 10), in urging his substitute for the presence-and-absence scheme, largely on grounds of convenience, agrees with East (1912) as to the general usability of the latter scheme as a ‘‘system of nomenclature’’ without cytological implications. He makes especially plain the undesirability of interpreting ‘‘ab- Sence’’ as a physical absence in the germ-cell, or (1915, p. 419) as “‘a hole in a chromosome.’’ In one respect, however, Morgan’s discussion seems less clear than it might be, and this is in the use of factor, in these articles, in a sense (the second here) which is not that of the presence-and-absence scheme, with only vaguely implied explanation of the distinction. It certainly is permissible to speak of ‘‘the absence of a factor from the germ-plasm,”’ if we mean the kind of ‘‘factor’’ implied by the presence-and- absence terminology. We must make it as clear as possible that factor (1) sometimes means a potentiality and (2) sometimes means a body, and that a factor is assumed to be paired either (3) with its absence or (4) with another (identical or different) factor. The combination of (1) and (3), then, gives the presence-and-absence scheme, while the combination of either (1) or (2) with (4) gives the scheme used by Morgan (1913; see also Castle, 1913) and other students of Drosophila. General objections to Mendelian analysis have been based largely on confusion of (1) and (2), which often leads to erroneous suppositions—for instance, that Mendelian analysis in general, or the presence-and-absence method of Men- delian analysis, requires the unnecessary and unwarranted as- sumption involved im the combination of (2) and (3). Prob- ably this confusion is also largely responsible for the persistence of another often discredited notion, the idea that ‘‘Mendelians”’ suppose their factors to be individually the basis of somatie char- acters, rather than simply necessary elements in an interacting complex which produces the characters. i SUMMARY The term factor has, in genetic use, two distinct meanings, which are continually interchanged or combined and often con- fused. It is essential to clearness in genetic discussion that these two meanings should be carefully distinguished. These mean- ings may be indicated by the following formal definitions: 1. A genetic (Mendelian) factor is a property or characteristic of the germ-plasm, more or less conveniently delimited for the purpose of analysis of segregating heredity. : 2. A genetic (Mendelian) factor, or gene, is an actual material 250 THE AMERICAN NATURALIST [Vou. LI unit of genetic segregation ; it is of unknown nature, but probably consists of a genetically indivisible portion of a chromosome (a locus) in a particular state. The presence-and-absence scheme of factor notation properly employs only the first of these meanings; the Morgan-Castle scheme, on the other hand, may use either. Howard B. Frost CITRUS EXPERIMENT STATION, RIVERSIDE, CALIF. BIBLIOGRAPHY Castle, hiran E. 1913. Simplification of Mendelian formulæ. Am. NAT., 47: 170-182. East, sansa M, 1912. The sagged Rp oie as a Description of ee Facts. Am. NAT., 46: 5. Emerson, py A. 1913. Simplified ea seta Formule. Am. NAT., 47: 307 a R. hia 1915. The Mutation Factor in Pr "n Par- icular Reference to Œnothera. 14 + 353 A ve Macm Pe Thomas H., and Clausen, R. E. T: endelian rare Dif- ferences versus diceateatenenbins ii in AAN Am. Nar., 51: 31-46, 92-101. Johannsen, W. 1909, Elemente der exakten Erblichkeitslehre. 6 + 516 p. n h Morgan, Thomas H. "1913. Factors and Unit Characters in Mendelian He- edity. Am. NAT., 47: 5-16 1915. The Bois of the Birra in the PE a s Sex-linked Mendelian Character in Drosophila. Am. NAT : 385-4 a Thomas H., Sturtevant, A. H., Muller, H. i ne Pres, Calvin B. The rekaan of Wondaiai ‘Heredity. 13 4+ 262 p. New Te Henry Holt & Co. THE SELECTION PROBLEM Untess history fails to repeat itself, geneticists, whose atten- tion is focused upon variation, sho sooner or later overem- phasize its impoftance as a factor in shaping the organic complex. There is, indeed, reason to believe that already a tendency for some among them to do so is becoming apparent. Dr. Pearl’s* recent paper under the title above affords an example in point. In that communication its author fails to discriminate sharply between two distinct phases of his subject. Whether selection may affect the course of evolution is a matter entirely apart from the possibility that it alters the germ plasm. Racial history may possibly be modified, if the genetic composition of a mixed population may be affected by selection based upon somatie dif- 1 Pearl, Raymond, 1917, ‘‘The Selection Problem,’’ THE AMERICAN NAT- URALIST, Vol 51, pp. 65-91. No.604] SHORTER ARTICLES AND DISCUSSIONS 251 ferences. But modification of the germ plasm by selection is impossible, if that agency acts only as ‘‘a mechanical sorter of existing diversities.’’ The present note is not concerned with the causes of variation. It refers only to the first-mentioned phase of the selection prob- lem. It accepts the statements in the quotation below as sub- stantially correct, and attempts in brief compass to evaluate the arguments by which Dr. Pearl supports his position regarding them. By transposition of a few phrases his ideas may be expressed in his own words as follows: The mere fact of elimination and survival . . . is capable, in theory at least, of bringing about evolutionary pete of a progressive sort, . if the elimination be selective, and the survivors transmit to their progeny those differences that mark them off from the eliminated. The theory that these two rules are always and everywhere in opera- tion, taken together with the observed fact that living creatures do die, is the Darwinian theory of Natural Selection as a factor in organie evolution. If, as is implied, Darwin gratuitously assumed the intolerable burden involved in the use of the words, always and everywhere, it is immaterial. It is not a vital issue whether the form in which he expressed himself will bear literal interpretation, but merely whether his idea is correct that natural selection effects notable changes in the course of evolution. Hence it seems suffi- cient to say, that if the ‘‘Dance of Death’’ is governed in gen- eral, or even in part, by the joint action of the two principles enunciated in the preceding paragraph, the changes described there should follow as surely, although more slowly, than if the conformity were complete. It is stated by Dr. Pearl, as one of three broad facts on ac- count of which natural selection is no longer regarded as a primary, or perhaps even a major factor in evolution, that even when selective elimination on the basis of somatie characters does occur, it does not follow generally and regularly that the somatic differences on which the selection acted will reappear in the progeny, . . . actual experience having abundantly demon- strated that a very great many of such somatic differences are not inherited. This may refer, first, to the fact that a single phenotype may include members of different genotypes. Yet, even so, no strict limitation is placed upon the possibility of changing the char- acter of a mixed population by selection based upon somatic qualities. If plus variants of an inferior strain seem superior to 252 THE AMERICAN NATURALIST [ Vou. LI individuals of a higher order of genetic worth, this will have its due effect in impeding progress; but since the essential point is, that, collectively considered, members of different pure lines do, in general, rather definitely reflect their different germinal constitution, advance may undoubtedly be made. However, it is unnecessary to labor the point, as it is not only admitted, but is urged by geneticists themselves in explanation of such results, for example, as those Castle and Phillips? obtained in selection ex- periments with hooded rats. Reference to the genetic behavior of such characters as side- sprig of the comb in poultry may also be included in the quota- tion above. If, in this case, emphasis is laid upon the fact that selection might apparently be exercised indefinitely without the least tendency toward evolution of a side-sprigged race, it needs only be pointed out that the argument is directed against the contention that selection is capable of modifying the germ plasm. It has no bearing whatever upon the possibility of its occurrence in nature, nor, aside from the point indicated, upon its influence in evolution. Upon the other hand it seems quite impossible that even if characters of which side-sprig is representative were highly useful, they should fundamentally modify the course of evolution under selection, or place any notable obstacle in its way, for they are not germinal variations and are as likely to occur in one line of descent as another. Hence, other things being equal, it is quite as probable that they should assure the survival of a subnormal representative of a superior genotype, as that they should tide over a superior representative of an inferior one; and the chance would be no greater that the attribute in question should appear in the offspring of one rather than in that of the other. A second of the three general groups of facts to which refer- ence has been made is summarized in the statement: Observation indicates that in many cases evolutionary changes have come about by relatively large, discontinuous steps, the new form being not merely fully differentiated at its first appearance, but also fully able to survive. In another connection it is stated forcefully by Dr. Pearl that if the game of survival is actually played by the quoted rules he formulates (and no others are necessary) the conclusion is logically irresistible that progress is bound to occur in the direc- tion of those differences which distinguish the survivors. But 2 Castle, W. E., and Phillips, John C., 1914, Carnegie Institution of Wash- ington, Publication 49. Castle, W. E., 1916, ‘Genetics and EES ” Har- vard University Press, Cambridge, 8vo, pp. vi + 353. No.604] SHORTER ARTICLES AND DISCUSSIONS 253 since obedience to his two rules is not in the least contingent upon the magnitude of the variations upon which selection is based, it must be admitted that the facts summarized above are entirely irrelevant to the present discussion. They neither bear upon the phenomena of inheritance, nor add anything to our knowledge of selective elimination. The third general fact cited in support of Dr. Pearl’s ries y tion regarding the diminished esteem in which natural selection is held as a factor in evolution is: All organisms possess in varying, but usually in very large, degree the power of personal, immediate, individual, somatic adaptation to the environment. It is affirmed in addition, that in consequence of this power of personal adaptation the survival expectation of an individual is not generally and regularly a function of any static, single- valued relation between its somatic structure, habits or physiol- ogy, on the one hand, and the impinging environmental stresses on the other. Yet, it is asserted, such a relation is implicitly assumed in that part of the theory of natural selection which affirms a ‘selective elimination on the basis of somatie char- acteristics. The reply to these various statements is, that their substantial truth may be admitted without the possibility that evolution is affected by selective elimination being thereby in the least diminished. The adaptive capability of any individual either rests upon or lacks a germinal basis. In the former case there is no obvious reason why it should not itself provide material upon which selective elimination might be based, with consequent change in the composition of the population. In the latter, individual adaptability is as incapable of exercising influence upon the course of evolution, as side-sprig should be, if it were a useful character of the same order of importance. If the accumulated results of genetic research provide no more effective arguments than these, it must remain an open question whether natural selection is not a primary factor, not in the origin of species, but in the determination of the elements composing the flora and fauna of the world at any period in its history. In other words, one who wishes to force an abandon- ment of that position must demonstrate that selective elimina- tion does not occur upon such a scale that it may account for the results ascribed to it. For many legitimate reasons Dr. Pearl has not treated this point at length in his article under discussion. But from scanty evidence available he derives the following conclusion : 254 THE AMERICAN NATURALIST [ Von. LI In some cases natural elimination is certainly in some degree selec- tive, while in other cases it certainly is not, and in the most favorable eases of all the selection is apparently not very rigorous. Gross terato- logical abnormalities are eliminated. But the smaller deviations from type, which in theory ought to furnish the basis of selection, appear upon quantitative study less generally and sharply determinative of survival than might have reasonably been expected theoretically. It is perhaps worthy of note that ever since it appeared that the larger, and rarer, discontinuous variations are in no danger of being lost through swamping, it has been beside the mark to ascribe especial theoretical significance to smaller deviations from type. Attention may also be directed to the important admis- sion that in some instances elimination is known to be selective. It will then be in order to examine reports of researches upon the basis of which it is confidently asserted that in other cases the same is not true. It happens that with five others one is cited which falls squarely within my own field of investigation. This is Pro- fessor Reighard’s* ‘‘Experimental Field Study of Warning Coloration in Coral-reef Fishes,” which Dr. Pearl seems to consider of particular significance for his own argument. But since it has been my good fortune to study the same material in the same place, more extensively and with better facilities, I feel justified in saying that Reighard’s results will bear no such interpretation as is here placed upon them. He proved that gray snappers possess powers of discrimination and of memory which would lead one to suppose that, if the bright colors of the smaller reef fishes possess a warning significance, the snappers’ should be aware of it and avoid them. He did not prove beyond possibility of doubt that they do attack such fishes freely, but that is of no importance in the present connection; his con- clusion that tropical fishes are not warningly colored with refer- ence to their commonest enemies is perfectly sound. His ideas regarding immunity coloration are, however, the only ones in his paper which have the remotest bearing upon the matter of selective elimination, and there is no reason to suppose that these are more than logical deductions from incorrect premises. Dr. Pearl’s* own report upon the natural elimination suffered 3 Reighard, Jacob, 1908, ‘‘ An Experimental Field-study of Warning Col- oration in Coral-reef Fishes,’’ Carnegie Institution of Washington, Papers from the Tortugas Laboratory, Vol. 2, pp. 257-325. 4Pearl, Raymond, 1911, ‘‘Data on the Relative Conspicuousness of Barred and Self-colored Fowls,’’ THE AMERICAN NATURALIST, Vol. 45, pp. 107-117. No.604] SHORTER ARTICLES AND DISCUSSIONS 255 by barred (Plymouth Rock) and black, or near-black, chickens living under the same conditions is also mentioned. In this paper photographs show the black fowls looming up against natural backgrounds much more distinctly than do the barred. This is accepted as ‘‘objective and unbiased evidence regarding the relative conspicuousness of the two types of plumage pattern.’’ It follows naturally, since the extensive record shows little difference in the rate of elimination of the two sorts of birds, that ‘‘the relative inconspicuousness of the barred color- pattern afforded its possessors no great or striking protection against elimination by natural enemies.’’ But photographs serve as accurate measures of the conspicuousness of the fowls in the eyes of color-blind enemies only. Therefore, if rats and pre- daceous birds are not color-blind, and there is perfectly good reason for supposing that creatures lower in organization than either have color vision, it is not at all certain that the elimina- tion in the two cases does not correspond fairly well with the actual difference in conspicuousness of the two t ; Kellogg and Bell’s® interesting ‘‘Studies of Variation in In- sects’’ deals with 24 species. In 23 of them, including the lady- beetle, Hippodamia, to which Dr. Pearl refers, the authors show that there is much variation in individuals which have success- fully run the gauntlet of natural selection. But since they have no knowledge whatever of the variation in the original populations, of which they have studied survivors only, these their results show nothing regarding the extent, or even the oc- currence, of selective elimination. In the honey bee alone duplicate studies were made of the variation of certain structures in individuals which were about to hatch, and in others, apparently from the same hive, after exposure to the vicissitudes of an active life. Among 200 drones in the first group the veins of the fore wings in 11 were im- perfectly developed, and as a result normal flight became difficult or impossible. The variation in others of the first series seems essentially the same as that observed in the 300 members of the second, among which none of the defective individuals were found. But if these facts prove anything, it seems to be that selective elimination does occur when unfavorable variations affect the normal functioning of an organ. Suggestion is en- tirely lacking that the mechanical efficiency of the wing is im- paired by the other variations noted, and it can scarcely be con- 5 Kellogg, V. L., and Bell, R. G., 1904, ‘‘ Studies of Variation in Insects,’’ Proc. Washington Acad. of Sciences, Vol. 6, pp. 203-332. 256 THE AMERICAN NATURALIST [ Vou. LI sidered a pregnant fact, in the present connection at least, that indifferent variations provide no basis for selective elimination. In the same two series of bees the variation in the number of hooks upon the costal margin of the hind wings was determined without significant difference appearing in the two cases. In this instance, since the hooks appear to operate to the insects’ ad- vantage in binding the fore and hind wings of each side together, it seems plausible enough at first glance, that the more hooks there are the more efficiently their function will be discharged. Upon second thought, however, a difficulty suggests itself. The number of hooks varies from 19 to 29 in different individuals, but even the smaller number may, for all that is known to the contrary, perform perfectly the function ascribed to them. In that event the others are superfluous and the advantage they confer entirely fictitious. But waive the objection, and what follows? Simply a conclusion which in its relation to the present argument is already invalidated: Variations of the magnitude indicated provide no ‘‘handle’’ for natural selection. t is unnecessary to carry the examination of the evidence farther. The three papers which have been reviewed are no carefully selected for criticism, but are the last, and apparently the most important, of six certified to be ‘‘fairly representa- tive.’’ If, however, this characterization is correct, it is apparent that the case against selective elimination is greatly exaggerated. In conclusion, it appears that neither genetic research nor studies upon elimination closely limit the possibility that selection has played a very important part in evolution. In addition, recent field-studies* demonstrate novel facts of common occur- rence which must apparently be ascribed to the action of this factor. Hence as was suggested in the beginning, Dr. Pearl would seem to over emphasize the importance of variation, and to attach too little significance to selective agencies in determin- ing the course of racial history. W. H. LONGLEY GOUCHER COLLEGE, BALTIMORE e Longley, W. H., 1914, Report upon color of fishes of the Tortu Reefs, Carnegie Inst. of Washington, Year Book No. 13, pp. 207-208; 1915, t‘ Coloration of Tropical Reef Fishes,’’ Carnegie Inst. of Washington, Year Book No. 14, pp. 208-209; 1916; ‘‘The Significance of the Colors of Trop- ical Reef Fishes,’’ Carnegie Inst. of Washington, Year Book No. 15, pp. 209-212; 1916, ‘‘Observations upon Tropical Fishes and Inferences from their Adaptive Coloration,’’ Proc. Nat. Acad. Sciences, Vol. 2, pp. 733-737. THE AMERICAN NATURALIST Vor. LI. May, 1917 No. 605 STUDIES UPON THE BIOLOGICAL SIGNIFI- CANCE OF ANIMAL COLORATION Il. A Revisep Worxine Hyporuesis or MIMICRY Dr. W. H. LONGLEY GoucHER COLLEGE, AND DEPARTMENT OF MARINE BIOLOGY, CARNEGIE INSTITUTION OF WASHINGTON ALTHOUGH zoologists know that detailed resemblance in outward appearance may occur between different species of insects which are not closely related, they do not agree in their interpretation of the facts they observe. Present knowledge, indeed, justifies nothing more than tentative explanations of mimicry ; but, in this matter, ob- servations recently reported! limit one’s freedom of choice, since they appear to bear directly upon the validity of current hypotheses reviewed in the following pages. The first attempt to interpret mimetic resemblance as a result of natural selection was made by H. W. Bates,? who writes: What advantage the Heliconide possess to make them so flourish- ing a group, and consequently the objects of so much mimetic resem- blance, it is not easy to discover. . . . It is probable that they are un- palatable to insect enemies. ... They have all a peculiar smell. I never saw flocks of the slow flying Heliconide in the woods persecuted by birds or dragon flies, to which they would have been an easy prey; 1 Longley, ‘‘The Colors and Color Changes of West Indian Reef-fishes,’’ Jour, Exp, Zool, 1917. It happens that the present paper will appear slightly before that cited. 2 Trans. Linn. Soc. Lond., Vol. 23, p. 510. 2 258 THE AMERICAN NATURALIST [ Vou. LI nor, when at rest on leaves, did they appear to be molested by lizards or the predaceous flies of the family Asilidæ, which were very often seen pouncing on butterflies of other families. If they owe their flour- w talidæ, whose scanty number of individuals reveals a less protected con- dition, should be disguised in their dress, and thus share their immunity. Bates himself points out the fact that ‘‘some of the mutual resemblances of the Heliconidz seem not to be due to the adaptation of the one to the other, but rather, as they have a real affinity, . . . to the similar adaptation of all to the same local, probably inorganic conditions.’’ Thus the application of his hypothesis was limited from the beginning. Fritz Miuller* showed how those instances of resem- blance which his predecessor ascribed to the influence of Lamarckian factors might be aligned with the Darwin- ian hypothesis. In Meldola’s* translation of his original paper his idea is expressed as follows: What benefit can one species derive from resembling another, if each is protected by “cag PUEA Obviously none at all, if insectivorous birds, lizards, etc., have acquired by inheritance a knowledge of the species which are ite) or distasteful to them—if an unconscious in- telligence tells them what they can safely devour and what they must avoid. But if each single bird has to learn this distinetion by ex- perience, a certain number of distasteful butterflies must also fall vic- tims to the inexperience of the young enemies. Now if two distasteful species are sufficiently alike to be mistaken for one another, the ex- perience acquired at the expense of one will likewise benefit the other; both species together will only have to contribute the same number of victims which each of them would have to furnish if they were different. Recent years have been marked by a tendency upon the part of some observers’ to extend the bounds of the Miillerian associations, until in a given fauna a large pro- portion of the insects which show the same color com- bination are included in one bionomie group. Scores of species have, indeed, been assigned to some, and it has 8 Kosmos, May, 1879, p. 100. 4 Proc. Ent. Soc. Lond., 1879, p. xxvii. 5 Marshall aed Poulton, ‘‘The Bionomies of South African Insects,’’ Trans. Ent. Soc. Lond., Vol, 41, pp. 287-584. No. 605] ANIMAL COLORATION 259 been alleged that very many types from the same region show the influence of one or another of the dominant Miillerian aggregations. Miiller’s hypothesis, however, has not attained preeminence solely by extension to newly discovered cases of resemblance, but has prospered in many instances at the direct expense of the Batesian con- ception. Its changed fortune depends largely upon Dixey’s® discovery of mimetic attraction, or reciprocal mimicry. This relation, which it is held exists between many in- sects previously considered typical Batesian couples, sug- gests that the observed resemblance involves mutual ad- justment. But the idea that each species introduces into its own pattern elements characterizing that of the other, and thus contributes actively to the development of a common type of coloration, is intelligible in terms of the mimicry hypotheses, only if the superficial resemblance so attained is mutually advantageous. Whatever justi- fies change in the interpretation of fact lies here; for, if it be assumed in the beginning that the likeness noted must have arisen either through the action of Batesian or Miillerian factors, it must be admitted that the notion of mutual advantage seems more fitly associated with the latter: two species, each of which enjoys marked immu- nity, seem better able to force reciprocal concessions in their achievement of resemblance, than they should, if there were great disparity in their means of defence. Perhaps the most obvious suggestion from Dixey’s . research is not, after all, that a closer approximation to truth may possibly be attained through reclassification of mimetic resemblances, but that an intergrading series of alleged Batesian and Miillerian mimics is perfectly con- ceivable, in which no known test could possibly determine the occurrence of a natural break. Hence there would seem to be only historical reasons for maintaining the two categories. It is noteworthy that pthc hypothesis to which ref- erence has been made was in its original form an attempt 6 Trans. Ent. Soc. Lond., 1894, pp. 249-334. 260 THE AMERICAN NATURALIST [Vou. LI to explain the existence of conspicuous creatures. The sole concern of each was the interpretation of resem- blances, which were later commonly considered mere in- cidents in the attainment and use of warning colors. This shifting of emphasis from ‘‘likeness which is perfectly staggering’ to conspicuousness, which failed to elicit a single outspoken comment in the original papers of Bates and Müller, is distinctly chargeable to Wallace, who ex- tended his hypothesis of the functional conspicuousness of bright colors, until it included the facts of mimicry. Abbott H. Thayer® has proposed an explanation of the resemblance between unrelated species of butterflies, which is consistent with his thesis that the foremost func- tion of animal coloration is concealment. This hypothesis is Darwinian in principle, but in practise is directly op- posed to current selectionist opinion. It is purely specu- lative, and is stated as follows: It is surely conceivable that in a certain region, one particular form of flower-scenery representation may furnish such advantage to butter- flies as to cause many widely separated species to become modified till they wear a common aspect, and it is conceivable also that there would be one common form of wing that would best lend itself to this scheme. More recently Thayer’? adds that in the paper from which the excerpt above is taken he ascribed more im- portance to butterflies’ resemblance to flowers, as com- pared to their rendering gf scenery, than he should at the | later date; but this necessitates no modification of his ‘idea that mimicry is mere incidental resemblance between species, which, through selection based upon the obliter- ative effect of their coloration, conform ever more closely to one ideal representation of their common background. If ‘the Darwinian ranks are divided upon the question of the prevalence of conspicuous types of coloration, sim- ilar though less Moe dissension appears among their op- 7 Bates, l ¢., p. t Dasiialen, p- pe pee: and Co., 1891. 9 Trans, Ent. Soc. Lond., Vol. 42, p. 557. 10 See Thayer, Gerald H., ee Coloration in the Animal King- dom,’’ Appendix B, p. O51. Maemillan and Co., 1909. No. 605] ANIMAL COLORATION 261 ponents. To Piepers,'! for example, specific coloration is in large part a visible token of internal organization de- termined from time immemorial, a result of orthoge- netic evolution capable, however, of being accelerated or retarded by external conditions. To Packard?? it repre- sented a racial response of organism to environment, and mimicry seemed an effect of exposure to conditions of the same kind. But if evolutionary processes be largely be- yond the control of external agents, and species spring from species through internal reorganization, as one con- figuration follows another in the kaleidoscope, one should anticipate that many color combinations of exaggerated conspicuousness might result. If, upon the other hand, the development of color and pattern be determined by animals’ environment, their coloration may well repeat dominant notes from their surroundings. Hence it is not surprising to find that Packard expresses hearty appre- ciation of Thayer’s discoveries, and Piepers, despite his anti-Darwinian attitude, might have much in common with Wallace and Poulton. Thus the full series of con- tradictions is rounded out, and the unsettled state of opinion concerning mimicry, or indeed the whole matter of animal coloration, is apparent, for the two qualities, utility and conspicuousness, openly or tacitly affirmed or denied, may be ascribed in every possible combination to _ the color and pattern of a single organism. It may apparently be stated safely without qualification that the bright colors of tropical fishes as a class are cor- related with the animals’ habits, and, in the case of all but red, distinctly repeat tones characterizing their normal environment. But, other things being equal, no one will maintain that any system of external pigmentation could be less conspicuous than one conforming to this principle. Therefore, among these creatures at least, the occurrence of bright colors in contrastive patterns is not inconsistent with the idea that the forms that display them are as in- conspicuous as may be under the conditions in which they 11‘ Mimikry, Selektion und Darwinismus.’’ Leiden, 1903. 12 Proc. Amer, Philos. Soc., Vol. 43, p. 421. 262 THE AMERICAN NATURALIST [ Von. LI live. It is this fact that necessitates a review of hypoth- eses of animal coloration that postulate conspicuous- ness; for one cannot safely disregard the suggestion that principles applicable to one group of animals may be valid also in the case of others. It is of interest to note that Chapman!* studied the birds of Trinidad under favorable conditions and ob- served that distinct types of coloration marked those of different habits. The most brilliant species occupy the most exposed positions in the treetops. More sedentary forms inhabiting the body of the trees are largely green, and brown predominates in the coloring of those that climb upon the tree-trunks, frequent the undergrowth near the forest border, or live upon the forest bottom. These ecological records are a mere incident, a by- product of their author’s activity. Through lack of de- tail they possess no great intrinsic value, but are highly significant in their present setting. Mention should also be made of Potts’!* observation that shrimps living sym- biotically with crinoids upon the Australian reefs repeat = the colors of the forms with which they are associated. That comparable facts regarding other groups of animals are not available is immaterial. If the bright colors of tropical birds, fishes and some crustacea repeat those of the animals’ respective environments and minister to the inconspicuousness of their possessors, it is of interest to inquire what data and reasoning support the contention that in insects similar combinations bear a different rela- tion to the colors about them and discharge another func- tion. It appears first, from the statements of a number of their more prominent advocates, that there are funda- mental theoretical objections to the hypotheses of warn- ing coloration and mimicry. Poulton!® remarks that the acquisition of an unpleas- ant taste or smell, together with a conspicuous appear- 13 Bull. Amer. Mus, Nat. Hist., Vol. 6, pp. 19-20, 14 Carn. Inst. Wash., Papers Dept. Mar. Biol., Vol. 9, pp. 71-96. 15 Proe. Zool. Soc. Lond., 1887, p. 192. No. 605] ANIMAL COLORATION 263 ance, is so simple a form of protection, and yet ex hy- pothesi so absolutely complete, that it seems remarkable that more species have not availed themselves of this mode of defence. He argues that if once their potential vertebrate enemies were driven to eat any such insects in spite of their unpleasant taste, they would almost cer- tainly soon acquire a relish for what was previously dis- agreeable, and the insects would be in great danger of ex- termination, having in the meantime become conspicuous by gaining warning colors. He concludes that if this reasoning is correct, it is clear that this mode of defence is not necessarily perfect, and that it depends for its ap- parently complete success upon the existence of relatively abundant palatable forms: in other words, its employment ` must be strictly limited. Dixey! encounters the same difficulty in his consid- eration of Batesian mimicry. He observes that in this relation the advantage is all on the side of the edible mimicking species, whose existence is, indeed, a source of danger to the form mimicked, inasmuch as any ex- perience gained by tasting the former would be used to the detriment of the latter. From these considerations he believes that such an association can exist only when the numbers of the one species are insignificant in com- parison with those of the other. Upon this point he is in essential agreement with Wallace,” who states that mim- icry has beén shown to be useful only to those species and groups that are rare and probably dying out, and would cease to have any effect should the proportionate abun- dance of mimic and model be reversed. Here, then, are two hypothetical types of association whose persistence admittedly depends upon the main- tenance of definite though undetermined ratios between their components. Hence it devolves upon those who hold that the assumed relations are real, to outline some sys- tem by which the due proportion of protected and unpro- 16 Trans. Ent. Soc. Lond., 1897, pp. 317-332. 17 Westminster Review, Vol. 32, N. S., p. 35. 264 THE AMERICAN NATURALIST [Vou. LI tected forms might be maintained through the sab a aba of natural factors. In an attempt to avoid this preliminary difficulty Poul- ton declares that it has always been recognized that an insect may be distasteful to one vertebrate enemy, but palatable to another. He suggests a different counter- balancing limit, which he admits would certainly in time become identical with the other. He argues that a verte- brate enemy may be forced by stress of hunger to eat an unpalatable insect, by implication asserts that this adapt- ability of potential enemies of forms tending to assume warning colors would limit the number of species de- veloping them, and considers the truth of his suggestion confirmed by experimental tests. It seems, however, that as long as warning coloration confers any advantage, other species should move in the same direction, for the same reasons that those did in which it first appeared. The number of warningly col- ored species should therefore increase until the situation conceived by Poulton materialized. But in that event the experimental proof that hungry animals are not so fas- tidious as those that are well fed is insufficient to meet the exigencies of the situation; for one has no reason to believe that animals which might become adjusted to the new condition would confine themselves to a diet of in- sects whose warning coloration was recently acquired, and would leave intact the vested interests of those that first attained it. Indeed, if one is permitted to speculate, it seems not wholly unreasonable to anticipate a swing of the pendulum in the other direction, so that conspicuous forms might face the possibility of almost complete ex- tinction. This is suggested by the well-known fact that interspecific adjustments are not rigid, and that a state of approximate equilibrium is frequently modified* by climatic or other factors, and restored only after a series of more or less definite oscillations. Examples in point are furnished by Howard’s'® observations upon ichneu- 18 ‘f New Nature Library,’’ Vol. 7, Pt. 1 (The Insect Book), p. 68. New York, Doubleday, Page and Co. No. 605] ANIMAL COLORATION 265 mon flies and their primary and secondary parasites, and more particularly by those of Forbes!’ and Bryant”? upon the feeding habits of birds. In his paper, which has been cited above, Dixey was not discussing the validity of the mimicry hypotheses and has not attempted to explain what limits the numbers of a mimic in proportion to those of its model. Wallace also failed to elucidate the mystery, leaving his readers to find the way out of their difficulty as they might best be able. Search elsewhere for helpful suggestion in the matter yields little of value. Poulton?! holds that ichneumon flies are particular enemies of the larve of ‘‘protected’’ Lepi- doptera, but this idea need not be taken very seriously at present, for it is apparently based upon the observation of a high proportion of parasitized individuals among such forms rather than upon comparative statistics cover- ing ‘‘unprotected’’ species as well. Even if it were true, while it would bolster the warning color hypothesis, it would aggravate the situation regarding Batesian mim- icry. Therefore only one conclusion is possible: The theoretical objections they themselves have raised are not adequately met by spokesmen for the two hypotheses. What observed facts support the contention that many animals are made conspicuous and profit by the striking combinations of color they display, should next be deter- mined. : In 1867 Wallace”? suggested before the Entomological Society of London that, as a rule, brilliantly colored larve are distasteful to birds, expressed his desire for infor- mation and his gratification, if any who kept birds, par- ticularly indigenous species, would make experiments with different larve to ascertain which were eaten and which rejected. Members of the society and others have repeatedly acted upon his suggestion and commonly 19‘ The Regulative Action of Birds upon Insect Oseillation,’’ Jil. State Lab. Nat, Hist., Bull. No. 6. 20¢«The Condor,”? Vol. 13, on p. 195-208. 21‘‘ The Colors of Animals,’’ p. 182. New York, D. Appleton and Co., 1892. For additional data, see Trans. Ent. Soc. Lond., Vol. 41, p. 337. 22 Proc, Ent. Soc. Lond., 1867, pp. oo 266 THE AMERICAN NATURALIST [ Von. LI - agree that the results of the experiments support his hypothesis, whose application has meanwhile been greatly extended. But the spirit in which the observed facts are interpreted precludes all possibility of the main inference drawn from them being seriously considered by any un- prejudiced critic who examines the argument. a dull-eolored insect be eaten, this is held by impli- cation to lend strong support to the hypothesis of warn- ing coloration; for palatable insects must be inconspic- uous or be destroyed. If, upon the contrary, such an insect be rejected, its distastefulness may be a useless character, an accident of metabolism, or vestigial, or may be related to functional distastefulness in preceding stages of the life history; in any case ‘‘it must be remembered that an unpleasant attribute must always appear in advance of the warning coloring,’’?® so the result is not inconsistent with Wallace’s suggestion. Again, if a bright-colored insect be rejected, this accords with his contention; and finally, if such a one be accepted, experimenters are prone to agree with J. Jenner Weir? who writes: But I am by no means inclined to attach undue importance to this fact, because the birds, being in a state of confinement, might readily be expected to eat insects, which in a state of nature, with a less lim- - ited choice, they would reject. It is unnecessary to review these experiments in detail and to attempt to evaluate them, for this has already been done thoroughly by McAtee,?° who shows their utter futil- ity by comparing typical observations upon caged birds with the facts revealed by analysis of the stomach con- tents of wild specimens of the species experimented upon. Regarding information from such sources as that last- mentioned he writes: Since this evidence is sufficient in itself and since experimental data must be supported by it, why perform the experiments? The same time spent in collecting trustworthy data regarding the natural food 23 Poulton, ‘‘The Colors of Animals,’’ p. 176 24 Trans. Ent. Soc. Lond., Vol. 7, Ser. 3, p. 22. 25 Proc. Acad. Nat. Sci. Phila., 1912, pp. 281-364. No. 605] ANIMAL COLORATION 267 habits of animals would bring much greater returns, and the result would be truth, not imaginative inferences from abnormal behavior. Before concluding the discussion of this matter it should be stated that even if it were proved that bright- colored insects are distasteful, it might not be inferred fairly that they are conspicuous, or that their coloration has a specific warning (aposematic) function. Feeding experiments under ideal conditions might determine the presence or absence of distastefulness, and show to what extent it is correlated with the display of color combina- tions of a particular sort. But even a high degree of ` correlation between unpalatability and gaudy coloration proves that the latter is conspicuous, no more than the demonstration that an unknown substance has the ap- proximate hardness of gold proves that it has the same specific gravity; for brightness, or vividness of colora- tion, and conspicuousness are incommensurable. Except in Thayer’s contributions, confusion upon this point has prevailed from the beginning, when Wallace used interchangeably the expressions, ‘‘brilliantly colored larve’’ and ‘‘caterpillars conspicuous by their lively coloration.’’ But the assumption that some animals are conspicuous, or, in other words, that while their habits remain the same their average visibility might be greatly diminished by another system of coloration than that they possess, can neither be adequately defended nor re- futed, until the results of such exhaustive studies of ani- mals’ habits as have rarely been attempted are available. The distribution of animals must be studied intensively, for the division of the world into provinces and the sub- division of these into their major components by gross dissection is not a technique of sufficient refinement to discover the essential relations between organism and en- ‘vironment, Passing to another phase of the matter, the value of recorded observations upon the conspicuousness of in- sects may be shown very clearly. Bates saw no Heliconide attacked by dragonflies or 268 ` THE AMERICAN NATURALIST [ Vou. LI other predaceous insects which often pounced upon but- terflies of other families. But Poulton? finds, to state it mildly, that ‘‘there is good reason for believing that such attacks are not rarely made, and that predaceous insects are important enemies of aposematic butterflies.’? He also writes?’ of the Batesian hypothesis: This was not, as has been generally supposed, originated by Bates during his years of observation in the Valley of the Amazon. It arose in his mind after his return home, when he came to examine his collec- tion and to reflect upon his experiences, Under these circumstances the uncorroborated testimony of this witness concerning matters which are not known to have been carefully investigated in the Brazilian wil- derness and are not determinable from the study of preserved material is of little immediate consequence. Among subjects regarding which his opinion can at pres- ent he held only in slight esteem, and concerning which he expressed himself in other communications than his original paper upon mimicry, the inherent conspicu- ousness of bright colors may be justly included. e positive assertions in the following quotation describing conditions observed by an entomologist in British Guiana are also instructive; the same is true of their author’s naive conclusion. W. J. Kaye? writes: The forest is dark and gloomy, and throughout the greater part of the year excessively damp owing to a superabundant rainfall. The character of the vegetation is always the same, as even in the dry season the trees are never otherwise than a fresh green. It is not sur- prising, therefore, that practically the whole of the Lepidoptera, except- ing, of course, the several species of Morpho, present a very uniform, somber tone of coloration. Even the very fine and brightly colored Heliconius catharine, H. astydamia and H. egeria do not strike one in their surroundings as being particularly gaudy, and one is bound largely to admit the assertion of Abbott H. Thayer that many species we call conspicuous are not really so in their natural surroundings. It must, however, have been quite impossible for nature to have evolved such 26 Trans, Ent, Soc. Lond., Vol. 41, p. 328. See also Vol. 44, pp. 323-409. 27 ‘í Essays on Evolution,’’ p. 211. Oxford, Clarendon press, 1908. 28 Trans. Ent. Soc. Lond., Vol. 44, p. 412. No. 605] ANIMAL COLORATION 269 minutely close resemblance in unrelated groups without the aid of Miillerian mimicry. It remains to state that among Lepidoptera different species have their characteristic attitudes of rest, fre- quent different places, fly at different levels, and are active at different times in the day. Even the two sexes of some species, and these are commonly dimorphic forms, do not haunt the same stations. It is certainly a pregnant fact, which we may accept since Wallace?® gives independent testimony to the same effect, that Bates*® observed many, apparently scores, of species, in which, as he says, the sun- loving males flaunted their gaudy hues in open places, while their respective females, soberly clad, frequented the forest shades. These facts and others that might be cited are indi- cations of diversity of habit among insects comparable with that which among fishes is correlated with the dis- play of different types of inconspicuous coloration. They suggest that in this group as well, external pigments are distributed among species according to an intelligible system other than that whose existence is commonly in- ferred. But if this should eventually prove to be true, we must have an explanation of mimicry without appeal to the concept of warning colors. Such a hypothesis has in fact been formulated by Punnett,*! for in spite of his apparent belief in the con- spicuousness of many species of butterflies it happens that he lays no stress upon it in his consideration of the origin of mimetic resemblance. His hypothesis, as he fully recognizes, is at present little more than naked sug- gestion. It is ingenious, is stated attractively in the cur- rent idiom of genetics, and is effectively displayed against a background of destructive criticism of its predecessors, from which it differs in minimizing the influence of nat- 29 Trans. Ent. Soc. Lond., Vol. 2, Ser. 2, pp. 253-264, 30 ‘The Naturalist on the River Amazon,’’ p. 291. Reprinted, New York, D. Appleton and Co., 81 ‘í Mimiery in ce ae fe Cumbildgs; University Press, 1915, 270 THE AMERICAN NATURALIST [Vou. LI ural selection. It proposes an entirely new explanation of mimicry in the following terms: If we assume that sudden and readily appreciable variations of the nature of “sports” turn up from time to time, and if these variations happen to resemble a form protected by distastefulness so closely that the two ean be confused by an enemy which has learned to avoid the latter, then there would appear to be good grounds for the mimicking sport becoming established as the type form of the species. ... On this view natural selection in the form of the discriminating enemy will have played its part, but now with a difference. Instead of building up a mimetic likeness bit by bit it will merely have conserved and ren- dered numerically preponderant a likeness which had turned up quite independently. . . . Why variations on the part of one species should bear a strong Dihin to other, and often distantly related, species is another question. . . . The occurrence of mimetic resemblances is the n hereditary factors of which the total number is by no means very great. As many of the factors are common to various groups of butterflies, it s to be expected that certain of the color patterns exhibited by one group should be paralleled by certain of those found in another. Upon examination these statements appear to embody a formal explanation of the facts to which it is difficult to take exception. Hence it must be admitted that this is perhaps the goal toward which with regard to this prob- lem naturalists have been working for more than half a century. But while his hypothesis may be correct, its author’s reasons for deeming it so seem quite insufficient. The chief points of support on which it rests are the following, which are not arranged in the order of im- portance assigned to them: 1. The difficulty of finding the appropriate enemy which shall exercise the discrimination postulated by cur- rent hypotheses. 2. The alleged fact, mathematically demonstrated, that reciprocal mimicry between two species can not be estab- lished by selection of a long series of slight variations. 3. The theoretical difficulty of the initial variation in cases other than that above, for it seems reasonable that if the ancestral types from which mimic and model are derived were in the beginning very unlike in appearance No. 605] ANIMAL COLORATION 271 no slight departure from one in the direction of the other could have selective value. 4. The non-appearance of intermediates when a form which is assumed to have been derived from another by selection of a series of heritable variations (mutations) is crossed back with the original. 5. The apparent fact that the three females of the poly- morphic oriental butterfly, Papilio polytes, occur in pro- portions which are approximately the same now as fifty, or possibly one hundred and fifty years ago, although it may be demonstrated mathematically that if either of the two mimetic forms possesses even a slight advantage over that from which they are assumed to be descended, this should have appeared in an altered ratio while they continue to live and breed together under the same con- ditions. 6. The fact that certain variations induced by differ- ences in temperature or humidity are not directly in- herited, since it is alleged that this limits the material upon which selection might be supposed to operate. Of these points the first three may be more conven- iently considered later with other common objections to current hypotheses of mimicry. The remaining three will be next examined in order. Punnett accepts Fryer’s®? conclusion that in P. polytes the two supernumerary females which resemble P. aris- tolochie and P. hector differ genetically from the third female (which resembles the male) to the extent of one and two Mendelian factors, respectively. He cites the fact that there are forms of sweet peas, for example, which are known to have arisen as sudden sports, and behave in heredity as though they differed. from the nor- mal by a single factor. Hence he infers from analogy, first, that the two mimetic butterflies sprang from the primitive type by one or two mutations, as the case may , and, as a corollary, that resemblance to their models was not attained by gradual shaping of their destinies 82 Phil. Trans. Roy. Soc. Lond., Vol. 204 B. 272 THE AMERICAN NATURALIST [ Von. LI through the accumulation of lesser variations by natural selection. But Castle** still contends vigorously that a single ge- netic character may undergo quantitative change under selection, and if this be true the difference finally attained by two forms differing only in one factor might follow from the accumulation of an indefinite number of slight variations. Still the complete abandonment of Castle’s position would not save Punnett’s argument. For if re- ported observations and inferences concerning Droso- phila** are correct, a red-eyed strain has given rise to white by mutation and this in turn to eosin, which being crossed with the original red gave in the F, generation offspring in the proportion of three red to one eosin. It is quite immaterial whether one explains these rela- tions upon the hypothesis of multiple allelomorphs, or, as Punnett?ë prefers, upon the assumption of complete coupling of factors. In the former case one must admit that by breeding experiments the end product of a series of mutations can not be indubitably distinguished from one that results from a single modification of the affected factor. In the latter, one thust grant that the occurrence of the grandparental types in the offspring of hybrid par- ents in the ratio of three to one is no proof that the grand- parents themselves differed in respect to one character alone, or that the difference between the two resulted from changes occurring at one time in the germinal constitu- tion of an ancestor of one of them. Hence, in so far as this argument is concerned, there is in the case of P. polytes, for example, no assured reason for supposing that the aristolochia-like form did not attain its present appearance by a series of steps, of which a number of the later at least were preserved by virtue of the advantage they conferred upon the individuals in which they ap- peared. 33‘*Genetics and Eugenies,’’ p. 188. Cambridge, Harvard University Press, 1916. 34 Morgan, Sturtevant, Müller and Bridges, ‘‘The Mechanism of Men- delian Heredity,’’ p. 164. New York, Henry Holt and Co., 1915, 35 Jour. of Geneties, Vol. 5, pp. 37-50. No. 605] ANIMAL COLORATION 273 Little need be said regarding the inference that the constancy of the ratio in which the females of P. polytes seem to have occurred for many years shows that natural selection does not exist for this species in Ceylon, or else that its force is so slight that in half a century, and perhaps in a century and a half, it has produced no effect appreciable to the method of examination employed. This is valid only if it is true as postulated that the various types of P. polytes constitute ‘‘a population liv- ing and breeding together under the same conditions.”’ But it is gravely to be doubted that this indispensable condition is fulfilled. We have some evidence: (that of Bates and Wallace already cited) that butterflies which differ in color differ in habit, and if it should appear that the colors of butterflies in general are correlated with and repeat those of their surroundings, Punnett’s fifth point is forever invalid. For it will be impossible to establish by observation the universal negative that is required, which is, of course, that the three types of female do not differ in any constant respect in their normal behavior. Regarding the sixth point, which has reference chiefly to the fact of seasonal dimorphism among butterflies, it must first be affirmed that although the induced changes differentiating the broods of the spring and summer, or wet and dry seasons, are not directly inherited, the capa- bility of responding definitely to the physical stimulus of changed temperaturé or humidity is a heritable racial trait. The seasonal variations in the coloration of but- terflies may be analogous upon the whole with the instan- taneous color changes of tropical fishes, which also occur in response to external stimuli. The latter, however, fol- low more quickly than the former upon appropriate stim- ulation; they are reversible; and are known to be nor- mally adaptive, since they reduce the conspicuousness of the individuals in which they appear. It may eventually prove to be a fact that instantaneous adaptive color ad- justments, the phenomena of seasonal and sexual dimor- 28 Professor Gerould first called attention to this fact in THE AMER. NAT., Vol, 50 (1916), pp. 310-316. 274 THE AMERICAN NATURALIST (Von. LI phism, and polymorphism all have the same biological significance, i. e., that they represent different ways in which the coloration of a species exercises its obliterative function in a greater variety of circumstances than would be possible if it were uniform. Upon this view of the matter there would seem to be no reason why color vari- ations of the seasonal sort should not provide material for evolution by natural selection. Before suggesting another possible explanation of the fact of mimetie resemblance it seems desirable to state more specifically why certain of those already mentioned seem improbable. Some color patterns are apparently limited to fishes whose habits are similar. Others occur which have sur- vived the introduction of marked structural changes and are now the common property of whole families or groups of families, whose manner of living varies decidedly from species to species.. There is one such system of coloring among grunts, groupers and snappers (Hemulide, Epinepheline, and Lutianide), and Labrids and Scarids share another. In each pattern modifications may be _ noted which seem particularly appropriate under the local conditions in which they appear. Individual elements may lose all semblance of the original, and yet the nature of the whole be not obscured. But these facts make one skeptical regarding Thayer’s hypothesis, for if, in but- terflies too, detail is less important than the appropriate effect of the whole, the probability is remote that for dif- ferent environments complex, ideal, protective or conceal- ing patterns exist, whose slightest spot is so significant that there is marked tendency for forms of different racial endowment to attain them, if their habits are similar. The same facts militate against such a conception as Packard’s, that mimicry is a result of similar reaction to the direct influence of one set of external conditions. For coloration is characterized by such conservatism or in- ertia, and the same elements of pattern appear in such a variety of habitats, that the power of environmental in- No. 605] ANIMAL COLORATION 275 fluence to induce uniformity of coloration seems dis- credited. If the possibility of the direct influence of local climatic factors be excluded, Piepers’ hypothesis, that mimicry is ` largely due to species having independently attained the same stage of development orthogenetically, leaves the facts of geographical distribution of mimics and their models enshrouded in mystery. That this is a very real difficulty follows from Punnett’s statement,*7 Examples of close resemblance between butterflies which live in dif- ferent parts of the world are relatively rare and serve to emphasize the fact that the great bulk of these resemblance cases are associated in pairs or little groups. Finally, instances of mimicry are, after all, only selected examples of resemblance, and it is desirable, if possible, to formulate an explanation that will apply to all equally well. But whether the likeness between them rises by sporting or otherwise, it is not to be supposed that Phas- mids and green leaves or dry twigs possess Mendelian factors in common. Therefore it seems profitable to pro- ceed for a little upon the assumption that mimicry is in some cases at least a visible token of the fact that the species manifesting it are linked by some bond other than common descent, common habit, or exposure to the influ- ence of a common environment. What this may be, does not appear, unless through mutual resemblance advan- tage accrues to some or all of the forms concerned. Evidence compiled by Marshall®* shows that birds un- doubtedly attack butterflies, but others deny that they feed upon the insects freely enough to affect the evolu- tion of their coloration, and more particularly the mimetic resemblances between different species. Hence there is plainly a question at issue concerning the sufficiency of an assigned cause to produce a stated effect. Under the circumstances any evidence tending to show that the fre- ~ quency of birds’ attacks has been underestimated, or that 87 L. €., p. 54. ; 3S Trans. Ent. Soc. Lond., 1909, pp. 329-383. 276 THE AMERICAN NATURALIST [ Von. LI their influence may be supplemented by that of other enemies is of the greatest interest. In this connection Swynnerton’s*® observation that of twenty small. bird excreta collected in the African forest no less than eighteen contained scales and small wing frag- ` ments of Lepidoptera has suggestive value. But, for the moment at least, it is more important that it appears that mimicry might be initiated and advanced by indiscrim- inate feeders, including lizards and insectivorous insects, provided only that they possess color vision. For to whatever extent such influence prevails it obviates the necessity of appeal to the effects of discriminate feeding by birds or other animals, and makes it possible to fore- stall the criticism to which reference has been made above. Therefore it is suggested as a tentative explanation of mimicry, that it has commonly arisen as a result of bio- nomic pressure applied first by discriminate or indis- criminate feeders, which by elimination of unadapted variants have forced their accustomed prey to assume color combinations which most effectually conceal it in its normal environment. In addition, for no demoystrated reason, in a few of the many thousands of cases in which colors adapted to the environment and habits of their possessor have been evolved, patterns have appeared which have been sufficiently like one another to deceive enemies which exercise discrimination in their choice of food. Beyond this point the evolution of resemblance may have proceeded according to accepted formule, but without conspicuousness being involved at any point in the process. : It is submitted that in our present state of ignorance this construction may be placed upon observed facts rationally and without exposure to the criticism that has been directed against other attempted interpretations. ` However, the chief classes of facts to be explained and the most serious objections registered against the Neo- Darwinian hypotheses of mimicry will be presented, that 39 Ibis, 1912. No. 605] ANIMAL COLORATION Zit the reader may judge whether a passage between Scylla and Charybdis may be made in safety. Professor Poulton’s extensive studies have convinced him that the evolution of mimetic resemblance has been directed by natural selection, yet the evidence upon which his conclusion rests may be taken over bodily and supports the’ revised hypothesis as consistently as that to whose service it was originally dedicated. There is nothing anomalous in finding mimic and model living under the same conditions, certain groups of insects show- ing the same series of local color varieties, or such diver- sity of coloration appearing in one group of butterflies or moths as allies them outwardly with different ‘‘pro- tected’’ genera. The same is true of the fact that insects with every variety of larval experience as adults possess the same type of coloration, that mimetic females are more common than males, or that the common coloration possessed by mimic and model is attained in the most diverse fashion, that is, that cases of mimicry are typical instances of analogy. Throughout the whole series of observations the points of agreement and difference are consistent, as far as is known, with the fundamental as- sumption that color and habit are associated variables. Passing to the negative side of the argument, we may first consider the statement that it is impossible that re- eiprocal mimicry should have been brought about by natural selection of small variations. Punnett has this idea from Marshall“ and uses it to emphasize the diffi- culty of the initial variation even in cases where it might seem that the theoretical advantage to be gained from mutual resemblance by two species would simplify the attainment of likeness. But Dixey,*? against whose posi- tion the argument was originally directed, has exposed its unsoundness by calling attention to a number of critical 40 See ‘‘ Natural Selection the Cause of Mimetic Resemblance and Common Warning Colors’’ in ‘‘ Essays on Evolution,’’ 1908. 41 Trans. Ent. Soc. Lond., Vol. 45, pp. 93-142. 42 Trans. Ent. Soc. Lond., Vol. 45, pp. 559-583. 278 THE AMERICAN NATURALIST [ Von. LI points which his opponent had failed to take into con- sideration. It may be added that Marshall’s reasoning rests upon what is without much doubt a baseless assumption, for he follows Müller in postulating that two species of distaste- ful insects will lose the same absolute number of indi- viduals through attacks of ignorant enemies which in the beginning recognize neither of them. Asa matter of fact, if two species differing in no respect except appear- ance are represented in the same area by 100,000 and 5,000 individuals, respectively, as Marshall assumes, the chances are 20:1 that any animal making an independent test of the food resources of its environment would first meet the more abundant form. Unless it learns its lesson perfectly from a single experience, the chances are essen- tially the same that it will kill another butterfly of the same kind before it encounters one of the second distasteful sort. But if most inexperienced enemies learn at the expense of one species that some butterflies are not edible, it is searcely to be supposed that they will undergo as many unpleasant experiences before they retain an im- pression of the disagreeable character of the other. Hence Marshall’s criticism can not be considered at pres- ent a valid objection. Packard** believed that the concept of Miillerian mimi- cry had been overextended. He thought that in ac- cumulating so many examples of warning coloration in their ‘‘Bionomics of South African Insects’? Marshall and Poulton* in particular attempted to prove too much. Why an association of some scores of species represent- ing many orders of Mashonaland insects should be pivoted upon the bitter-flavored beetle, Lycus, though some mem- bers of the group seemed more amply protected from at- tack by birds and lizards, was not clear. Yet one dares not be dogmatic in such matters, for the wasp, Pompilus, though more adequately equipped for defense than any other member of the association, may have drifted toward 43 Proc. Amer. Philos, Soc., hig es p. 424, 44 Trans, Ent. Soe. Lond., Vol. ` No. 605] ANIMAL COLORATION 279 it at a comparatively late date, when the relatively slight distastefulness of a large number of insects of one type of coloration subtended a larger angle in the conscious- ness of insectivorous animals than the greater unpal- atability of any single form. However, the idea that what has been considered mimicry is too common, and that in general the most effectively protected types should be the nuclei of the Miillerian combinations, is certainly not wholly unreasonable. . One of the chief reasons for believing in the existence of warning colors, and particularly of common warning colors, is the fact that some families of insects have slight range of color and pattern compared with others. Mayer*® found that ‘‘the 200 species of Papilio in South America display 36 distinct colors, while the 450 species of Danaoid Heliconide exhibit only 15,’’ and that ‘‘there is -= no lack of individual variability among the species of the latter, yet as a whole they vary but little from the two great types of color-pattern represented by Melinea and Ithomia.’’ To explain these facts he felt obliged to resort to Miiller’s hypothesis, but if instead of thinking of Itho- miine and Papilionide one considers Holocentride and Labride, an alternative solution appears. The squirrel fishes seem to be of red or reddish coloration the world over, but their habits are equally invariable, while the Labrids’ diversity of coloring is no greater than that pre- vailing in the varied environments in which they live. Such facts indicate the necessity of making detailed studies of the coloration of tropical Lepidoptera and cor- relating the facts discovered with the insects’ distribution and behavior. When this is done there is reason to sup- pose that combinations of the same colors will be found upon animals of the same habit, which would have been as they are in many species, if any or all the others which display the same combinations had never existed. That is to say, it is probable that much that has masqueraded as Miillerian mimicry is nothing but the result of con- 45 Bull, Mus. Comp. Zool. Harv. Coll, Feb., 1897, p. 225. 280 THE AMERICAN NATURALIST [ Vou. LI vergent evolution, which has been difficult to explain be- cause of the deep-seated misconception that has prevailed regarding the function of animal coloration. Dewar and Finn** cite a number of instances of resem- blance between mammals, and others between birds, whose ranges coincide at no point. For the most part these likenesses do not seem comparable with the clearest cases of mimicry among insects in the degree of detailed re- semblance they involve, and scarcely seem to rise above the level of interesting coincidences. It is unquestionably true, nevertheless, that such degree of likeness as may spring up between two species whose bionomic association is impossible on account of differences in geographical distribution, may also arise between species of one region without reference to the action of natural selection di- rected toward the production of resemblance. Lock* states that Syrphid flies, which closely mimic small bees and wasps whose habits are similar to their own, are surprisingly numerous in southern Japan, and that their resemblance to bees is particularly noticeable, though these are conspicuous by their absence. Hence the question arises, how the flies can benefit by their re- semblance to them: to which one must apparently answer, that under the conditions stated, the bee-like disguise can, as such, be of little value. But this query is overshadowed in interest by another: If the Syrphids are unprotected and driven by their enemies to assume the appearance of defended forms, how do they survive in regions where their disguise possesses no suggestion of unpalatability. The idea is not to be entertained for a moment that Lock would be at a loss for an answer. But if the concept of warning coloration be abandoned, there is no reason to suppose bees less perfectly adapted in color and form than other animals to their respective modes of life. Bee- like flies whose habits resemble those of bees should there- 46‘*The Making of Species,” pp. 242-245. London and New York, J. Lane, 1909. 47 ‘í Recent Progress in the Study of Variation, Heredity. and Evolution,’’ p. 58. London, John Murray, 1907. No. 605] ANIMAL COLORATION 281 fore be well able to exist beyond the range of models, which they may have mimicked in other times and places, if their particular type of coloration is as well suited to the new environment as to the old. An apparent inconsistency in the Batesian and Miiller- ian hypotheses as at present interpreted has been fre- quently noted by hostile critics, To Reighard‘’ it ap- pears, for example, that if insectivorous vertebrates have pushed the resemblance between mimics and their models to the point of apparent identity, ordinary specifie dif- ferences should suffice to warn them of the unpalatability of prospective and familiar prey. This objection. is so fairly met by the revised hypoth- esis, and the ground for criticism so completely removed, that further comment is unnecessary. But even when no inconsistency is involved in the explanation of the facts, some will doubtless consider the resemblance of the mimic to its model, or of insects to other objects, hypertelic. It is doubtful, however, whether hypertely embodies a real difficulty. For just as two streams flowing down a tol- erably smooth inclined plane of infinite length will even- tually unite, if all deviations of one or both which exceed a given magnitude are blocked when they tend to increase the distance between them, so, if heritable variations in the color and pattern of a given mimic are distributed ac- cording to Quetelet’s law, for example, and only the ex- treme forms most unlike the model be eliminated in suc- cessive generations, closer and closer resemblance between the two may appear and approach identity with- out appeal to that over-refinement of vision whose ex- istence among insects’ enemies is at least problematical. It is a standing objection to the mimicry hypotheses, and indeed to the explanation of any highly complex adaptation by natural selection, that at every stage the degree of resemblance attained must have been service- able in order to assure its survival. It is understood, however, that this objection is applicable only to stages 48 Carn. Inst. Wash., Papers from Tortugas Lab., Vol. 2, p. 315. # 282 THE AMERICAN NATURALIST [ Vou. LI following the first to which the selectionist ascribes de- ceptive value. Resemblance resulting from undirected variation, or existing for other reasons, is not subject to this criticism. Darwin recognized this fact and attempted to throw upon another cause than natural selection a large part of the burden of producing functional resemblance. His idea may best be expressed in his own words: The process of imitation probably never commenced between forms widely dissimilar in color. But, starting with species already some- what like each other, the closest resemblance, if beneficial, could readil or coloring wholly unlike that of the other members of the family to which it belonged. There is, however, some difficulty on this head, for it is necessary to suppose in some cases that ancient members belonging to several distinct groups, before they had diverged to the present ex- tent, accidentally resembled a member of another and protected group in sufficient degree to afford some slight protection, this having given the basis for the subsequent acquisition of the most perfect resem- blanee.*° Weismann”? attempted to avoid the same difficulty in another way. He makes no assumption that the original difference betwen mimic and model was distinctly less than that appearing at present between typical members of their respective families, but magnifies the importance of the first slight resemblance and subsequent positive variations. He had been deceived repeatedly, at least for the moment, by similarity in the flight of different species _ whose colors were not the same, and held as a consequence that mere variation in the manner of flight combined with the habit of associating with the form mimicked might have prepared the way for selection. Wallace®! would have it that certain butterflies, having 49‘ Origin of Species,’’ Chap. XIV. 50‘¢The Evolution Theory,’’ Vol. 1, p. 93. London, Edward Arnold, 904, 51 ‘* Darwinism,’’ p. 243. No. 605] ANIMAL COLORATION 283 become unpalatable through the possession of disagree- able juices, developed distinguishing marks, whether in color, form or mode of flight. He then plunges in medias res with the assertion that ‘‘during the early stages of this process, some of the Pieridæ, inhabiting the same dis- trict, happened to be sufficiently like some of the Heli- conide to be occasionally mistaken for them.’’ There- after, as may be anticipated, evolution proceeded merrily, and examples of Batesian mimicry were perfected in due time. Wallace’s pronouncement begs the whole question. Weismann’s hypothesis is conceivably true, but lacks the support necessary to carry conviction. Darwin’s idea, finally, seems to be at variance with fact, since Poulton®? infers from his own studies that the conclusion that emerges most clearly is the entire independence of zoological affinity exhibited by mimetic resemblance. unnett also shows most clearly how impossible the Darwinian suggestion is, but errs when he supposes that it can not be true in many cases that model and mimic were closely alike to start with. His demonstration may be accepted that the development of mimetic resemblance has not been commonly facilitated by preexisting like- ness due to racial affinity, but he has wholly disregarded the fact that the degree of likeness which it is necessary to presuppose, if mimicry has been brought about by a series of comparatively small variations, might occur for other reasons. May we not assume, for example, that the Pieride and Heliconide are usually distinctly different in their habits, and that the coloration of typical members of each 52 Proc. Linn. Soc. Lond., Vol. 26, p. 570. 53 This ean scarcely be considered a rash assumption, since Wallace states (Trans. Ent. Soc. Lond., Vol. 2, Ser. 2) that the Pieride of the Amazon valley generally are open-ground butterflies, two genera only, Lep- talis and Terias, being true denizens of the forest. He also remarks that most of the species of Heliconia prefer the forest shades, It is also of in- terest to note that he comments upon the in ievousness of some species at least of Ithomia, in which connection one should recall the observation of W. J. Kaye already quoted. 284 THE AMERICAN NATURALIST [Von LI group is a combination of hues well suited upon the aver- age to render them inconspicuous in such places as they commonly frequent. If this be so, the initial step toward the production of new cases of mimicry might be any one of many variations in mode of nutrition or reproduction, which would lead repr tatives of the first family to spend their lives after the manner of the second. Reason has already been given for believing that convergence in color would probably accompany or follow convergence in habit. _ The new colors would undoubtedly appear in patterns largely determined by and reflecting the Pierian ancestry. Among fishes, as has been stated, a primitive color pat- - tern peculiar to one or common to several closely related families is sometimes readily recognizable, in which dis- tinct elements are apparent, now definite, now diffuse, mere stains of dyes that are not permanent. It is to be expected no less in insects that the family patterns, like finely wrought ornaments cast into the melting-pot, will be reshapen and serve new purposes. But from the welter of change and recombination which this involves may come once in many times a new grouping of characters, which suggests the pattern of another race. At this point natural selection directed toward the production of a pro- tective design painted in colors appropriate to the en- vironment may yield to selection working in the direction of resemblance. If so, a new pattern may be developed in the same protective colors and coupled with such change in the shape of the wings, or in other characters, as confers the additional advantage of being mistaken for a species which enjoys some measure of immunity. Either in organization or development most animals - give evidence of changes in habit much greater than the initial one herein postulated. Yet admit that these may occur, and what is already partly proved, that color is correlated with habit throughout the animal kingdom, and a theoretical difficulty that has engaged the attention of adherents and opponents of the mimicry hypotheses van- No. 605] ANIMAL COLORATION 285 ishes. No matter how wide the original gap between mimic and model, it may be bridged; no matter what de- gree of similarity between two forms may be necessary before natural selection may become effective in heighten- ing their resemblance, it may be attained without appeal to chance that is wholly blind, for there appears to be an automatic feature in the mechanism which has hitherto escaped observation. The ideas outlined in the preceding pages are neither a pure product of reflection nor a compromise suggested by an examination of the literature upon mimicry. They are a normal outgrowth of studies which had no preconceived relation to the problem of mimetic resemblance. They constitute a working hypothesis, and as such are sub- mitted to those biologists, particularly entomologists, who may have opportunity to test them rigorously. NUCLEUS AND CYTOPLASM AS VEHICLES OF HEREDITY! L. C DUNN BUSSEY [INSTITUTION Turre have been of late several attempts to effect a compromise between theories of heredity through the cytoplasm and theories which regard the chromosomes as the vehicles of inherited characters. Conklin (’08) was the first to suggest that egg, embryonic and general phyletic characters of any stage of the organism were determined in the egg cytoplasm while the determiners in the chromosomes made their presence known only through the specific or individual adult characters. Shull (1916) has elaborated this suggestion, and has brought to its support not only the older data on maternal inheritance, matrocline hybrids and the facts of development which relate to polarity, symmetry and organ-forming sub- stances, but has added new evidence of his own from ex- periments with rotifers. Most recently, Loeb, in his book ‘The Organism as a Whole,’’ has advanced a similar com- promise theory, based on similar evidence. Before examining in detail the experimental basis for such a compromise, it is important that the terms to be used be clearly and unmistakably defined. The first of these is the word ‘‘determined.’’ That a character is de- _ termined in the germ cell means that the differential, causal antecedents of that character are present in the germ cell. It does not mean that the character itself is present in the germ in any form, but rather that it is _ represented by substances or forces which not only stand _for the character but in some way bring about its expres- sion. : | 1A review made at the suggestion of Dr. H. W. Rand to whom grateful acknowledgement is due. : 286 * No. 605] NUCLEUS AND CYTOPLASM 287 If this definition of ‘‘determined’’ is accepted, two kinds of continuity in organisms are immediately differ- entiated. The first sort may be called substantial con- tinuity. It is the carrying over from one generation to the next of autonomous organizations of protoplasm in a manner analogous to the carrying of the bacilli of certain diseases (e. g., syphilis) in the germ cell. Here the germ cell is a passive vehicle. The character is present, not determined; and its changes from fertilized egg to adult are mere proliferations. If hereditary characters were to be so viewed, and the view carried to its logical con- clusion, the result would be something very like an ‘‘em- boitement’’ theory, which facts of development have proved to be untenable. Substantial continuity is hence only a concomitant rather than a part or a method of heredity. The second type of continuity may be called ‘‘genetic”’ continuity, and characters which are genetically con- tinuous are those which show a new coming into being with every generation. They are developed anew, and their resemblance to homologous characters in the pre- ceding generation is due to their development not from those characters but from homologous determinants. Characters of this type are truly determined and all hereditary characters are reducible to this type whether they are exhibited in egg, sperm, embryo or adult. It is now possible and desirable to define the expres- sions ‘‘inheritance through the ecytoplasm’’ and ‘‘inher- itance through the chromosomes.’’? The first properly means that the locus of the determiners or representa- tives of a character is the cytoplasm, and since it is the egg alone which contains any significant amount of cyto- plasm, the expression usually means the presence of these — determiners in the egg cytoplasm. ‘‘Inheritance through the chromosomes’’ means that the chromatic substance of the nucleus is the locus of determiners, and since the nuclear content of egg and sperm is equivalent this must also mean an equal determinative share by egg and sperm in heredity. 288 THE AMERICAN NATURALIST [Vou. LI Are now both of these theories compatible with one definition of ‘‘determined’’? Are they both possible and both necessary? Loeb has stated the problem and the ‘‘compromise’’ in these words (1916, p. 245): Question: “ Is the organism nothing but a mosaic of hereditary char- acters determined essentially by definite elements in the chromosomes; and if this be true what makes a harmonious whole TEREN out of this mie a assortment? ” Answ . . . the cytoplasm of the egg is the future embryo in the rough, aad the factor of heredity in the sperm only act by impressing the details on the rough block.” Shull’s statement is as follows (p. 6): The cytoplasm often (perhaps usually) determines the type of cleav- age, the early course of development, and in large measure the larval characters, while the adult characteristics are determined by the chromo- somes Conklin’s conclusions are (1915, p. 176): There is no doubt that most of the differentiations of the egg cyto- plasm have arisen during the ovarian history of the egg and as a result of the interaction of nucleus and cytoplasm; but the fact remains that at the time of fertilization, the hereditary potencies of the two germ cells are not equal. All the early stages of development, including the polarity, symmetry, type of cleavage, and the pattern or relative posi- tion of future organs, being foreshadowed in the cytoplasm of the egg cell while only the differentiations of later development are influenced by the sperm. In short the egg cytoplasm fixes the general type of development and the sperm and egg nuclei supply only the details. We are vertebrates because our mothers were vertebrates and produced eggs of the vertebrate pattern—but the color of our skin, eyes and hair . were determined by the sperm as well as by the egg from which we came, The same author has reiterated and somewhat elaborated the same views in an unpublished paper presented before the National Academy of Science in November, 1916.! 1 Since the above was written Conklin’s paper (1917) has been published. In its closing paragraphs he modifies materially the view which he had earlier expressed, admitting that cytoplasmic differentiation of the egg-cell probably arises under nuclear influence exerted equally by the egg and the sperm nuclei of the previous generation, the view mainfained by the present writer. No. 605] NUCLEUS AND CYTOPLASM 289 The evidence and a criticism of parts of it follows: 1. SHULL’s EVIDENCE (a) Cases of maternal inheritance. Under this head- ing Shull places such experiments as those of Correns (1909) on Mirabilis jalapa var. albomaculata. This plant — the common four-o’clock—has variegated leaves, green and white, the white being due to inhibited development of green in the chromatophores. The amount of green and white varies in different plants and furthermore whole branches may be green while other whole branches may be white. Flowers borne upon green branches, if self-fertilized, give seed that produces only green off- spring. Flowers from white branches, if selfed, give seed that produces only whites, which die because they are unable to carry on photosynthesis. Flowers on vari- egated branches yield offspring some of which are green, some variegated. Crosses among these green, white and variegated plants reveal the fact that the offspring re- semble invariably the female parent. White females pol- linated by any green or variegated pollen yield only whites which die. Green females pollinated by white or variegated pollen yield only green descendants. The paternal character never reappears in subsequent gen- erations. : Correns has assumed in explaining these remarkable occurrences that a disease transmitted by the cytoplasm of the ovule is the cause of the color differences, inasmuch as the white color in either self condition or as mottling on the green is a pathological condition. The chromo- somes are assumed to be immune to this disease. If the disease is caused by a germ (which is very likely) this germ acts only on the chromatophores and may well be passed through the egg like the germs of syphilis and other diseases. If this is the case, true heredity is not involved. The egg has simply acted as the passive bearer of a foreign body. And if the disease is due to a defect in the chromatophores, the case is not very different. The 290 THE AMERICAN NATURALIST [ Vou. LI chromatophores, as Shull says, are probably ‘‘autonomous bodies arising only from other autonomous bodies like themselves.’’ On such a view they are simply structures enclosed within the cytoplasm, having a continuity parallel with but independent of the continuity of inherited char- acters. (b) Shull’s next evidence is drawn from experiments resulting in so-called ‘‘matrocline hybrids,’’ which he de- fines as ‘‘unequal reciprocal hybrids which resemble the mother more than the father.’’ Under this head he cites the well-known cases of species and genera crosses among echinoderms. He says of the first generation from sea- urchin 2 X starfish £ (Loeb, 1903), and from sea-urchin 2x crinoid ¢ (Godlewski, 1906) that the embryos were purely maternal in character. In contrast to this, other observers of species and genera crosses among echino- derms (Morgan, Boveri, Baltzer and Herbst) have de- scribed the F, embryos as intermediate between the parents wherever there was normal union of maternal and paternal chromatin. Shull emphasizes especially the maternal character of embryos possessing no maternal chromatin. These were produced by Godlewski, by fer- tilizing enucleated fragments of sea-urchin eggs with crinoid sperm. The larval stages were reported to have been purely of the sea-urchin type. However, Boveri, Bierens de Haans and Herbst have obtained results which showed either the reverse condition, viz., paternal em- bryonic characters, or else intermediate larve. Moreover in the cross fertilization of giant sea-urchin eggs pos- sessing twice the normal amounts of cytoplasm and chro- matin, the larve inclined to the maternal side. That this was not due to the doubled amount of cytoplasm was demonstrated by Boveri, who fertilized nucleated half- fragments of normal eggs with sperm of another species and found no paternal inclination from the halved amount of cytoplasm. That the phenomena are due rather to the chromosomes is indicated by the same experimenter’s work with dispermie fertilization. When more than one No. 605] NUCLEUS AND CYTOPLASM 291 sperm enters the egg abnormal mitotic figures and ab- normal chromatin dieteibntii are directly correlated with abnormalities in the larvæ, although the egg cyto- plasm remains constant. Baltzer’s fine work on species crosses may also be cited as showing that Loeb’s and Godlewski’s observations penetrated only part way toward the truth. As an example, only one of Baltzer’s crosses need be cited. When eggs of Spherechinus were fertilized by sperm of Strongylocentrotus, the mitotic figures and dis- tribution of chromatin were normal and the larvæ were intermediate, not maternal. From the reciprocal cross ‘‘matrocline hybrids’’ (most of them abnormal) did re- sult, but their maternal resemblance was not due to cyto- plasm. That it was due to irregularities in the chromatin distribution was proved by Baltzer, who followed the his- tory of the maternal and paternal chromosomes in the hybrid embryos. He found that the majority of the paternal chromosomes (15 out of 18) were inactive at the first cleavage. They were extruded from the develop- ing odsperm nucleus, and degenerated. The cells of the hybrid had then 21 chromosomes, 18 maternal and 3 pater- nal, and their maternal resemblance is easily explicable on these grounds. It is not explicable on any other for no abnormal conditions obtained in either cytoplasm or the surrounding medium. Herbst repeated the first of Baltzer’s crosses (Spher- echinus 2 X Strongylocentrotus 3) but chemically induced parthenogenetic development in the egg before the en- trance of the sperm. Thus the ¢ pronucleus was behind at the first division and failed to be incorporated into the nucleus of one of the first daughter cells. One side of the developing hybrid had, then, merely paternal chromo- somes, while the other had both maternal and paternal, and in striking sequence to this distribution were embryos which had only paternal characters on one side, while those of the other side were intermediate. On the evidence thus far, Shull himself has not placed 292 THE AMERICAN NATURALIST [ Vou. LI the maximum of emphasis, and the foregoing criticism has been intended to indicate that its support of the cyto- plasmic view has been and may continue to be so mini- mized as to be non-existent. However, Shull does place much emphasis on some carefully collected evidence of his own. This does not, I believe, support his theory to any greater extent than his quoted cases. The evidence is briefly as follows: Shull crossed two lines of rotifers which differed in two egg characters— time of hatching of sexual eggs, and the proportions of sexual egs which actually hatched. The eggs of line “A” hatched on the average in 1-2 weeks, about 50 per cent. emerging. Line ‘‘B’’ eggs took 5-6 weeks to hatch and only 5 per cent. emerged. Line ‘‘A’’ females fertilized by line ‘‘B’’ sperm laid eggs which hatched in 1-3 weeks, 50 per cent. emerging. The eggs thus resembled the mother’s line in both respects. Line ‘‘B’’ females fer- tilized by line ‘‘ A’’ sperm laid eggs, 30 per cent. of which emerged in 4-5 weeks, resembling more the mother’s line than the father’s. The reciprocal hybrids are thus very unequal, says Shull, and since in crossing parents which differ by Mendelian, chromosome-determined characters, the resulting reciprocal hybrids are equal, the characters under observation are non-Mendelian and determined in the cytoplasm of the egg. But it is to be objected that in reality these hybrid eggs of the first generation are not hybrid in these characters at all. The characters are egg characters and as such can be exhibited only when the hybrid zygote produces its eggs, not when the hybrid zygote is formed. The expres- sion of the character is thus delayed until the hatching of eggs laid by these F, zygotes. And Shull’s data shows this to be the case. But when new lines were obtained from these hybrid (F,) eggs, and these lines produced sexual eggs of their own, the two reciprocal hybrid lines were fully equal. Now the usual occurrence is observed, viz., the reciprocal hybrids are equal and the contributions to the character No. 605] NUCLEUS AND CYTOPLASM 293 by g and 2? are proved to be also equal. Shull’s conten- tion for the participation of the egg cytoplasm rests en- tirely on the maternal character of the first egg-genera- tion. These eggs were matured, since the mother was homozygous in the characters, under the influence of the like chromatic contributions of her parents; the hybrid mother matured hers under the influence of the unlike chromatic contributions of her parents and showed the participation of her paternal chromosomes only in the behavior of her eggs. The peculiarities of the case lie not in that we are dealing with a cytoplasmically deter- mined character, as Shull contends, but in (1) the fact that the characters are exhibited only by females; (2) in the fact that the characters are egg-characters, which places segregation one generation farther away from the original cross. The case is quite analogous to the case of the inherit- pace of red pericarp color in corn, which, although a ‘maternal character,” was shown to Mendelize by Lock (706). It is also comparable to the egg-character ‘‘uni- bivoltinism’’ in silk moths, which Castle (’10) proved from Miss McCracken’s data to be a Mendelian character. These cases will be discussed more fully in later para- graphs. Shull’s conclusion that cytoplasm determines egg and larval characters is, I believe, unnecessary. It has been shown that characters exhibited only by females and only in the egg may be equally determined both in the egg and in the sperm. The sperm contribution being predomi- nantly chromatic, and the chromosomes being the ac- cepted carriers of the determiners of other characters, it is to be concluded that the determiners for the characters investigated by Shull are also to be sought in the chromo- somes. Loeb, in his book (1916), has given considerable space to this question, closely following the earlier treatment by Conklin. His conclusion which has been stated and the evidence on which it rests may be briefly criticized by the 294 THE AMERICAN NATURALIST (Von. Ll addition of further evidence which warrants a changed interpretation of certain facts. Loeb first calls attention to the exclusively maternal character of the early development of enucleated frag- ments of eggs when fertilized by sperm of a different species. Such evidence has already been treated above (p. 5). His second claim—that the rate of cell division and de- velopment is determined only by the egg cytoplasm— warrants further consideration. An egg from a line in which segmentation of the egg takes place eight hours after fertilization was fertilized by sperm from a line in which segmentation begins in 30 minutes. The rate of these cross-fertilized eggs was 8 hours, like the mother’s line. The careful and long continued work of Newman (714) has, on the other hand, shown that the entrance of sperm of a different species does materially alter the rate of de- ~ velopment. In both reciprocal crosses between Fundulus heteroclitus and F. majalis the rate of development of the hybrids was intermediate between that of the two parent species. This was true of cleavage rate, rate of germ-ring formation, ete. In the cross F. heteroclitus X F. diaphanus ‘‘both recip- rocal crosses have a higher rate than the pure bred strain, Similarly, when we make reciprocal crosses between Cy- pronodon and any species of Fundulus we find a marked retardation in developmental rate in both crosses. .. .’’ It is of the greatest significance that in all three cases the results of reciprocal crosses were equal. Either both were intermediate, both were accelerated or both were retarded regardless of which species was used as the egg- parent. In the face of such evidence, a theory of exclusive control of the egg over early development is untenable. Newman’s fundulus hybrids, while demonstrating the conclusion just stated, do not form critical evidence for determining the action in heredity of such rate-characters because only the F, generation is known. Another series No. 605] NUCLEUS AND CYTOPLASM 295 of experiments on a hatching time character has, how- ever, been carried through the F, generation and as an illustration such a case may be cited in detail. It consists of Miss MeCracken’s (1909) experiments with silk moths. Castle later (1910) called attention to some facts in her data which indicated that although a female-exhibited character and confined to the egg in its expression, it nevertheless gave evidence of Mendelizing in crosses. Toyama (712) concluded that dominance was present, and both of the latter investigators agreed that the males, al- though unable to exhibit the character, gave evidence by their genetic behavior of having an equal determinative influence with the females. The data follows: Silk moths lay one batch of eggs, always in the spring. The eggs of some batches hatch out immediately, pro- ducing another brood of larve and moths in that season. The parents of such batches of eggs are hence known as bivoltins. The eggs of other batches do not hatch for twelve months, and since in this way there is but dne brood or flight each season, the parents of such eggs are known as univoltins. If a univoltin female is crossed with a bivoltin male, the spring batch is laid as usual and hatches in 12 months. This is just what would have oc- curred if the mother had been fertilized by a male of her own sort. When these eggs hatch, a hybrid brood emerges which lay their egg batches immediately but the univoltin character is again exhibited in that all of these eggs are of the 12 months type. But, these eggs now differ among themselves as is shown by the behavior of the zygotes which emerge from them. Some of these females are bivoltin, laying eggs which develop immediately, while others are univoltin, laying eggs which hatch the follow- ing spring. The expression of the paternal contribution is delayed, but its activity in determining the time of hatching is quite apparent. The inheritance of red pericarp color in corn follows exactly the same course as that outlined above, with red dominant over white. The F, embryos must be raised 296 THE AMERICAN NATURALIST [Vou. LI before the segregation of pericarp color among them can be seen, for it is exhibited only in the seed coats. The conclusions follow: (1) The egg and all its determinative content is produced under the double influence of the sperm and egg chromatin contributions which united to form the producing zygote. (2) Hybridization experi- ments with egg characters, to be critical, must be carried as far the F, generation. (3) In all experiments which I have seen reported, in which this condition obtained, the influence of the sperm on the characters in question has been observed. Loeb’s evidence, however, introduces also crosses be- tween Strongylocentrotus purpuratus and S. francis- canus, and the statement is made that the development of the hybrid up to the formation of the skeleton resembles exclusively the development of the mother’s species. But Loeb also finds that the cross-barring in the spicules of purpuratus behaves as a dominant character in reciprocal crosses. He assigns this character to a factor, which he imagines to be a ferment or enzyme. This statement fol- lows: ‘‘Since the pure purpuratus has two determiners for the development of the cross-bars, and the hybrids only one, the pure purpuratus should have twice the en- zyme and develop twice as fast’’—and it did. He pro- vides here not only evidence that avowed chromosome characters do affect the rate of development, but even furnishes an enzymatic mechanism by which they may do it. And yet soon after the above quotation, we read: We can therefore be tolerably sure that wherever we deal with a hereditary factor which is determined by the egg alone, the cytoplasm of the latter is partly or exclusively responsible for the result. We have already mentioned that rate of segmentation is such a character. The whole case of the supporters of any theory which views the cytoplasm as determinative rests on either their refusal to go back and inquire the source of this cyto- plasm, or on their refusal to give due emphasis to the source, even though they recognize it. Conklin recognizes No. 605] NUCLEUS AND CYTOPLASM 297 the double influence which is exerted on the developing egg better than any of the others who have adopted his ‘‘ecompromise theory.’’ He admits that ‘‘most of the dif- ferentiations of the cytoplasm have arisen during the ovarian history of the egg and as a result of the inter- action of nucleus and cytoplasm.’’ He has demonstrated better than any other one man how complex and definite these differentiations in the egg cytoplasm are. All will agree with him when he says that they ‘‘foreshadow”’ the future organism. But ‘‘cytoplasmic organization, while affording the immediate conditions of development, is itself a result of the nature of the nuclear substance which represents by its inherent composition the totality of heritable potence.’’?’ These last are the words of E. B. Wilson (1895, p. 25), although he has translated and adapted them from an earlier paper by Driesch. They represent the opinion of Wilson and of Driesch in full accord. ‘‘The nuclear substance’? referred to was even then known to contain equality of maternal and paternal chromatin. j Wilson himself had been able to demonstrate that the structure of the cytoplasm in sea-urchin eggs was ac- quired during ovarian life, and on the basis of this and of a considerable body of similar evidence he was able to conclude quite definitely: That a preorganization of the cytoplasm can not be regarded as the primary factor in heredity is conclusively proved by the old argument based on inheritance from the father through the sperm nucleus. The only link which is needed to make the chain com- plete is some substantial body of evidence, demonstrating the effect and the mechanism of action of the nucleus on the cytoplasm. This, it must be admitted, has not been entirely filled. Nägeli, to be sure, held a theory of a dynamic effect of the nuclear idioplasm on the cytoplasm, while Driesch contended that the mechanism was chem- ical. The nucleus, in his opinion, exercised its governance by means of ferments or enzymes.. There are facts in r 298 THE AMERICAN NATURALIST [ Vou. LI development which point to effects of the nucleus even on the visible differentiation of the egg before fertilization. In the sea-urchin, for instance, this differentiation is pre- ceded by the absorption into the nucleus of part of the fluid content of the cytoplasm, altering the chemical com- position of the latter and greatly increasing the bulk of the nucleus. The membrane of this enlarged nucleus then dissolves and part of its contents by their color may be traced to a clear cap of fluid which later gives rise to the skeleton of the echinoderm. Such absorptions and ming- lings probably play a large part in the reactions of nucleus and cytoplasm, either at the successive disap- pearances of the nuclear membrane during mitosis or through that membrane. Nucleus and cytoplasm may certainly be regarded as forming a reaction system analogous to that which might exist between a series of chemical substances (Jennings, 1914). The cytoplasm in turn is linked with the extra- cellular milieu in a quite comparable way and forms the intermediary between the nticleus and the exterior. Evidence on this interaction is accumulating. As an example I may quote the work of Cameron and Gladstone on cells other than ovarian, and, to be sure, by the static, histological method. But they have observed fine prep- arations and have concluded that the cytoplasm is visibly differentiated into two grades of endoplasm. The first is next to the nucleus, is clear and refractive. This is the nascent material of the cell and is the first visible stage in the genesis of protoplasm. It is, they postulate, a derivative of the nucleus itself, and to the nucleus is ascribed the final elaboration of nutritive material which has been ingested by the cytoplasm. This nascent endo- plasm. they conceive to be the active material of the cell grading into passivity toward the cell periphery, through the maturer endoplasm and the ectoplasm. Whatever relations may exist between the two, the fact remains that the cytoplasm is necessary. Without it the nucleus, deprived of its milieu, can not live, and develop- No. 605] NUCLEUS AND CYTOPLASM 299 ment can not take place. The investigation of the finer physiological reactions which take place between the nucleus and cytoplasm is badly needed, and the restate- ment of them in terms of physics and chemistry. Such evidence as is available indicates that the importance of the cytoplasm is, in the main, subordinate to that of the nucleus. The evidence from egg-characters, it might be noted in conclusion, is one-sided. I have no doubt that if sperm- characters were to be studied as intensively as egg-char- acters have been (which has not been the case due to the microscopic size of most sperms) the differential char- acters in the sperm would be found to behave in heredity like the differential characters of eggs, and would be de- termined as largely by the egg nucleus as by the nucleus of the sperm of the preceding generation. CONCLUSIONS Direct continuity of substances in the cytoplasm is not a method of heredity. It simply provides for the autono- mous proliferation of materials with no determinative significance. No compromise, then, is possible between the two views outlined as ‘‘eytoplasmic’’ and ‘‘chromo- some’’ theories of heredity. The first is non determi- native; the second is the primarily effective method of heredity and of development. The working of the ef- fective method is known for heredity, if heredity be prop- erly only concerned with the way in which the hereditary factors are distributed in the germ cells. For develop- ment, its mechanism is but grossly known, but we have learned enough of the determinative effect of the nucleus and of the possibilities for interaction between cytoplasm and nucleus to foster a suspicion that one day the gov- ernance of the chromosomes over development will be ex- plained in PERTE terms. BIBLIOGRAPHY Cameron and Gladstone. 1916. Journ. Anat. and Physiol., pp. 207-227. Castle, W. E. 1910. Journ. Exp. Zool., Vol. 8, No. 2. 300 THE AMERICAN NATURALIST [Vou. LI Conklin, E. G. 1908. Science, N. S., 27. Conklin, E. G. 1915. Heredity sa Tao. Princeton Univ. Press. Conklin, E.G. 1917. Proc. Nat. Acad. Sci., Vol. 3, No. 2. Correns, C. 1909. Zeit. tn ae u. Vererb., 1 and 2. Godlewski, E. 1906. Arch. Ent. ch., 20. Lock, R. H. 1906. pose Roy. Bot. Gara, Peredeniya, 3. Loeb, J. 1903. Univ. Cal. Pub. No. Loeb, J. 1916. The eee as a Woot, nr N Y, Newman, H. H. 1914 . Exp. shop McCracken, I, 1909. ce urn. tas Zoo oe T ate 4, hull, A. F. 1916. Ohio Journ. of or p DERA Toyama, K. 1912. Biol. Centralbl., Vol. 32. Wilson, E. B. 1896. Arch, Ent. Me oh: Vol. 3. SHORTER ARTICLES AND DISCUSSION MODIFYING FACTORS AND MULTIPLE ALLELO- MORPHS IN RELATION TO THE RESULTS OF SELECTION In the prevailing controversy as to the effectiveness of selec- tion, those who reject such effectiveness put forth multiple modi- fying factors as the explanation of the results observed. The given character (for example, the coat color in Castle’s hooded rats) is held to depend on one main factor (determining the presence or absence of the character) and upon numerous modi- _ fying factors; the number of the latter present in a given case determines the degree of expression of the character. Selective breeding is then held to act, as it does with relation to all other Mendelian factors, merely by making diverse combinations of such factors. Many are gathered together in certain individuals, few in others, and the degree of expression of the inherited char- acter varies accordingly. : Much evidence is presented that this is actually the mode of operation in many cases where selection is effective. Thus are explained the visible results of selection in Castle’s rats; thus the unexpected fact that many of the mutations of Drosophila have shown themselves (in aecordance with Castle’s prediction) to be amenable to selection, although in other respects they be- have like alterations of a single unit factor. This is indeed the usual explanation for that effectiveness of selection which is coming to light in so many eases; and in many of the cases it ap- pears clear that the explanation is correct. But in which direction does this explanation carry us? To answer this question we must know what these multiple modi- fying factors are. If they are mere examples of the statie con- - dition of diversity observed among so many closely related or- _ ganisms, the answer falls in the negative direction, so far as _ the effectiveness of selection in accumulating actual alterations of hereditary characters is concerned. But if they should them- selves turn out to be the actual alterations of hereditary consti- tution that are accumulated by selection, then the answer con- 301 302 THE AMERICAN NATURALIST [Vou. LI firms the effectiveness of selection and adds greatly to our knowledge of how it is brought about. What are the facts? For these we may turn to the organism of which the genetics are best known; to the fruit fly Drosophila; and we may accept the accounts presented by those most uncompromising opponents of the effectiveness of selection, the investigators of Drosophila in the Columbia laboratory. Their account we cannot suspect of being colored to favor the selectionist point of view. We find data as to certain known modifying factors in the recent im- portant paper of Bridges (1916) on ‘‘Non-Disjunction of the Chromosomes in Drosophila.’’ Bridges tells us that he has found no less than seven diverse factors that modify the single primary grade of eye color known as eosin. These seven factors are located in parts of the chromosomal apparatus different from the spot on which the presence or absence of eosin depends, and each is inherited in Mendelian fashion. One of these factors lightens the eosin color in a fly with eosin eyes, nearly or quite turning eosin to white; this factor Bridges calls ‘‘whiting.’’ An- other has the effect of lightening the eosin color a little less, giving a sort of cream color; this is called ‘‘cream b.” A third factor dilutes the eosin color not so much; it is called ‘‘cream ”? Tn addition to these, Bridges tells us that he has discovered three other diluters of the eosin color; we will call them the fourth, fifth and sixth diluters. And finally Bridges tells us of another factor whose only effect is to make eosin darker; this factor he calls ‘‘dark.’’ We get therefore the following list of the modifying factors for eosin color: 1. Whiting 2. Cream b 3. Cream a 4. Fourth diluter 5. Fifth diluter 6. Sixth diluter 7. Dark = We have then in Drosophila minutely differing conditions of ’ a single shade of color, brought about by seven modifying factors. Concerning these, Bridges makes the following remark, which is worthy of particular attention: A remarkably close imitation of such a multiple factor case as that of Castle’s hooded rats could be concocted with the chief gene eosin No. 605] SHORTER ARTICLES AND DISCUSSION 303 for reduced color, and these six diluters which by themselves produce no effect, but which carry the color of eosin through every dilution stage from the dark yellowish pink of the eosin female to pure white (Bridges, 1916, p. 149). Thus we see that in Drosophila we could get the same sort of graded results that Castle does with his rats, only in Drosophila this is by multiple modifying factors, whereas Castle believes that in the rat it is by actual alterations in the hereditary con- stitution ! But what are these modifying factors? And here we come to the astonishing point. These modifying factors are themselves alterations in the hereditary constitution! Bridges leaves no doubt upon this point. He lists and describes them specifically as mutations; as actual changes in the hereditary constitution. Now so far as I can see, the question involved in the selection controversy is as to the occurrence of minute changes in the hereditary constitution, and their accumulation by selection; so that by selection various grades of a given external character can be obtained. In Drosophila, according to Bridges, such changes occur; changes which give, so far as our present imper- fect knowledge goes, at least seven diverse grades of a single tint (that is itself, as we shall note, only one grade in another series of seven known grades). By means of these graded. changes one could obtain, by the mutationist’s own statement, the continuously graded visible results which selection actually gives. Is not then the controversy as to the effectiveness of selec- tion at an end? As to just where the graded hereditary changes occur there remain indeed certain differences of opinion; some selectionists, like Castle, hold that the various grades of a given external char- acter are due to diverse minute modifications of the same unit character—of the same locus in the chromosome; while, as we have seen, the modifying factors are due to changes in diverse parts of the hereditary material. This matter of detail does not - touch the main point, but it is of interest to ask what the work of the mutationists gives us on this question. It is curious to find that their studies of Drosophila furnish almost all that could be asked by the radical selectionist as to the existence of a single unit character in a series of numerous hereditary gradations. The best instance here is again the color of the eye, which fur- nished us also our example of modifying factors. The color 304. ~ THE AMERICAN NATURALIST [Vou. LI eosin, of which the modifying factors give us seven grades, is itself only one of another series of seven grades that .are due to diverse alterations in the same unit factor—in the same chromo- somal locus. As we know from the studies of Drosophila, this locus is a certain region of the X-chromosome. When this locus retains its normal condition the eye is red. Some years ago a variation was observed by which the eye lost its red color, be- coming white. Somewhat later another variation came, by which the eye color became eosin. By the wonderfully ingenious methods which the advanced state of knowledge of the genetics of Drosophila has made possible, it was determined that the mutations white and eosin are due to changes in a particular part of a particular chromosome, and that indeed the two are due to different conditions of a particular region of the X-chro- mosome. In other words, they show different conditions of the same unit. Moreover the normal red represents a third con- dition of this same unit. Later a fourth condition of this same unit was found, giving a color which lies nearer the red, between the red and eosin; this new color was called cherry. We have now four grades of this single unit character. And now, with the minute attention paid to the grades of eye color, new grades begin to come fast. In the November number m Genetics, Hyde (1916) adds two new grades, one called ‘blood, ”” near the extreme red end of the series; the other, called ‘‘tinged,”” near the extreme white end; in fact, from the de- scriptions, it requires careful examination to distinguish these two from red and white respectively. So we have now six grades of this unit. And in the same number of the same journal, Safir (1916) adds another intermediate grade, lying between tinged and eosin; this he calls ‘‘ buff.’’ So, up to date we know from the mutationists’ own studies of Drosophila that a single unit factor presents seven gradations of color from white to red, each grade heritable in the usual Men- delian manner. These grades or ‘‘multiple allelomorphs’’ as they are called, are the following: 1. Red 2. Blood 3. Cherry 4. Eosin No. 605] SHORTER ARTICLES AND DISCUSSION 305 5. Buff 6. Tinged 7. White Three of these seven grades have been made known to us within the last six months. It would not require a bold prophet to pre- dict that as the years pass we shall come to know more of these gradations, till all detectable differences of shade have been made out and each shown to be inherited as a Mendelian unit. Con- sidering that the work on Drosophila has been in progress but eight or nine years, we have already remarkable progress toward a demonstration that a single unit character may present as many heritable grades as can be distinguished; that the grades may give a pragmatically continuous series. This is precisely the situation that the selectionist postulates. Furthermore, as we have seen, besides this primary series of seven grades, due to alterations of a single unit factor, there is a secondary series containing seven more grades, all affecting the central grade (eosin) of this primary series, but due to al- terations of other parts of the germinal material. How much more does the selectionist require? This situation in Drosophila is not exceptional. To mention one or two other examples, Castle and Wright (1916) find a large series of such diverse conditions of a single factor (‘‘mul- tiple allelomorphs’’) determining various shades of coat colors in rodents. Emerson (1917) in his recent account of the extra- ordinary condition of affairs in the genetics of pericarp colors in corn, talks of ‘‘a series of not less than nine or ten multiple allelomorphs,’’ which moreover leap back and forth from one condition to another in bewildering fashion. To sum up, it appears to me that the work in Mendelism, and particularly the work on Drosophila, is supplying a complete foundation for evolution through the accumulation by selection of minute gradations. We have got far away from the old notion that hereditary changes consist only in the dropping out of complete units, or that they are bound to occur in large steps. The ‘‘multiple allelomorphs”’ show that a single unit factor may exist in a great number of grades; the ‘‘multiple modifying factors” show that a visible character may be modified in the finest gradations by alterations in diverse parts of the germinal material. The objections raised by the mutationists to gradual change through selection are breaking down as a result of the thoroughness of the mutationists’ own studies. 306 THE AMERICAN NATURALIST [Vou. LI The positive contribution of these matters to the selection problem is to enable us to see the important rôle played by Men- delism in the effectiveness of selection. Hereditary variations, such as give rise to the multiple allelomorphs and multiple modi- fying factors, occur in some organisms rather infrequently, as measured by the time scale of human happenings. If there were no interchange of factors among individuals and stocks, it would take a long time to obtain in one individual all the six diluters of the eosin color of the Drosophila eye; one arises in one in- dividual, another in another. But by selective crossbreeding it is possible to bring together into one stock all the modifiers that have been produced in diverse stocks. Mendelism acts as a tremendous accelerator to the effectiveness of selection. PAPERS CITED Bridges, C. S. 1916. Non- aen as Proof of the Chromosome Theory of Heredity. Genetics, 1: 1-52, 107-1 63. Castle, w. E., and Wri i S. 1916. Studies of Inheritance in Guine Pigs EN Rats.. Publ. No. 241, RREN Institution of Washington, 192 pp. l Emerson, R. A. 1917. Genetical Studies of Variegated Pericarp in Maize. 35. Hyde, R. R. 1916. Two New pina ‘a Sex-linked Multiple (sex- H aa oin System. Genetics, 1 580. fir, S. B. 6. Buff, a New ew ATlelomorpl p White Eye Color in Dro- sophila. ae 1: 584— H. S. JENNINGS THE JOHNS HOPKINS UNIVERSITY A WING MUTATION IN PIOPHILA CASEI In the early part of December, 1915, I began to breed the ‘‘cheese skipper’’ Piophila casei, in order to see if mutations were to be found in this fly. The source of my stock was a small piece of Italian cheese containing a dozen or so larve.t As these were doubtless the offspring of one female, inbreeding has been very close. Up to June 22, 1916, only one heritable mutation had been found among the thousands of individuals bred; this was the wing defect described below, which was first noted on March 12, 1916. 1 This work was ee on at the ay Zoological Laboratory, Yale Uni- versity, New Haven, Connecticut. It was in New Haven that I obtained the cheese. Contribution No. 135, Zoological Laboratory, University of Texas. No. 605] SHORTER ARTICLES AND DISCUSSION 307 The Defect—When viewed from the dorsal surface the defect appeared as a blister of variable size on the proximal and posterior part of the wing. From the ventral surface it appeared as a pit. Occasionally a real blister filled with fluid was ob- tained. The position of the defect was constant; when small it lay in the posterior cell just below the discal cell. When large it involved nearly the whole wing including the axillary, anal, second basal, discal and posterior cells. Usually both wings were affected alike, but here and there flies were found with one wing normal and the other wing severely affected. This factor is strikingly similar, both in its appearance and the variability of its behavior, to the ‘‘balloon wings’’ found by Morgan? in Drosophila and more recently fully described by Marshall and Muller.* The flies carrying the defect, in my cul- tures, were very frequently sterile, and in no case did their fertility begin to approach that of normal stock. In breeding, the character behaved as a mendelian recessive. Normal crossed with balloon gave, in the F, generation, 196 normal and no affected individuals. (This included 4 matings.) When brother and sister were mated, in the F, generation, 312 normal and 111 balloon offspring were obtained This is very close to the expected 3:1 ratio, of a monohybrid cross. When balloon flies were crossed, all individuals were affected (74 off- spring obtained) but the character showed itself extremely vari- able; in some cases the flies appeared normal until very closely examined. The defect was not sex-linked as is shown i the following mating. A defective female was mated with a normal male of normal stock. Of the 50 offspring resulting both males and females were normal. The variation in the appearance of the balloon flies suggested either that the size of the blister was dependent upon some unknown environmental factors, or else, was due to multiple allelomorphs or multiple factors. A great number of matings were made to gain light on this point, but due to the sterility of the affected individuals, the evidence is not sufficient to allow us to draw any conclusions. Two individuals both of whom were severely affected were crossed. The 20 offspring resulting were all severely affected. Two individuals, both of whom were only 2 In Morgan’s ‘‘A Critique of the Theory of Evolution.’’ 3 Marshall and Muller, Jour. of Exp. Zool., Vol. 22, 1917. 308 THE AMERICAN NATURALIST [ Von. LI slightly affected, were crossed. Of the 29 offspring resulting, 17 showed the defect in a severe form, and 12 showed only small blisters. A female which had only one wing affected, was mated to a male, one of whose wings was severely affected while the other bore-a very small blister. Of the 45 offspring resulting, 27 bore the defect in a severe form on both wings and 18 showed small blisters again on both wings. Further experiments with this new character were under way when the work was stopped by the mobilization of the Militia in June. The work with these flies, however, is again being resumed. THEOPHILUS S. PAINTER UNIVERSITY OF TEXAS A CASE OF REGENERATION IN PANULIRUS ARGUS! THE occurrence of regenerative processes in the crustacea has been a matter of record for a number of years, but the instances have been mostly confined to the regeneration of appendages and portions of the nervous system. Observations on the regeneration of portions of the exoskeleton of the trunk are far less numerous. The present observations on the regeneration of a portion of the rostrum of Panulirus argus, the common crayfish of the Bermuda Islands, were made during the summer of 1916 at the Bermuda Biological Station. Panulirus argus when full grown is about 14 to 16 inches in length. It lacks chelipeds, their place being taken by the ordi- nary type of walking appendage. None of the walking append- ages is provided with nippers, all being tipped with a single hook, as, e. g., in the fourth pair of appendages of the crayfish Cam- barus. The rostrum of Panulirus, instead of being a single median projection, -consists of a pair of long (30-35 mm.), sharply pointed spines, slightly compressed laterally, and grow- ing out from the carapace just posterior and slightly dorsal to the base of the eye-stalks. The animal in question was a half-grown male, eight and one half inches long. When caught, June 20, the left spine (com- pare figure and explanation) of the rostrum was entirely miss- ing. The carapace around the base was jagged and rough, as though the break had been recent ; but a thin, soft membrane had 1 Contributions from the Bermuda Biological Station for Research, No. 58. No. 605] SHORTER ARTICLES AND DISCUSSION 309 formed across the surface of the break. Five days later, June 25, the protecting membrane had hardened, so that it could not be dented with the point of a scalpel. No further change could be noted until after the molting, which occurred four days later, June 29. The casting occurred at night, and the next morning the new shell showed no signs of any wound. By one o’clock a very slight hump appeared, and by ten o'clock at night a little rudimentary spine 2 mm. in length had formed. The next morn- ing another millimeter had been added to its length. Meantime the normal spine had increased 1.5 mm. in length. No further growth followed before the new shell had hardened. Fie. 1. From a photograph of the left side of the head region of Panulirus argus, showing (norm.) the normal, and (regn.) the regenerated rostral spine. As the figure is reproduced from a r c print, the picture is reversed, the right spine appearing like a left Sixteen days later, July 15, another molt occurred. As before, the old shell was cast at night and by the following morning the regenerating spine had added 2 mm. to its length, being now 5 mm. long. By the next evening all growth had been stopped by the hardening of the new shell, but the total length of the spine was at this time 7 mm. The spine now showed a sharp point and also a slight lateral compression like that of the normal spine. At this casting the normal right spine added 1 mm. to its length, showing that, while the whole animal was growing, the 310 THE AMERICAN NATURALIST (Vou. LI regenerating part was increasing at a much faster rate than other parts. Thus in the period of twenty-seven days during which the ani- mal was under observation, it had undergone two molts and had regenerated a missing rostral spine of normal form, 7 mm. in length, while the normal spine had added 2.5 mm. to its length in the same period. These results show that the period between molts for this animal under laboratory conditions is six- teen days; that a rostral spine of normal form can be regen- erated; and that the rate of this regeneration was nearly three times the rate of normal growth of a similar spine during the same period. A. C. WALTON HARVARD UNIVERSITY NOTES AND LITERATURE THE COAL MEASURES AMPHIBIA OF NORTH AMERICA? THE excellent monograph by Dr. Roy L. Moodie is a worthy successor of the long series of works by Dawson and by Cope on the air-breathing vertebrates of the Coal Period in North America. The extremely varied amphibian fauna of the coal swamps as described by Moodie contains representatives of no less than 7 orders, 19 families, 46 genera and 88 species, the animals ranging in size from the minute Humicrerpeton, less than two inches long, to the great Leptophractus obsoletus, which was as large as an adult Florida alligator. The wide dif- ferentiation and high specialization of these amphibians shows that the class even at that early epoch had evolved very far from its first adaptive radiation, so that, as Dr. Moodie well observes, the origin of land vertebrates from fishes must be looked for in a much earlier time, perhaps the Silurian. Carboniferous Amphibia are reported from various localities in North America, but only four of these have yielded large or important collections. From the South Joggins coal mines in Nova Scotia Sir William Dawson secured most of the specimens of Microsaurs described by him, many of the skeletons being found in the rotten stumps of Sigillaria trees. This material is preserved chiefly in the Museum at McGill University, Montreal. From the Linton, Ohio, coal seams Newberry and his collectors secured the great collections described by Cope and which are now chiefly in the American Museum of Natural History. At Mazon Creek, Illinois, the fossils are found in ironstone nodules in a stratum of shale; the specimens have been described chiefly by Newberry, Cope and Moodie and are scattered in various museums. At Cannelton, Pennsylvania, the fossils occur in slates and have been described by Moodie, the material being in the National Museum. On account of the fragmentary nature of most of the material and the fact that generic and specific names have been based on 1 Carnegie Institution of Washington, Publication No. 238, 1916. 311 312 THE AMERICAN NATURALIST [Vou. LI many different and non-comparable parts of these animals, the author’s task was an exceedingly difficult one, and only those who have occasion to study this work very closely can appreciate either the magnitude of the undertaking or the thoroughness with which it has been carried out. Dr. Moodie’s monograph will NETE invite comparison with the well-known works on the Permian Amphibia of Bohemia and Saxony by Fritsch and by Credner. It must be admitted, however, that many of the illustrations are inferior to those of the works mentioned, partly on account of the difficulty of showing the real character of these fragmentary specimens by means of photographs. The author’s method is so intensive that he has left even readers who may have some first-hand knowledge of Paleozoic Amphibia in need of many broader facts and comparisons which may rea- sonably be expected to result from such a conscientiously exe- cuted investigation ; it is the aim of this review in some measure to supply this deficiency, in the hope that Dr. Moodie him- self may be induced to write a general article covering more fully the points here raised. In the chapter on stratigraphic and geographic distribution the author shows that the four chief Amphibia-bearing for- mations in North America mentioned above are all in the Alle- ghany or Lower Coal Measures and are thus much older than those deposits (Salt Fork, Pitcairn) of the Upper Productive Coal Measures at the top of the Pennsylvanian series, which have collectively yielded Cricotus, PNE Heyope and other genera characteristic of the Texas ‘‘Perm The author does not discuss the piste seria of the Lower Coal Measures fauna either with the ‘‘Permian’’ fauna of Texas and other states, or with the Carboniferous and Permian faunas of Ireland, Scotland, England, France, Saxony and Bohemia. Even the Permian and Triassic amphibian faunas of South Africa invite comparison with the varied Temnospondyli of the Carboniferous and Permian of America and Europe. The Lower Coal Measures fauna of America includes a long series of branchiosaurs, microsaurs and primitive labyrintho- donts (Spondylerpeton, Dendrepeton, Macrerpeton, Eoba- phetes), and it is totally lacking in pelycosaurs, poliosaurs, cotylosaurs, or any other reptiles except Eosauravus. The Texas fauna, on the other hand, has only a single microsaur (Crossotelos) and no branchiosaurs; its varied labyrinthodonts No. 605] NOTES AND LITERATURE 313 (ineluding Cricotus, Eryops, Dissorophus and many others) are all genera not found in the Lower Coal Measures, and it abounds in reptiles of several orders and many families. Some of this difference may be due to the fact that the Lower Coal Measures fauna represents only the life of the coal swamps, while the Texas fauna represents the life of the pools and streams of a wide delta country (Case); but all authorities agree that the former is much the older of the two. The Lower Coal Measures fauna is far more similar to the Per- mian fauna of Bohemia, which according to Fritsch’s classifica- tion comprises a similar series of 13 families, 26 genera and 63 species, of branchiosaurs, microsaurs and temnospondyls. But no genera are common to the two countries and many of the ‘*families’’ (as listed) are peculiar to one or the other. The families peculiar to America are the Cocyctinide, Peleontide, Tutidanide, Ptyoniide, Molgophiide, Sauropleuride, Amphiba- mide, Ichthyeanthide, Stegopide, Macrepetide, while those peculiar to Europe are the Apateonide, Limnerpetide, Micro- brachide, Dolichosoma, Ophiderpeton, Melosauride and Arche- gosauride. The families common to both continents are Nyraniide,? Cricotidee (Diplovertebride), Anthracosauride, Professor Case has directed attention* to the marked resem- blance of two of the genera (Diplovertebron, Macromerion) from the lowest Bohemian horizon (Nyran) to Cricotus of the Upper Coal Measures of North America, as furnishing ewdence that the Bohemian deposits are of Upper Carboniferous age. Subsequent research may well show on the one hand that some of the American ‘‘families’’ are more closely related to European groups than is now recognized and on the other hand that some of the ‘‘families’’ classed as common to both continents are artificial or ill defined (Hylonomide?, Nyraniide?); yet even ` with our present imperfect knowledge it appears that the Lower Coal Measures fauna of America and the ‘‘Permian’’ fauna of Europe represent nearly identical life conditions and similar 2 Brachyderpeton of the English Coal Measures, as shown by Watson, appears to be related to Diplocaulus. 3 The presence of this family in America is doubtful, and Dr. Moodie’s reasons for assigning the genera Ichthyerpeton and Cercariomorphus to this family are not stated and difficult to infer. 4 Science, Vol. 42 (Dee. 3, 1915), pp. 797-798. 314 THE AMERICAN NATURALIST [ Vou. LI adaptations on the part of two divergent associations derived from some older and common source, possibly of Mississippian age and of wide distribution; and it further appears probable that the American Lower Coal Measures fauna is somewhat the older of the two. The chapter on the morphology of the Coal Measures Am- phibia contains a careful description of the characters of the skull and other parts of the skeleton, but the author is extremely chary of generalizations. He might have mentioned, for in- stance, the interesting fact that the skull-pattern of these am- phibians is a shifting mosaic, one in which several of the dermal elements have different contacts and different positions in the various families. In some microsaurs, for example, the post- orbital grows backward and secures a broad contact with the tabular; in others it retains its primitive position. The jugal and lacrymal also differ widely in their form and contacts. The nasals and adjacent elements are small and much crowded in many branchiosaurs and microsaurs, long and wide in most labyrinthodonts. Certain dermal elements are present in some and absent in others, especially the intertemporal and the rare interfrontal and internasal elements. The shape of the occiput differs widely, sometimes truncate posteriorly, with the auditory notch obsolete, sometimes angulate posteriorly, retaining the primitively wide auditory notch. Very curious is the tendency f the different families of microsaurs to develop ‘‘horns’’— sharp backwardly projected apophyses in the occipital region— growing sometimes from the tabular, sometimes from the squa- mosal and sometimes from both at once. These remind one of the backwardly directed processes from the ‘‘epiotic’’ and supra- occipital in the skull of teleost fishes and perhaps they-may have served for the attachment of longitudinal ligaments or muscles in wriggling, aquatic types All the differences in skull pattern may be regarded as minor readjustments which were taking place after the more profound transformation of a generalized pro-ganoid skull into the am- phibian type, the greatest alteration including the loss of the opereular bones so as to leave the gill chamber covered only by membrane and the change of the preopereulars or cheek plates into the squamosals. The author expresses the opinion that the membrane bones may have originally been derived from scales ‘‘which later became consolidated into large bony scutes,” but on histological grounds the reviewer regards it as far more prob- No. 605] NOTES AND LITERATURE 315 able that in the ancestral fishes each membrane bone and each scale grew from clusters of cosmine tubercles underlain by tracts of vascular and stratified bony tissue, and that there never was a time when the elements of the dermo-cranium were scale-like in form (i. e., rhombic or polygonal), although the several tissues involved were histologically identical in the body scales and in the dermo-cranium. On page 85 the author uses the name ‘‘squamosal’’ for the clamant which he and most other authorities now designate as ‘‘supratemporal.”’? In the description of the hyobranchial elements of Coocy Giants (a genus doubtfully assigned to the Proteida), the reader looks in vain for a comparison with the same elements in the Permian ‘‘Urodele’’ Lysorophus as described by Williston. It may be noted, by the way, that the branchial arches in that genus are extremely primitive and almost Polypterus-like in form and arrangement, although doubtless homologous also with those of the modern Amblystoma. The author has given a very thorough study of the dermal scales and scutes of the branchiosaurs, microsaurs and tem- nospondyls. The ventral ‘‘seutelle,’? which appear to be homologous with the abdominal ribs of reptiles, are formed, the author holds, as ossifications in the connective-tissue septa or myocomata of the ventral muscles, vestiges of these having been found in modern urodeles. The highly differentiated charac- teristics of this ventral armature affords many family and generic characters; it is sometimes absent or reduced to needle- like ossicles, sometimes highly developed, forming heavy median ` V’s and wide lateral shelves (Ctenerpeton). Some of the micro- saurs had rounded, slightly imbricating fish-like body scales with concentric markings which recall the similar armature of certain Bohemian and Saxon types, such as Ricnodon and Dis- cosaurus. Vestiges of such conditions may be represented in the scales of modern excilians (as shown in the enlarged figures of cæcilian scales by the Sarasin brothers he reviewer ventures to doubt the correctness of Dr. Moodie’s reconstruction of the shoulder-girdle of branchiosaurs and miero- saurs, in the matter of the position of the seapula. Many of the specimens figured by Fritsch and by Credner seem to indicate that the concave border of the scapula was posterior in position, as it is in Eryops and in modern urodeles, and that it did not form the glenoid border as in Dr. Moodie’s reconstructions. 316 THE AMERICAN NATURALIST [ Von. LI Dr. Moodie’s history of the classification of the Amphibia ap- pears to the reviewer to be rather meager, since he simply lists the classifications of his predecessors without giving any critical discussion. It is surprising that in this chapter he did not men- tion the work of Fritsch with which he must be extremely famil- iar. Fritsch’s classification of the extinct Amphibia, although it was adapted and extended from the classification proposed by the British Association Committee in 1870, was, in the judgment of the reviewer, a distinct contribution to the subject which cer- tainly deserves notice in an historical review, especially since Fritsch erected several new families and gave definitions of all the European groups. The author’s own classification is an interesting attempt to divide the Amphibia of the Coal Measures into two major series or subclasses, the first (Euamphibia) including all those which may be related to modern types; the second comprising all the wholly extinet groups (microsaurs, aistopods and labyrintho- donts of all suborders). He derives most of the modern urodeles (Caudata) from the branchiosaurs, for which he has given con- siderable evidence; he follows Cope in provisionally deriving the modern Proteida from the Coceytinide of the Coal Measures. He regards the strange Diplocaulus, an amphibian with a head like a colonial cocked hat, as a member of the Euamphibia, prob- ably because its vertebre bear short, straight, double-headed ribs which are attached to paired lateral apophyses springing from the middle of the vertebre, after the fashion of those of branchio- saurs and Caudata and quite unlike the hour-glass centra of microsaurs, which bear long, slender ribs between the vertebre. But Watson and Williston regard Diplocaulus and Brachyder- peton as microsaurs, the last named genus showing in the ver- tebre and in the skull how the Diplocaulus type may have been derived from primitive microsaurian conditions. Indeed it may well be argued that the branchiosaurs and urodeles (Caudata) themselves, in spite of the retention of gills in the young, may have been derived from primitive microsaurs, that is that the vertebre and ribs of microsaurs are on the whole much more primitive than those of branchiosaurs and Caudata. The systematic relations and origin of the frogs and toads re- main doubtful. Dr. Moodie gives an excellent discussion of the resemblances of Pelion lyelli Wyman, from the Linton, Ohio, Coal Measures, to the modern Anura but leaves the phylogenetic problem open. Pelion is so little known that it may or not be No. 605] NOTES AND LITERATURE 317 ancestral to the Anura, and the Jurassic Anura are so entirely modernized that they do not bridge over the wide structural gap between the Paleozoic Amphibia and the modern frogs and toads. It seems to the reviewer, after repeated comparisons of the —osteology of the Anura with that of many of the temnospondyls, that some members of the latter group, in the brain-case, the dermo-cranium and even in the vertebre and limbs retain many characters which may reasonably be looked for in Paleozoic an- . cestors of the frogs and toads; and that such forms as Brachyops, Cacops and Dissorophus, although not directly ancestral, differ from the Anura chiefly in the retention of many primitive am- phibian characters. It may be that some of the short-headed Triassic temnospondyls of South Africa will furnish the linking forms; but at any rate it is interesting to note that the existing frogs and toads retain a long series of characters in the skull and skeleton which are seen in the Paleozoic temnospondyls, and that they differ from the latter in such modernized characters as the following: the wide fenestration of the occiput and palate, the resulting slenderness of the skull bones, the loss of the dermo- supraoccipitals, tabulars, ectopterygoids, pre- and post-frontals, the completion of the auditory ring, the development of extreme ’ saltatorial adaptations in the skeleton, including the modifica- tion of the vertebre from the rhachitomous into the notocentrous and epichordal types, the development of a long continuous uro- style coincident with the forward shifting of the sacrum and lengthening of the ilium. r. Moodie’s arrangement and sequence of the families of microsaurs appear to the reviewer to be highly confusing. would perhaps have been better, after beginning with the newt- like types, to pass at once to the long-bodied Urocordylide and the snake-like Molgophiide and Ptyoniidx, instead of interjecting in the middle of the series the Stegopide, which appear to the reviewer to be more nearly allied with the Temnospondyli, and the Amphibamidx, which are heavy-limbed offshoots of the prim- itive microsaurs. 7 The author’s ordinal and family definitions are extremely full, but the reader will find so many characters that are com- mon to several families and sometimes orders, that it is difficult to cull out the most striking ones. This the reviewer has at- tempted to do in the subjoined table in which he has also included the principal European families of branchiosaurs and micro- saurs. The families of microsaurs are arranged so far as pos- 318 THE AMERICAN NATURALIST [ Vou. Ll sible in the general order of their specialization, proceeding from the more primitive newt-like forms to the snake-like microsaurs or Aistopoda. ` SYNOPSIS OF THE PRINCIPAL BRANCHIOSAURS AND MICROSAURS OF AMERICA ND EUROP A. Vertebræ oe i. e., having the notochord expanded in the middle each vertebra; transverse process in dorsal region large; ne short, RRS and heavy and borne on the trans- verse processes, Sac near the middle of the vertebre. Skull broad, obtusely rounded. B. Auditory notch camer rather than lateral in arna Bra osaurid BB. Aae notch extended laterally, the squamosals ving A for- MEO hae oe E EA eee Rane oi oa Ap AA, Vertebre ris ae brig i. €., With the centra igi e yore con- ted cylinders; ribs intercentral, typically long and curved. B. Digits (when present) 4 in manus, 5 in pes; carpus and tarsus artilaginous C. Body newt-like. D. Body covered with cycloid scales or sculptured. scutes. mney longer than fore limbs. E. Taan DOW s ee Hylonomide. EE. Skull broa a . Ribs short, slightly ago nerpetide.5 F. Ribs thin, curved ...... Maroca hida: DD. Body- Pao reduced or absent; neural and hæmal spines caudal vertebræ often expanded. E. eae scutelle absent, nes moderate in length, skull without ‘‘horn Tutidanide. EE. Ventral scutelle weak or moderately i tail long. F. ‘‘Horns’’ on squamosals. Diceratosaurus. FF. ‘‘Horns’’ on tabulars. Goati EEE. Ventral armature highly developed, consisting of s, plates or stout bristles. Skull (so far as known) without ‘‘horns.’’ Ribs broad and heavy. Limbs well developed with claw-like phalanges: 604.4566 8% Sauropleuride, F. Skull very wide and obtuse; teeth heter- OOONG. cs5 chaise Saurerpeton. 5 Permian of Europe. ê Europe. No. 605] NOTES AND LITERATURE 319 FF. Skull moderately oo teeth O ropleu FFF. Skull very large; teeth a cakes caus apex and anterior cutting edge. Le ie a FFFF, Skull unknown; abdominal ribs very heavy with poa shelf- tika pera rE spines of vertebræ pectinate as in Urocor- dylus, Tc ciekehis and Ptyo gion Ee CC. Body abn like (serpentine). Limbs much reduced or ab- Re ea CREST eS Ta ee ( Aistopoda). D. tect armature weak or absent. E. Bi ee heavy and broad; neural tg hemal short or absent ....Molgophiid EE. mis s delicate, single- Seated neural ae wide low spines. Skull n geet pointe sa ichoso EEE. ae well developed; neural set nat. ‘solace audals expan soi scapes! skull lanceolate isi long, slender . .Ptyoniide. DD, Ventral armature consistin i ht he oat-shaped scu- telle; ribs forked, two-headed; neural enya bo low ae lower tra Hata process expanded i wide plate in the anterior half of the vertebra. Stati thee and hunker: than in Dolichosoma Ophiderpeton.? CCC. Body stout; oe well oo (but still with pyes ginous carpus and us). Tail short, hea rge. V miba ong. D. No ‘‘horns’’ on squamosal, skin covered with rounded or . hexagonal tuberculated seales. Ventral scutelle pres- Ont lesa) ae oes Chae shes cae i i DD. Squamosals produced into ‘‘horns.’’ Ventral scutellæ ap- parently similar to those of Abaas Eoserpeton tenuicorne.8 BB. Digits of manus unknown; pes with well ossified tarsus. (Limb aper Erypos). Ventral scutellæ delicate. Centra amphice- lou n broad and heavy ........... Ichthycanthide. AAA. Vertebral centra discoidal. B. TATE short, thick and ag apa Body covered ds small cycloid scales ........... ‘í Nyranide’’( D 2 C. hEn MOWGIIEG ie igs ee ee a Ichthyerpeton bradleyi. Cercariomorphus. urope. 8 Placed by Dr. Moodie in the Urocordylide, but possibly related to Am- phibamus. 9 The skull of Nyrania as figured by Fritsch (Vol. II, p. 34) appears to ` the reviewer to relate this genius with the ee but it is placed provisionally by Dr. Moodie in the Microsau 320 THE AMERICAN NATURALIST [Von. LI CC. Body Proteus BEE np eh. SOR i eee. Ichthyerpeton squamosum. . Vertebræ aac aioe deeply amphiceelous. . ae Vertebrae nown. Skull of pretty sari ae ‘with separate eee ee eis rymal and nasals large, orbits central rather han anterior; squamosals produced into m divaricate f O EE a NN S e ciple les Stegopide (Stegons). 10 10 May be remotely related to the Temnospondyli (W. K. G.) Wo. K. Grecory AMERICAN MUSEUM OF NATURAL HISTORY THE AMERICAN NATURALIST Vout. LI. . June, 1917 No. 606 BIOLOGICAL ENIGMAS AND THE THEORY OF ENZYME ACTION DR. LEONARD THOMPSON TROLAND HARVARD UNIVERSITY I Durine the past twenty years the sciences which deal with inorganic physical phenomena have made astound- ing progress in the logical synthesis of their facts and theories.1 The beginnings of this synthetic tendency lie, of course, in the middle part of the last century, in the work of such men as Faraday, Maxwell, and Mendelejeff. The discovery of radio-activity by Becquerel in 1896, and the demonstration by Thomson of the electron, in the fol- lowing year, let loose the pent-up forces of an intellectual avalanche which swept scientific conservatism quite off its feet, and seems to be carrying our thought with thrilling rapidity towards a goal which metaphysical philosophers have for ages regarded with wistful longing. This goal is the comprehension of the physical universe in terms of a few simple conceptions. The lines of demarcation which once were so rigidly drawn between the departments of physical science are disappearing before our eyes. The discovery of radio- activity, instead of adding a new science to the list, has 1D, F. Comstock and the present writer have attempted to give an ele- mentary, but comprehensive presentation of the modern theory of matter in their book, ‘‘The Nature of Matter and Electricity,’’ 1917. 321 322 THE AMERICAN NATURALIST [Vou. LI brought us very close to a cancellation of all of the names except one; the demonstration of the existence of par- ticles of negative electricity smaller than any known atom, instead of further complicating the facts of chem- istry, has introduced a hundred simplifications. Mechan- ics, chemistry, optics, and the sciences of heat, electricity and magnetism are rapidly fusing into a single logical system, the ultimate terms of which are minute particles of positive and negative electricity, the ultimate laws those of electro-dynamics, and the ultimate problems those of the structures formed by these particles in space and of the changes which these structures undergo in time. This startling progress in physics during the last two decades has not been the product of unadulterate em- pirical research. On the contrary, it has been made pos- sible only by acts of daring speculation, which to certain more orthodox scientists of an earlier period might have seemed inexcusably foolhardy. However, their justifica- tion has often come so quickly and in such unequivocal terms, that methodological critics have been obliged to remain modestly silent. To indulge in a definite and de- tailed account of the structure and behavior of single atoms of hydrogen—particles far beyond the visual range of even the ultra-microscope—may seem no more a scien- tific undertaking than the fabrication of a fairy-tale; and yet when from this account there emerges by inevitable logic a mathematical formula corresponding exactly with the constitution of the complex spectrum of hydrogen, our minds are opened to the possibility that the specula- tion is pointing the way to a fundamental truth.2 This impression becomes especially forcible when we consider that the constitution of this same hydrogen spectrum had for twenty-eight years defied the intellects of the best scientists, and by some had been regarded as incapable of explanation upon any simple hypothesis. It is a fact of fundamental logical significance that the 2 The reference is to the theory of N. Bohr, published in the Philosoph- ical Magazine, 1913, 26; 1, 476 and 857. No. 606] BIOLOGICAL ENIGMAS 323 progress represented by the modern electro-molecular conception of the physical universe has been achieved by the utilization of a few general conceptions, such as those of the electron and electrical action at a distance, These conceptions, although general, i. e., universally applicable, are nevertheless extremely definite. They are also as tangible, or concrete, as it is possible to make them. It is nearly as characteristic of the modern theory of matter to eliminate abstractions as it is for it to gather up scat- tered facts and theories to unite them into an integral system. Although elements of abstraction still remain, they are reduced to a minimum by the increasing tend- ency to demand not only an algebraic symbol, but a visual picture of the processes of nature. e H It is perhaps not surprising that the astonishing prog- ress of general physics during recent times should thus far have failed to exert any very notable influence upon the science of biology. From the point of view of the physicist, biological problems must be regarded as ques- tions of special material structure, usually of a very in- tricate character, and involving the arrangement and his- tory of units of matter for the most part larger than those upon which his attention is immediately concentrated. The program of modern physics is to build up the theory of all material structures by means of geometry and the dynamics of electrical particles. The first problem, logically, is that of the constitution of the atom, and as the solution of this problem is still unfinished, too much should not be expected of our knowledge of the config- uration of particles and forces in higher aggregates of matter. However, a critic who sees current events in the light of the history of science can hardly escape a twinge of disappointment at the recrud in biological theory, at the present time, of the doctrine of vitalism. The pres- ent, of all periods in the history of thought, is an hour of triumph of the monistic theory of nature, and yet now, 324 THE AMERICAN NATURALIST [Vou. LI more frequently than during the nineteenth century, men eminent in biology seem to quail before the complexity and delicacy of the life process, and, while uttering mechanistic truths about life, to offer them as sacrifices to a spirit of vagueness and discouragement.* It is my belief that this rejuvenation of mysticism and Aristotelian teleology is due not so much to a natural ad- miration on the part of biologists for obscure ways of thinking, as to their neglect of modern physics and of the methods of thought pursued in that science. It is the purpose of this paper, which is intentionally polemical in manner, to rebuke this tendency by commending to the attention of biologists a general speculation concerning the life process, which—although incapable of immediate verification in all of its aspects—does answer the most perplexing questions raised by vitalism, and at the same time forms a perfectly distinct bond between biological theory and the modern theory of matter. It is not improbable that the future will look back upon contemporary theoretical biology as a reactionary phase in the history of the science. The great synthetic energy of the Darwinian theory has been spent, has accomplished its magnificent results, but has left many tattered ends, by means of which a few of its enemies are attempting to tear down the entire structure once more. Even the re- markable discoveries which are classed under the name Mendelism are sometimes turned against the mechanistic conception of evolution. These discoveries, although patently of fundamental importance for the theory of life- processes, have as yet provided us with no new synthetic instruments of thought, but instead have generated an amazing and ever-growing list of abstract concepts. How- ever, these concepts do furnish us with a means for the analysis of species in terms of their genetic determination and the recent studies of Morgant and Goldschmidt® in 8 Consider, for example, the contents of Haldane’s recent address on 0-632, ‘The New Physiology,’’ Science (1916), N. S., 44, 62 « Morgan, T. H., and others, ‘‘ The Sachantans of Mendelian Heredity,’’ 1915, 5 See Goldschmidt, R., ‘‘Genetie Factors and Enzyme Reaction,’’ Science No. 606] BIOLOGICAL ENIGMAS 325 this field are pointing the way to synthetic considerations of far-reaching significance. That biologists recognize the need of new light in the theory of heredity and of evolution, is clearly shown by the following quotations, from Bateson’s Silliman lec- tures: In Z ses OF the general attention devoted to the uae of varia- knowledge of the chemistry and physics of living things which at pres- ent is quite beyond our reach. It is however becoming probable that if more knowledge of the chemical and physical structure of organisms is to be attained, the clue will be found through genetics, and thus that even in the uncodrdinated accumulation of facts of variation we are pro- viding the means of analysis applicable not only to them, but to the problems of normality also. Again: Somewhat as the philosophers of the seventeenth and eighteenth cen- turies were awaiting both a chemical and a mechanical discovery which should serve as a key to the problems of unorganized matter, so have . biologists been awaiting two several clues. In Mendelian analysis we have now, it is true, something comparable with the clue of chemistry, but there is still little prospect of penetrating the obscurity which en- velopes the mechanical aspect of our A ye morro Again: When with the thoughts suggested in the last chapter we contemplate the problem of evolution at large, the hope at the present time of con- structing even a mental picture of that process grows weak almost to the point of vanishing. We are left wondering that so lately men in general, whether scientifie or lay, were so easily satisfied. Our satis- faction, as we now see, was chiefly founded on ignorance.’ It will be perceived that the demand made by Bateson in these passages is not for new biological facts, but for physico-chemical conceptions in terms of which a chaos of biological facts, already at hand, can be explained, or systematized. Moreover, the emphasis is laid entirely upon the inability of the mind to conceive an explanation, ee 43, 98-100, Also ‘‘Experimental Intersexuality and the Sex Prob- >”? AMERICAN NaTuRALIST (1916), 50, 705-719. ga ony W., ‘Problems of Geneties’’ (1913), 31, 32, and 97. 326 THE AMERICAN NATURALIST — [Vor. LI or a synthesis of these facts, rather than upon the neces- sity of detailed proof of some explanation which has al- ready been offered. The contents of genetics would verify the proper conceptions if the human mind were only capable of suggesting them. In another place,” Bateson says, with reference to the mechanism of cell division: It is, I fear, a problem rather for the physicist than for the biologist. The sentiment may not be a popular one to utter before an assembly of biologists, but looking at the truth impersonally, I suspect that when at length minds of first rate analytical power are attracted to biological problems, some advance will be made of the kind which we are awaiting. As a matter of fact, in the school of the physical chemists there has been in preparation, since the days of Thomas Graham, a system of knowledge which, even in its present unfinished form, has a most important and direct bearing upon mooted biological problems. This is the science of the colloidal state. The difficult abstrac- tions and elaborate classificatory scheme, in terms of which the theory is now stated, will tend to be cleared up as our study of colloids comes definitely under the do- minion of the general electro-molecular theory of matter. Intimate contact with the latter has already been estab- lished, indeed, through recent remarkable contributions by Langmuir,’ dealing with the atomic constitution of solids and liquids. It is to colloidal chemistry that we must look for answers to the large majority of the fun- damental problems of vital activity. These answers will be slow in appearing, however, if we refuse to look. _ In fairness, it must of course be admitted that many biologists are keenly alive to the importance of the theory of matter, and especially of the theory of colloids, for the advancement of their science. However, possibly because the majority of these men are specialists in biochemistry, there seems to be a lack of coherent applications of mod- 7 Loo. cit., 41. 8 Langmuir, I., ‘‘The Constitution and Fundamental Properties of Solids and Liquids,’’ Journal of the American Chemical Society (1916), 38, 2221- 2295; and other forthcoming papers in the same journal and in the Phys- ical Review. No. 606] BIOLOGICAL ENIGMAS 327 ern physico-chemical ideas to the problems of evolution and heredity, which make up the heart of the biological mystery. It has for some years been my conviction that the con- ception of enzyme action, or of specific catalysis, provides a definite, general solution for all of the fundamental bio- logical enigmas: the mysteries of the origin of living matter, of the source of variations, of the mechanism of heredity and ontogeny, and of general organic regula- tion.® In this conception I believe we can find a single, synthetic answer to many, if not all, of the broad, out- standing problems of theoretical biology. Itis an answer, moreover, which links these great biological phenomena directly with molecular physics, and perfects the unity not alone of biology, but of the whole system of physical science, by suggesting that what we call life is funda- mentally a product of catalytic laws acting in colloidal systems of matter throughout the long periods of geologic time. This view implies no absurd attempt to reduce every element of vital activity to enzyme action, but it does involve a reference of all such activity to some en- zyme action, however distantly removed from present activity in time or space, as a necessary first cause. Ca- talysis is essentially a determinative relationship, and the enzyme theory of life, as a general biological hypoth- _ esis, would claim that all intra-vital or ‘‘hereditary’’ de- termination is, in the last analysis, catalytic. The conception of enzyme action is, of course, one with which all biologists, including students of genetics, are extremely familiar.° Probably there is no student of - morphogenesis who would not consider it absurd to deny that enzymes play a very important rôle in individual development. In a number of cases such participation has been clearly demonstrated by experiment, and the suggestion that the germ-cell contains ‘‘determiners’’ for 9See my two papers: ‘‘The Chemical Origin and Regulation of Life,’’ Monist (1914), 22, 92-134; and ‘‘The Enzyme Theory of Life,’’ Cleveland Medical Journal (1916), 15, 377 ff. 10 On enzyme action in general, see Bayliss, W., ‘‘The Nature of Enzyme _ Action,’’ 1914. 328 THE AMERICAN NATURALIST [Von LI the production of enzymes, which; in turn, regulate cer- tain aspects of the development, is a common one.! Several Mendelians have even hinted that the ‘‘ unit charac- ters” themselves are enzymes,'* but so far as I am aware, no worker in genetics, with the exception of Goldschmidt, has regarded this conception as an important one. In- deed, in the face of the nearly self-evident, they have turned away to vitalism and despair. Consider, for example, the Sii quotation from Bateson. We must not lose sight of the fact that though the factors: operate by the production of enzymes, of bodies on which these enzymes can act, and of intermediary substances necessary to complete the enzyme action, yet these bodies themselves can scarcely be genetic factors, but consequences of their existence. What are the factors themselves? ence do they come? How do they become integral parts of the organism? Whence, for example, came the power which is present in a White Leghorn of destroying—probably reducing—-the pigment in its feathers ?14 It is my contention in this and previous papers that statements of this sort can hardly represent anything less than intellectual blindness. On the supposition that the actual Mendelian factors are enzymes, nearly all of these general difficulties instantly vanish, and I am not acquainted with any evidence which is inconsistent with this supposition. An Up to very recent times, although a great number of hypotheses to explain catalysis were in existence,’* no 11 See, for example, the following: Loeb, J., and Chamberlain, M. M., ‘‘ An Attempt at a Be arr bice git Explanation of Certain Groups of Fluctuat- ing Variations,’’ Journal of Painia Zoology (1915), 19, 559-568. Moore, A. R., ‘‘On e delian Dominance,’’ Archiv für Entwicklungs- em (1912), 34, 168-175. Riddle, O., ‘‘Our Kno f Melanin r Formation and its Bearing on the Mendelian Description of Hered- yay ? pesca Fyllelis (1908); 16, 316 ff. 12 See Bateson, ‘‘ Mendel’s Principles of Heredity,’’ 1909, 268. 13 The ests publication of experimental results of great importance in Ep connection is promised by Goldschmidt. See above references. 14 Bateson, ‘‘ Problems of Geneties,’’ 86. 15 A comprehensive review of these theories and of the facts of catalysis and fermentation is given by Mellor, J. W., ‘‘Chemical Statics and Dy- namies,’’ 1914, 245-383. j No. 606] BIOLOGICAL ENIGMAS 329 completely satisfactory general theory of the process could be formulated. In this state of affairs, the use of the conception as a general explanatory agent in biology, could not be said to establish an unequivocal bond be- tween biological regulation and the theory of matter. At the present day, however, it is possible to frame a hy- pothesis to account for catalytic action, which has general applicability and at the same time rests directly upon the ideas of modern molecular physics. Ostwald defines a catalytic agent as ‘‘a substance which changes the velocity of a reaction without itself being changed by the process.’° In the older terminology of the pioneer, Berzelius, it is ‘‘a substance which, merely by its presence and not through its affinity, has the power to render active affinities which are latent at ordinary temperatures.’”7 According to Ostwald, catalytic power is a universal property of matter, for he says: There is probably no kind of chemical reaction which cannot be in- fluenced catalytically, and there is no substance, element, or compound which ean not act as a catalyzer.18 This being the case, it should often occur that a substance will catalyze a reaction which generates further quan- tities of the same substance, a process known as auto- catalysis. Catalytic relationships may thus be classified into the autocatalytic and the heterocatalytic. I shall at- tempt to show that the former may be the more funda- mental of the two relationships, and that reasons can be adduced for regarding autocatalytie power as a necessary property of every complex form of matter. Perhaps the simplest illustration of a catalytic effect of any sort is that of the production of crystallization in a supersaturated solution of some substance by the intro- duction of a small erystal of the same substance. This of course has the form of an autocatalytic process. Al- though effects of this kind are included in Ostwald’s classification of varieties of catalysis,!® up to recent times 16 Mellor, loc, cit., 250. 17 Ibid., 246. 18 Ibid., 254. 19 Ibid., 255. 330 THE AMERICAN NATURALIST [Vou. LI it might have been possible to raise a legitimate objection to the illustration on the ground that the induced change is not a chemical one. However, this objection is defi- nitely disposed of by the recent work of the Braggs, and others, on the constitution of crystals,2° which has shown that the unit of structure in solid bodies is usually the single atom, and not the molecule, since in crystals there is, as a rule, no exclusive arrangement of the atoms into molecular groups. The spacing of the atoms is such as to make it clear, moreover, that the forces which hold the total crystal system together are identical with those which we regard as underlying chemical affinity. In other words, in the crystal there is either no distinction between inter-atomie and inter-molecular forces (i. e., between ‘ chemical affinity and cohesion), or else the entire crystal must be considered to be a huge polymeric molecule.2! It is therefore perfectly legitimate to treat the process of crystallization as a chemical change, and to regard the initiation of this process under the conditions above de- seribed, as an example of autocatalysis, which may well be typical. Although the results of crystal analysis indicate that no distinct molecules are to be found in the solid state, this is not true of the dissolved, or of the gaseous state. Moreover, on account of the fact that their component particles are held in place by forces of electrical attrac- tion and repulsion, all molecules must possess their own fields of electrical force, and the field of any molecule must have a spatial form which is characteristic of that molecule. These field patterns will thus be different in the molecules of substances which differ chemically, and will be similar in molecules of the same or of an allied chemical substance.?? The forces of cohesion in a crystal may be thought of as resulting from the fusion of a large number of these molecular fields into a continuous mosaic, and in such manner that their several axes are parallel. 20 See Bragg, W. H., and W. L., ‘‘ X-Rays and Crystal Structure,’’ 1915, 21 Of. Langmuir, loc. cit., 2221-2222. 22 See Comstock and Troland, loc. cit., 86-89. No. 606] BIOLOGICAL ENIGMAS 331 However, such fusion can not fail to have an influ- ence upon both the form and the strength of the fields in question, since it involves a redistribution of the atomic forces. This will take the form of an opening out, or ex- pansion, which will necessarily reduce the coherence of the group of atoms originally forming the individual molecule. The degree of this ‘‘opening’’ of the field which occurs in crystallization must vary with the nature of the molecule, and is probably smaller for organic sub- stances than it is for the majority of inorganic com- pounds. = The mechanism of the autocatalytic process of crys- tallization may be described somewhat as follows: In a solution, or a gas, the molecules of the dissolved substance move about at random among the molecules of the solvent, and the orientation of the axes of their fields is entirely haphazard. However, as soon as a crystal of the solute is introduced, the field forces of the surface layer of atoms attract the dissolved molecules and at the same time tend to turn them on their axes so that, as they condense, they will fall into the pattern of the ‘‘space lat- tice” upon the plan of which the crystal is built.** As this is the most stable position which they can assume, they will tend to remain there and form a new surface layer of the crystal, to act in turn upon further molecules in the solution, until all of the surplus dissolved substance has been deposited. The primary force bringing the molecules to the erystal face is of course not the surface field of attraction— surface tension field—but their temperature motion—or osmotic pressure. A similar force, of lower magnitude in the case of a supersaturated solution, is constantly disengaging molecules from the crystal and throwing them back into the solution. The action apparently ceases when the number of molecules deposited upon the crystal surface in unit time becomes reduced—owing to decreasing concentration—to an equality with the num- ber leaving in the same interval. 23 Cf. ibid., 113. 332 THE AMERICAN NATURALIST [Von. LI The essential feature of the above described mechanism for the autocatalytie production of polymeric molecules may be illustrated to the eye by means of a model con- sisting of a board with a large number of small compass needles mounted upon it. If these needles are freed from the action of the terrestrial magnetic field and are then shaken into a random orientation, they may remain in this condition indefinitely. However, if a small number of adjacent needles be turned by some outside force so as to acquire a common direction, their combined magnetic fields will cause other neighboring needles to swing into line, so that the action must spread to all of the needles on the board. The field of an ideal compass needle has a simple bipolar pattern, and a symmetrical distribution of forces. In the cases of specific atoms and molecules, how- ever, this is probably seldom true. Nevertheless, the general principles involved in their dynamic interaction would remain the same as those for the case of the com- pass needles. It is clear that the explanation of autocatalysis above given accounts immediately only for the synthesis of poly- meric molecules from individual units which are all alike. As a rule, chemical changes involve the interaction of dif- ferent units, and it can easily be seen that the same gen- eral mechanism will apply to the catalysis of reactions of this sort as to that of simple crytallization. The prin- ciples involved in the process have been made especially clear in the recent articles of Langmuir.** Consider first a solution containing two kinds of molecules which can be deposited upon a erystal surface consisting of an orderly arrangement of these two molecular groupings in mosaic. or lattice form. The second species of molecules may be considered, for example, to be those of the solvent, as in the case of ‘‘ water of crystallization.’ There will be cer- tain ‘‘elementary spaces’’—as Langmuir calls them— upon the surface of the crystal, which will especially at- tract and orient the water molecules, and adjacent ele- | mentary spaces which will act in the same way upon the 24 Loc. cit., 2286-2292. No. 606] BIOLOGICAL ENIGMAS 833 molecules of the solute. In this way the erystal or poly- meric molecule will be built up out of two components by the simultaneous and parallel action of two initially com- _ bined species of molecular fields. This change is cat- alyzed by the crystal, and is an autocatalytie process in- volving the synthesis of two substances. It is clear that any number of substances may be influenced in this way by a similar, but more complex initial crystal form. However, our explanation still remains somewhat spe- cial in its application, as in the majority of cases the products of catalysis do not adhere permanently to the catalytic surface. The extension of the explanation to cases of this sort is not difficult, since we have already seen that, even in the case of crystallization, the heat vibrations of the atoms are constantly throwing off molec- ular groups from the surface of the solid. As pointed out by Langmuir, the attraction between the surface and two molecular groups which have a strong affinity for each other may be less than the sum of the attractions of the surface for each of the groups, when separate.2> This is due to the ‘‘closing up’’ or contraction of the fields of force of the groups as they come together. Hence com- bined groups of this sort will be more easily detached from the surface than will the uncombined groups, which will tend to be held in place until their mates fall into the right positions. The catalytic surface thus acts like an orienting sieve which on account of its special structure forces a chaotic crowd of individuals which come into con- tact with it, to fall into a special configuration. Many machines which accomplish exactly this effect are in use in the industries. Thus far we have dealt only with the mechanism of autocatalysis. Heterocatalysis is probably to be regarded as an extension of the process of autocatalysis. It is obvious that exact similarity of the force patterns of the catalyzing and catalyzed systems is not essential. In- deed, the catalytic effect which is based upon direct simi- larity of structure between the two systems should be 25 Ibid., 2257, 2264-2266. 334 THE AMERICAN NATURALIST [Von. LI much weaker than that which accompanies certain types of structural correspondence, such as that existing be- tween a body and its mirror-image, or between a lock and a key. Special structural relations of this sort probably exist between stereochemical isomers, between acids and bases, ete. It is easily conceivable that the patterns of certain surfaces may be capable of distorting other special configurations which come under their influence, so that they fall into new equilibrium figures, without these figures being of necessity identical with those of the catalytic system. The general principles of the action, however, remain the same. Catalytic synthesis is a less common process in the laboratory than is destructive catalysis, but the laws of energy necessitate both effects, if either one is possible. Consequently the mechanism which we have described above must be an exactly reversible one, and must assist in the decomposition of molecular complexes as much as it aids in their synthesis. The deposition of the mole- cules to be decomposed, upon the catalytic surface would naturally follow the same principles as those stated for simple polymerization. In this state of deposition the . field forces of the crystal surface would inevitably have a tendency to open up the field of the deposited molecule, thus rendering it more unstable than before, in which con- dition the temperature vibrations of the system could break it up more easily than in the undeposited state. — This weakening of the internal bonds of the molecule in the field of the catalytic surface corresponds with the weakening of forces of electrical attraction by increasing the ‘‘dielectrie capacity’’ of the medium in which an elec- trical system is contained. It is the same action which permits water to dissociate neutral molecules into ions,?° and is probably responsible for the high catalytic power of water, in general. However, in detail, the process must be a ‘‘personal’”’ affair between individual water mole- cules and molecules of the dissolved substance, just as in the case of the crystal surface, since the ionizing effect 26 Cf, Comstock and Troland, loc. cit., 139-140. No. 606] BIOLOGICAL ENIGMAS 335 of water does not appear to depend merely upon the chemical instability of the solute. The increase in reaction velocity which characterizes catalysis is to be attributed to three more or less sep- arable influences exerted by the catalytic surface, (1) the local increase in the concentrations of the reacting sub- stances at the surface, (2) the impressment upon the attached molecules, of a relative orientation which is favorable to chemical union, or which in part constitutes such union, and (3) the spreading and weakening of the fields of force of the molecules, due to their interaction with the surface fields. The first factor, alone, would be of primary importance for the combination of free atoms —a relatively rare process—while the last factor, alone, would be responsible for the acceleration of simple de- compositions. Reactions between two or more molecular groups, whether synthetic or metathetic, should be influ- enced by all three factors. Strutt” -has shown that in | certain typical chemical reactions, only one out of many millions of collisions between potentially reactable mole- cules results in chemical interaction. The active collisions probably coincide with the presence in the colliding sys- tem of favorable relative orientati and states of tlie molecular fields, which in the absence of a catalyzer de- pend upon chance, but which in the presence of a catalyzer ` are encouraged by the nature of the catalytic surface. It is of course not possible in a paper of this sort to enter into the mathematics of the theory of catalysis which is outlined above.?® Catalytic influence is obviously only one among many factors which affect a chemical reaction. Catalysis is possible only when the appropriate raw ma- terials are provided, and when the energy relations of the system are such as to make the reaction thermodynam- ically conceivable. The heterocatalytic effect of a given substance may far outweigh its autocatalytic effect either 27 Strutt, R. J., ‘‘ Molecular Statisties of Some Chemical Actions,’’ Pro- ceedings of the Royal Society (1912), A, 87, 302-309. 28 Cf. Mellor, loc cit., 250-254, Also Bayliss, loc. cit., 49-71. 29 For a development of the mass action relationships involved, see Lang- muir, loc. cit., 2287 ff. 336 THE AMERICAN NATURALIST (Von. LI because the energy changes do not favor the-.latter, or be- cause in a given system the raw material for the auto- catalytic reaction is absent, while that for the hetero- catalytic reaction is present in abundance. However, the above considerations would lead us to be- lieve that all substances should show some tendency to form polymeric molecules or crystals. This appears to conflict with the classical division of substances into erystalloids and colloids, but this division, like all others, can not be expected to stand unmodified by the modern analysis. There is plenty of evidence from direct observation that many colloidal particles are simply very small crystals.” On the other hand, the molecules of polymeric substances of high molecular weight, such as starch and certain pro- teins, are probably of the same order of magnitude as small colloidal particles. From the point of view of the theory of matter, there is no fundamental difference be- tween the general plan of a starch molecule and that of a crystal of sugar, and it is highly probable that the dis- tinction between colloids and ecrystalloids rests upon purely quantitative relations, respecting the size of the polymeric structures (crystals) produced under ordinary conditions. Large crystals are formed easily by simple substances whose molecules have open fields of force or highly un- saturated attractions. Small crystals are characteristic of more complex substances, common among the com- pounds of carbon, having relatively closed fields. Large mosaics of such molecules become unwieldy and are easily disrupted by the temperature vibrations. They are also built up more slowly than are mosaics of molecules with open fields. The distinction between these two classes of molecules is of course merely quantitative; no type of molecule has a completely closed field, and on the other hand no substance is capable of forming indefinitely large crystals in a finite length of time. The atomic structure of the solid phase of a colloidal gel is probably analogous 30 See Ostwald, Wo., ‘‘A Handbook of Colloid-Chemistry,’’ English trans- lation, 1915, 56-66. No. 606] BIOLOGICAL ENIGMAS 337 to that of the mass of small crystals, compacted together, which always results from the rapid crystallization of a supersaturated solution of a substance like, e. g., sodium thiosulphate. It is evident, then, that the general theory of catalysis which has been outlined is applicable to enzyme action, which almost certainly depends upon the deposition, or adsorption of the reacting substances upon the surfaces of colloidal particles.*t Such adsorption, the molecular mechanism of which has been made very clear by Lang- muir,®* will tend to be specific, and the more specific the more complex is the structure of the units making up the mosaic of the surface. Molecules the field patterns of which fit closely into the fields of the surface will tend to displace others having a cruder correspondence. This follows from either electro-dynamies or thermodynamics, and obviously coincides with Fischer’s classical concep- tion of the lock and key relation between enzyme and sub- strate.** It will be perceived that our theory of the catalytic process is simply a refinement and extension of the clas- sical theory of ‘‘intermediate compounds,’’ which has been proven true in so many instances.** ‘‘ Adsorption compounds,” which play the principal rôle in enzyme action, do not differ dynamically from chemical com- pounds in general, since the forces causing adsorption are the same as those responsible for chemical union. Conversely, catalytic action in which the catalyst is in a molecular or unpolymerized state will not necessarily differ in its mechanism from that characteristic of en- zymes or of metallic surfaces. IV The suggestion that the fundamental life-process of growth is the expression of an autocatalytie chemical re- 31 See Bayliss, loc. cit., 104-123. .32 Loc, cit., 2267-2278. 33 See Mellor, loc. cit., 363. 34 Cf., e. g., Kendall, J., and Booge, J. E., t‘ Studies on Catalysis. I. The Addition Compounds of Esters with Grpiste Acids,’’ Journal of the Amer- ican Chemical Society (1916), 38, 1712-1736. 338 THE AMERICAN NATURALIST [Von. LI action has been made independently by a number of in- vestigators.*> It will be perceived that on the basis of the foregoing theory of autocatalysis, this suggestion becomes closely allied to the familiar and ancient com- parison of vital growth to the growth of a crystal. The customary objection to this comparison, viz., that a erys- tal grows by accretion whereas protoplasm increases by intussusception, loses its force as soon as we regard liv- ing matter as a complex mixture of substances suspended by colloidal subdivision in water, since there is no evi- dence that the individual colloidal particles do not grow by accretion. On the contrary, it is almost inconceivable that these bodies, which are the real chemical units of pro- toplasm, should grow in any other way. The growth of a system like a cell could be regarded as the resultant effect of a very large number of component growths, each governed by its specific autocatalytic mechanism. It has been shown by T. B. Robertson** that growth curves, with respect to the time, actually do coincide in general form with the curve characteristic of an autocatalytic reaction. A multitude of observations substantiate the belief that the internal determination of cell-life rests primarily with the nucleus,®? or with the chromatin substance of the cell, when no well-defined nucleus is present. Even in the highly organized cell, this substance can be seen to pos- sess a mosaic structure, and it ean be shown that for a given species this structure is sensibly constant,** so that it is necessary to suppose that a reduplication of chromatin units occurs with each cell-division. This process of re- duplication is apparently made visible to us in mitosis. 35 See, for example, esis W, ‘t Ueber die zeitlichen Epe ane der a Vorträge und Aufsätze über wicklungs- E oe herausgegeben von W, Roux ( sor » Hoeft 5, Leipzig. 36 Robertson, T. B., ‘On the Normal Rate of Growth of an Individual and its Seppia Significance, ’” Archiv fiir sbi sir nae (1908), 25, cael iy hi ep ee articles in the sa n, E. B., ‘‘ The Cell in Develbinds wiy Inheritance,’’ second 54. edition, uren l 1911), 30-31, 341 38 Cf. Boveri’s Individualitäts Hypothese and ‘‘law of proportional nuclear growth. ”? No. 606] BIOLOGICAL ENIGMAS 339 The simplest hypothesis to account for such reproduction lies in the supposition that each unit can give rise to an- other unit substantially identical with itself. The Weismannian theory of the constitution of the germ plasm,*® which is typical of the so-called ‘‘corpus- cular theories’’ of the life-process or of heredity, also de- mands the existence of vital elements, each possessing the power of reproduction ad infinitum. The general conceptions of this theory appear to find verification, first, in the facts already mentioned, and second, in the discov- eries of Mendelism. The recent work of Morgan and his collaborators,#? moreover, reveals clearly the intimate connection which exists between the corpuscular ‘‘unit characters” of Mendelian heredity and the histological units present in the chromosomes. Consequently, it would seem to be a fairly safe generalization, or at least an extremely probable hypothesis, which states that the distinctive properties of cells, tissues, and species are primarily determined by the nature of systems of colloidal particles contained in cell-nuclei and, originally, in some germ-cell nucleus. However, in spite of the seeming strength of the evi- dence, some biologists are of the opinion that such a view as this must be rejected because it paralyzes thought. Consider, for example, the following quotation from Chid.“ It is scarcely necessary to call attention to the fact that these [cor- puscular] theories do not help us in any way to solve any of the fun- damental problems of biology; they merely serve to place these prob- lems beyond the reach of scientific investigation. The hypothetical units are themselves organisms with all the essential characteristics of the organisms that we know; they possess a definite constitution, they grow at the expense of nutritive material, they reproduce their kind. In other words, the problems of development, growth, reproduction, and inheritance exist for each of them, and the assumption of their existence brings us not a step nearer the solution of any of these p ems. These theories are nothing more nor less than translations of the phenomena of life as we know them into terms of the activity of 39 Weismann, A., ‘‘The Germ Plasm,’’ English translation, 1893. 40 Loe. cit. 41 Child, C. M., ‘‘ Senescence and Rejuvenescence,’’ 1915, 11-12, 340 THE AMERICAN NATURALIST [Von. LI raultitudes of invisible hypothetical organisms, and therefore contribute nothing in the way of real advance. No valid evidence for the existence of these units exists, but if their existence were to be demonstrated we might well despair of gaining any actual knowledge of life. We have in this passage a clear statement of the essen- tiality of growth, as self-reduplication of specific sub- stance, in the life-process. Consideration of Child’s re- marks will show that.the difficulties which he raises are almost completely dissolved as soon as we postulate for the biological corpuscles the power of autocatalysis. In the light of our previous discussion, it can not be claimed that this is purely a verbal solution of the problem, as we have advanced definite reasons for believing that auto- catalytic activity is a property of all chemical substances whatsoever, given the appropriate chemical’ environment. Since the environment of the chromatin particles has been made to order by evolution, the force of Child’s criticisms would seem to be nil. Moreover, he certainly underesti- mates the importance of the facts which point to an actual corpuscular determination of vital functions. In view of this, it would appear advisable to aecept the Elementarorganismen'? as if they were clearly estab- lished facts, and proceed to consider what further light ean be thrown upon biological problems by the conception of specifie catalysis. It is well known that in many cases, at least, the nucleus controls cell activity by liberating enzymes,** and the mass activity of cells in the form of specific tissues has been satisfactorily proven to depend, so far as it is di- rectly chemical, upon the presence in these tissues’ of specific enzymes. In the field of general adult metabolism the determinative importance of catalysis would appear to be no longer a matter of debate. Analogy would lead us to believe that the same principle is of prime impor- tance in the metabolism of development. In laying emphasis upon the cardinal importance of the riicke, Se Akademie der Wissenschaften, Wien, 42 See PA, 44, (2), 381-4 43 See ‘Ment nu, G., * irs Chemistry of the ma g Rie 454 ff. Also, Loeb, J., ‘‘ The Dynamics of Living Matter,” 1906, No. 606] BIOLOGICAL ENIGMAS 341 enzyme for organic regulation we must of course recog- nize that the exact effects produced by catalysis depend, at all stages of development, upon the manner of its co- operation with other physical principles, which may in- volve the functioning of molar structures already pres- ent. However, as we retrace the course of ontogeny and of the evolution of any specific germ-cell, we should find that the number and importance of such structures de- crease, although the construction of any given tissue- form always depends upon the action of specific enzymes in conjunction with preéxisting tissue structures. It is not to be doubted for an instant that important preéstab- lished structures exist even in germ-cells, and enter into the determination of their activity. It is therefore un- fair to demand a catalytic explanation of such a complex process as karyokinesis, which shall not take into con- sideration the history or evolution of the cell. The task of elucidating the exact mechanism by means of which vital regulation is maintained, and especially of showing how, in accordance with recognized principles | of physics, a complex of specific, autocatalytic, colloidal particles in the germ-cell can engineer the construction of a vertebrate organism, is truly'so formidable that it is unkind for the vitalist arbitrarily to deny us the use of any of these recognized principles. For example, we must be permitted to suppose that a large number of variables can unite in the production of a single effect. The greater part of the modern vitalistie worry over ‘‘or- ganization” and vital ‘‘equilibrium’’** appears to depend upon a tacit assumption, either that physical laws are not reliable, or that it is impossible for a number of variables — to control simultaneously a single process. Both of these assumptions are self-evidently counter to the most funda- | mental presuppositions of physical analysis. - Although the fundamental life-property of the chro- matin units is that of autocatalysis, it is necessary and legitimate to suppose that the majority of them sustain specific heterocatalytie relationships to reactions oc- 44 Consider, for example, Haldane, loc. cit. 342 THE AMERICAN NATURALIST [Vou. LI curring in living matter. This is because nuclear ma- terial makes up a relatively small percentage of proto- plasm, and because the reactions governed by enzymes are ordinarily heterocatalytic. It is a remarkable fact that the chemistry of the cell- nucleus has reached a stage of advancement superior to that attained by the chemistry of the cytoplasm. It ap- pears that the essential constituent of chromatin is a sub- stance called nuclein, which is composed of a basic, pro- tein factor and nucleic acid. The facts indicate that the acid factor is the permanent and essential component of the nucleus, and organic chemical analysis seems to prove that only one kind of nucleic acid exists in animal tissues, although a different variety is to be found in the cells of plants.*® If, as now seems probable, the genetic enzymes must be identified with the nucleic acids, we shall be forced to suppose that these substances, although homo- geneous—in animal or plant—from the point of view of ordinary chemical analysis, are actually built up in the living chromatin, into highly differentiated colloidal, and colloidal-molar, structures. The apparent homogeneity results from the fact that ordinary chemical analysis pro- vides us only with the statistics of the fundamental radicles which are involved. To some minds, the idea that a portion of matter as small as a germ-cell can contain sufficient catalytic sub- stance to control the destinies of a complex organism, seems hardly plausible. However, considering the slow- ness of such processes as growth, it is clear that the quan- tity of catalyzer required will usually be smaller than that used in laboratory experiments; and it is a truism in chemistry that radical alterations of reaction velocities ean be caused by the presence of almost infinitesimal amounts of catalytic material. From the nature of the process, it is evident that only a few molecules of sub- stance will be required to furnish the basis for an auto- catalytic reaction which may eventually result in the pro- 45 See Jones, W., ‘‘ The Nucleic Acids,’’ 1914, 46 See Mellor, loc. cit., 248-249. No. 606] BIOLOGICAL ENIGMAS 343 duction of any desired amount of this substance; and a simple calculation shows that the chromatin of the human zygote has sufficient volume to contain about one quad- rillion (101°) molecules the size of that of oxygen.*7 In order that the enzymes of the germ-cell should be able to determine the form of the mature organism, they must have the power to govern (1) the physical and chem- ical properties of specific tissue material, (2) the posi- tion of specific tissues, (3) the size of these tissues and (4) their form. Since the physical properties of any piece of matter depend upon its chemical constitution, and since any chemical change can be regulated by catalysis, the mere presence of a specific catalyzer in a favorable mixture is sufficient to determine the produc- tion of matter of any possible variety, in any possible amount. It is always necessary to assume that the his- tory of an organic system is such as to have provided it with the raw materials necessary to its activities. If this is not the case, the system naturally perishes of ‘‘star- vation.” The most primitive form of cell-division involves noth- ing more than reduplication, and this is the law of mul- tiplication of the germ-plasm. Driescht8 argues that to explain the reproduction of a nuclear ‘‘machine’’ which determines development, we must postulate another ma- chine to carry out the operation, and so on ad infinitum. The nature of the autocatalytic process, however, shows that this conclusion is in error, since pure autocatalysis would tend to bring about an exact qualitative reproduc- tion of any given plane or linear mosaic of specific units. In a nutritive medium such a mesaic would tend to grow in all of its parts by the deposition of similar substance. Primitive nuclear division (as, e. g., in the Protista) may depend solely upon the physical instability of colloidal particles greater than a certain size, but it can hardly be 47 This calculation is based on the following assumptions: (1) that the diameter of the germ-cell nucleus is .05 mm., and (2) that the molecules fill only one-sixth of the total volume of the nucleus. 48 Driesch, H., ‘‘The Science and Philosophy of the Organism,’’ 1908, 2, 341. 344 THE AMERICAN NATURALIST [Vou. LI doubted that the complex mechanism of mitosis rests upon definite structural machinery, established by long periods of evolution. In order to account for the differentiation of cell-nature which occurs in ontology, Weismann was led to assume a thoroughgoing differential segregation of the biophores of the original germ-cell in the course of embryological development; in other words, he supposed that in this process the rule of nuclear division is differentiation and not reduplication. This assumption, although undoubt- edly a partial truth, is neither necessary nor in harmony with general biological probabilities, in the form in which it was made by Weismann.*® Consequently the difficulties into which it has led his general theory, can be regarded as without important bearing upon the acceptability of corpuscular hypotheses at large. Since reduplicating division is the established rule among unicellular organisms—which must have had a long evolutionary history—we should expect this rule to be conserved as far as possible in multicellular evolution. According to the general law of recapitulation, this should be especially true for the primary stages of ontogeny, for which Driesch’s principle of the ‘‘equipotential system”’ appears often to hold. The blastula may well be simply an undifferentiated mass of germ-cells, analogous to a homogeneous colony of unicellular forms. Rudimentary differentiation may be brought about and determined by specific enzyme constitution, without differential par- titionment of enzymes in segmentation, since the forces acting upon any cell must depend upon its position in the - mass, and the activation or inhibition of a given enzyme may be conditioned by the presence of definite stimuli in definite intensities. This being the case, any cell could assume germinal characters if isolated from the total mass.°° 49 Weismann’s theory, it must be recognized, assumes ‘‘doubling divi- sion’’ for the early “ae of segmentation, a law which continues to hold for the ‘‘ germ-tracks.’ 50 Of. Hertwig, O., ‘Evolution or Epigenesis,’’ English translation, 1896. No. 606] BIOLOGICAL ENIGMAS 345 However, the facts of ‘‘crossing over’’ observed in re- cent studies on the relation between Mendelian charac- ters and chromosome constitution®! show that the latter is not inviolate, even in purely germinal segmentation. Since the power possessed by cells to assume germinal character, even to the limited degree of. being able to regenerate a single organ or tissue, seems to vary in in- verse proportion to the degree of specialization of the cells, it is reasonable to suppose that Weismann’s prin- ciple of differentiating division actually does operate in the higher stages of development. However, at no stage is it the only mechanism of differentiation, and it cer- tainly is not the primitive means. It is possible that cancer represents a return of tissue cells to a germinal or semi-germinal stage, due to the failure of a ‘‘stimulus of differentiation’’ to remain ef- fective. The control by the genetic enzymes of the position, size and form of specific tissues must involve, first, a quanti- tative regulation of the process of differentiation, which can be effected by the establishment of definite relations between the chemical constants of the catalytic reactions and the conditions under which the course of development necessarily places them; and, second, control of the planes of segmentation of the cells. To attempt a specification of the exact process by which this latter factor can be governed by the chemical constitution of the cell-nucleus, lies beyond the scope of the present paper, but it should be pointed out that in the last analysis chemical constitu- tion means nothing but a definite spatial arrangement of electrical forces, so that there is nothing paradoxical in the determination of ‘‘pure form’’ by chemical agents. As is evident in the quotations made above from Bate- son, the dominant problem in the modern discussion of evolution is that of the origin of variations. ` It is the failure of Neo-Darwinians to explain the appearance of variations, and especially of new unit characters, which has led such writers as Driesch, Korschinsky and Wolff 51 See Morgan, loc. cit., 346 THE AMERICAN NATURALIST [Vón LI to speak of the ‘‘episode of Darwinism’’ and of Das Sterbelager des Darwinismus.®°2 The enzyme theory of vital determination brings new life to the doctrine of evo- lution by accidental variation and natural selection, first, by showing that all fundamental variations should be dis- continuous, or heterogenetic, as demanded by the muta- tion theory of De Vries, and second by revealing the exact mechanism of the production of these variations. The discontinuity follows from the existence of qualitative gaps between all specific chemical substances, such as those making up the system of genetic enzymes. The mechanism of production of variations is simply that of the initial production of any new chemical individual, i. e., the fortuitous encounter of the appropriate molecules with the right relative orientations and at the correct speeds (vide supra). The ‘‘chance’’ nature of variation thus is made to depend upon that ‘‘molecular chaos”’ which is so very familiar to all physicists, but the impli- cations of which for biology have thus far been largely neglected. A moment’s thought will show that, on the basis of the enzyme theory, variation should be additive, since an autocatalytic individual, once established, will tend auto- matically to maintain itself. The complete elimination of such individuals will oceur only through the destruction of the entire germinal mosaic of which they form a part, an effect accomplished by natural selection unless the new enzyme is in harmony with functions which preserve the organism. It is very important to bear in mind that the catalytic complex which is supposed to underlie organic development and regulation has been determined in its nature by excessively exhaustive practical tests and, as a complex, by nothing else. It is therefore not surprising that the practical delicacy of the regulation which it actu- ally subserves should be very great. In other papers, I have discussed somewhat in detail the bearing of the enzyme theory upon the problem of the origin of life. On the basis provided by this theory, the 52 See Kellogg, V. L., ‘‘ Darwinism To-day,’’ 1907. No. 606] BIOLOGICAL ENIGMAS 347 origin of life can not be regarded as a catastrophic event; life depends upon an organized complex of selected cat- alytic material, and hence some life originates with each new, successful mutation. Of course, if we trace the proc- ess of the evolution of any given species back sufficiently far, we must eventually come to the first mutation, which would consist in the molecular production of an auto- catalytic particle sustaining relations with its environ- ment such as to make possible its continued growth and reproduction. I have used the name protase to stand for the ‘‘first enzyme’? of the archebiotic process, but there is no particular reason for supposing that there was only one enzyme to which this name could apply. There is considerable evidence that free autocatalytic enzymes exist in our biological universe even at the pres- ent day. Such an hypothesis would serve to account for the specific contagious diseases, such as measles, rabies, and smallpox, which have been demonstrated to possess ‘‘filterable viruses.’ The so-called Chlamydozoa prob- ably fall in this class. That the Chlamydozoa consist of free chromatin ma- terial is suggested by the late Professor Minchin, in his admirable paper on the evolution of the cell,®* with the main outlines of which the enzyme theory would entirely agree. The single cell, and so-called simple protoplasm, must be regarded as the products of a detailed process of evolution, and hence can not form the ultimate explanatory units in biology. Next to the free autocatalytic particle, the simplest typical life-structure would consist of a single particle of this sort surrounded by an envelope of semi- liquid and chemically homogeneous substance with which it sustains a heterocatalytic relationship. “The most primitive substance of this kind might be called eoplasm, to distinguish it from complex protoplasm, and the phys- ical system made up of protase and eoplasm would repre- sent a living cell in its most reduced form. Minchin says, in the article referred to :** 53 Minchin, E. A., ‘The Evolution of the Cell,’? AMERICAN NATURALIST (1916), 50, 5-39, 106-119. 54 Loc. cit., 35-36. 348 THE AMERICAN NATURALIST [Von. LI The biochemist renders inestimable services in elucidating the chem- ical mechanisms of living organisms but the problem of individuality and specific behavior, as manifested by living things, is beyond the scope of his science, at least at present. Such problems are essentially of distinctive vital nature and their treatment can not be brought satis- factorily into relation at the present time with the physico-chemical in- teractions of the substances composing the living body. S in a certain historical epoch, and that in the future the chemist will be able to correlate the individuality of living beings with their chemico- physical properties, and so explain to us how living beings first came into existence; how, that is to say, a combination of chemical substances, each owing its characteristic properties to a definite molecular composi- tion, can produce a living individual in which specific properties are associated with matter in a state of flux. It is my contention that the enzyme theory of life satis- factorily meets these general requirements. y To arrive at a proper estimate of the importance of a general theory such as the one discussed in the present paper, necessitates considerable reflection. The path of scientific progress is beset by the pitfalls of conservative empiricism, on the one hand, and by those of radical spec- ulation, on the other. ‘To the radicals the enzyme theory presents an aspect of a priori self-evidence; to the con- servatives it seems to be a vague generalization with no particular or specific facts to support it, approximately on the same plane as the statement that ‘‘life is motion,” which Driesch says is about as useful as the proposition that ‘‘Kant was a vertebrate.’’ Regarding the general enzyme theory, the following opinion has been expressed to me privately by an eminent zoologist. The idea... is a perfectly familiar one. The trouble comes when we attempt to make a specific application of this idea. to a concrete problem, which is what science demands if a pure speculation is to become a valuable working hypothesis. For instance, how an auto- pratense molecule could produce the phenomena seen in the division of chromosomes in the cell is by no means clear; nor is it clear why pies molecules brought together during fertilization separate from each other at the maturation division. It is these specifie questions that must first be answered, I think, before we can make much advance, i in No. 606] BIOLOGICAL ENIGMAS 349 regard to the nature of the phenomena. ... Of course, I realize that general ideas are always important for the development of science, only I think they should be advanced with caution and all attempts to make them appear as specific explanations should be avoided. _ There are a number of principles of scientific meth- odology which have a bearing upon criticisms of this sort. In the first place such a statement as that ‘‘life is motion’’ would have value at a time when the connection between these two ideas had not been noted, or had been underemphasized to such an extent that eminent scien- tists were bemoaning the inability of the human mind to account for properties of life which the most superficial examination would show to be identical with those of motion. From my own point of view the proposition that ‘*life is determined by specific catalysis’’ appears to be somewhat of the nature of a truism, indeed so much so that even if we had no direct evidence for the existence of enzyme action, we should be forced to invent the con- ception to account for the most general properties of liv- ing systems. It seems to me that this is exactly what a considerable number of biologists actually have done, and the only important error in their thinking lies in the ap- plication to the concept of such names as ‘‘biophore,”’ ‘‘determinant,’’ ‘‘unit character,” ‘‘formative factor,” ‘*Elementarorganism,’’ ‘‘élan vital, or ‘‘entelechy.’’ This error, however, is fatal to progress, as it multiplies terminology and delays the synthesis of actual ideas which is the goal of scientific endeavor. Furthermore, it is not true that the establishment of a general principle necessitates an examination of all of the concrete details of specific systems. If this were so, none of the fundamental laws of mechanics, such as the first and second laws of energetics or Hamilton’s prin- ciple, would possess- any rigidity, since they are derived from a study of what may be called the ‘‘entrance and exit’’ properties of mechanical systems, without reference to their exact contents. It is to be admitted, of course, that we can not rest content with this kind of knowledge, - and that principles of this sort receive complete elucida- 350 THE AMERICAN NATURALIST > [Vou LI tion only when the details of all systems are made clear, but the security of the principles themselves is attected searcely at all by this analysis. I do not claim that the enzyme theory of life possesses a general basis as adequate, for example, as that of the principle of least action. I do claim, however, that this is because the latter can be stated in terms of an exact mathematical formula, whereas the enzyme theory has to be given a qualitative description. The enzyme doctrine is supported at the present time by a considerable number of specific facts of cell chemistry, but it possesses a far more substantial bulwark in the general facts of vital function. Shall we deny that these facts are adequately established, or that they are important, or that they merit explanation? Shall we reject a definite physico-chemical conception which at one stroke explains the majority of the mass relationships of living matter, on the ground that the details of some special life-processes have not yet been described in terms of this conception? Or is it preferable to preserve the inexplication of these same generalities to furnish a basis for vitalism? There are an indefinitely large number of ways in which the principle of the conservation of energy can be exem- plified in special pieces of machinery, and there are just as many ways in which the principle of specific catalysis can operate. Instead of holding the energy principle in abeyance until we have seen how the action of a special mechanical system can be explained in terms of it, we usually assume it to be true, and shortly find the action in question very easy to understand. This would seem to be the only feasible method for employing any theoretical proposition, even if it is merely a novel working hypoth- esis. The trouble which arises in the attempt to apply the enzyme theory to specific problems is a normal result of the inertia of the human imagination, which does not im- mediately outline a plan for a machine to accomplish a definite purpose, even when it is provided with all of the principles of mechanics. Surely, however, the plan can never be developed if such principles are neglected. MUTATION IN DIDINIUM NASUTUM S. O. MAST ZOOLOGICAL LABORATORY OF THE JOHNS HOPKINS UNIVERSITY Tae origin of heritable variations or mutations consti- tutes one of the most fundamental problems of biology. It has long since been recognized that evolution depends upon such variations, and they have consequently been extensively studied by a considerable number of inves- tigators, e. g., Darwin, DeVries, Batson, Kammerer, Tower, Stockard, MacDougal, Jennings, Morgan, et al. These studies have resulted in the accumulation of a mass of facts of great importance, but the nature of the origin of the variations in question is still shrouded in mystery. In a series of experiments on the effect of conjugation and encystment in Didinium, extending from April, 1910, to May, 1914, there suddenly appeared, in the latter part of July, 1912, a marked difference in the rate of fission in the progeny of a single individual. This difference ap- pears to have been permanent, as the results presented herewith indicate. And it seems to show, in opposition to the conclusions reached by a considerable number of in- vestigators, that variations in organisms reproducing asexually are at times heritable. The difference in rate of fission mentioned was discov- ered in a group of five pure lines, all of which had been carried from the beginning of the experiment. These lines all originated from the same individual, and before the mutation occurred they had produced, without con- jugating, an average of 721 generations; and without en- eysting, 197 generations. Throughout this entire period there was remarkably little difference in the rate of fission in the five lines. The total number of fissions produced by these lines during the 40 days.immediately preceding the appearance of the mutation was respectively 164, 171, 351 352 THE AMERICAN NATURALIST [Von. LI 168, 166, and 168. Thus it is obvious that in the ancestors of the mutants nothing in the nature of mutations in the rate of fission had occurred for many generations. This indicates that mutations do not ordinarily occur in asexual reproduction. During the period of 40 days mentioned above, ending July 10, all of the lines were in excellent condition and not a single individual died. On July 12, however, one line died out and on the 14th three more died, leaving but one line. For several days preceding the temperature was very high. It was recorded twice daily and these records show that it reached a maximum on July 9, when it was 28.5° at 7:30 a.m. and 31° at 6 p.m. It was, how- ever, high continuously from the 4th on, and during this period reproduction was exceedingly rapid, as Table I indicates. It was at the close of this period of rapid mul- tiplication that the four lines mentioned above died out and it is probable that this extraordinary environmental condition had much to do with the nature of the variations in the progeny of the remaining line, although similar variations did not occur in four other groups of lines that were running parallel with the one under consideration. From the remaining line mentioned above five new lines were started on July 15. For the first five days, the rate © of fission in these lines was nearly the same, the total number per line being 26, 27, 28, 26 and 28, respectively. During the next five days the difference became somewhat greater, the total number of fissions per line being 21, 19, 18, 17 and 17, respectively. On the day following this period, the line which had produced 19 fissions died out and was replaced by a new line from the one which had produced 21. There were thus two lines having more rapidly, and three lines having less rapidly, dividing an- cestors, all, however, originating from the same indi- vidual. On October 14 the lines in both groups were in- ereased to five and thus they were continued until the close of the experiment, which extended through 315 days. The individuals in all of the lines in each group repro- No. 606] MUTATION IN DIDINIUM NASUTUM 353 TABLE I THE RELATION BETWEEN THE FISSION-RATES OF Two STRAINS OR GROUPS OF Lines ISOLATED FROM THE PROGENY OF A SINGLE DIDINIUM June 1, 1912 to May 27, 1913 Each column under the brackets represents a line and each number in the column the total number of fissions for five days; d indicates that the line died out; c, that it eneysted; the brackets show the ancestry of the new lines lines that died out are indicated in the table. Whenever a line died out more than once in a five-day period, as sometimes happened, it is recorded only once. : Average Total No. Fissions for Five-day Periods Pure Line with 553 + Generations without Conjugation First Second Group of Group of Lines Lines 23; 23 22 21 22 14 15 14 15 15 17 18 18 af 17 18 19 18 18 21 20 21 21 21 20 23 22 22 21 21 22 21 21 23 28 31 32 31 29 Total 164 171 168 166 168 30 d d d d 26) 27 28 26 264 21 19 18 17 17 20 17 d 22 d d 22 22 i7| S 15 15 15 17} 15 20 15 16 16 20 154 d 26 20 22 21 26 24 24 a, 19 24 184 25 18 24 19 ri 20 21 13 18 15 204 15 c 19 15 16 15 19 54 26 | 20 21 21 25 24 19 19 20 244 19} A bys ‘17 14 15 14 17 if 15 16 12 12 11 1 11 d 13 9 9 9 11 9 7 3 5 1 15 16 12 11 9 153 104 d 1 13 13 15 18 134 8 10 | 8 || 6 7 5 5643, 38 1-38 neS 9 9 12: 8% 354 THE AMERICAN NATURALIST TABLE I (continued) [Vou. LI at wW el 42 121-12 mH n 9i I0 9-10 10 | 10 9 9 10 | 10 10 d 10) 11 11 | e 41 | 12 d 13 d | 14 eer 15} 8 12 13 7-48 12-41 10i 10 41541 10; 10 H 10 -13 11 | 14 10: 10 7 8 91 138 | 7 | 12 10 14 8 | 12 ade 9 101 11 8 8 9 9 Average Total No. Fissions for Five-day Periods Pure Line with 553 + Generations without Conjugation Second Group of Group of Lines Lines iis) eT % 8 ia Gen 74 BTR] 12 9 9 d 12 9 Coach Tee aes) pi iaee 13} 1S: | 18 d | 8 8 9 9 128 83 11 WFH 9 8 3l 8 7 11 8 tlw wl daj elot T| 8 9$ 6 10 d 11 7 5 6 6 d 10 6 EIO RATEN ENE te ee oe ee ee ee ee 9} 4} ss Smuspisaan hace. 1o Gi olib adl bl dl 8 9 51 a EEN A N ie tN, a A e A a E 6} dla ilo nr | Tor 9} 64 mMM | | 12} 11 | ie 81 ee | s| 8 | 104 8} 3i elisi eel T 8l a 8 eee 6 agnosis Eie o EAEE. peA Nieman; 12 | Ri eilel S å 11 64 d Bi n 9 6 9 8 8 12 8 Et 440534. H 10 088 a0 18 16 14 102 H nHn HL B 9 19 10 14 94 1241S D 9 8 8 8 8 12 84 19: lei Mmi II 910 P isi 94 Fear a a a eee na On i eC eT e g iv 7 dies SE | 2 8 8 8 8 t 1 7$ HiG 6 7i ja 10} 6 peu OnE gin 4 10) eai g] e 7, g W 72 2 i) B 9u 8 8 8 114 84 dlld 10) o niii Bi . 98 fem Oh Ob ro diilad D a - GL Ol ey 4 n el Gg 8} 7 IZ: 12 d 10 8 9 9 9 114 9 Ca P 1B) 18 | in ae 8 8 9 8 114 . 88 12) 12 | I4 9 8 9 9 10 122 9 2 IEB 9 f 8 9 7 il 8 8 8 Ol 64-8 1 eB $ 83 6 A A d Hinn 9 7; € 1.8 6 10} 7 Si elél g 27) oe 74 54 Pid id) 7j d | 13. 5. 8 7 No. 606] MUTATION IN DIDINIUM NASUTUM 355 TABLE I (concluded) verage Total No. aa for Five-day Periods Pure Line with 553 = Generations without Conjugation Spee SET TCT Te Group of fanes of in } | tna. | pe seve Warne | | 9 9 1.38 7 7 | 6 6 ee ee. | 5 8 54 Drar aT ONC ete 7 12 8} | Re arene. n AARNE aE GES 7 9 | 10 9 9 | 6 5 $ d | 7 84 5} en, 16 | Ik | 13.1°14:| 14 d 10 % d | 9 133 82 | Pani ai eras wang d eh eg 8 6 ee ee 5 4| 4 7 4 Se Pa eT ea Te Se Pa Te | 4 83 42 8 |} 71 i 9 8 ich & thd 9 73 ——_—_ d 8 | 10 d | 10|| 6 d 8 5 | d 92 61 Total average generations per line in 315 days 838 + 634 + Number of Hage died out | 30 33 es encys | 8 3 Number of aed miem gag ockurred before transfer was | made | 0 duced at practically the same rate, but those in the former, considerably more rapidly than those in the latter. Dur- ing the 315 days each line in the one group produced ap- proximately 838 generations, 2% per day, and each line in the other group approximately 634 generations, 2 per day. The averages for the five lines in each group for five- day periods are presented in Table I and plotted in Fig. 1. By referring to this table and the figure it will be seen that the difference in rate of fission in the two groups remained fairly constant throughout the entire 315 days, and that in both groups the rate was high in July and August, 1912, after which it decreased considerably and then remained fairly constant. The fluctuations in rate of fission were closely asso- ciated with variations in temperature. This was true for twenty-four-hour periods as well as for the five-day periods given in the table. During July and August, when the fission rate was high, the temperature was in general much higher than it was during the rest of the time, when the fission rate was relatively low. At the [Vou. LI THE AMERICAN NATURALIST 356 ‘skup Gg Əpuə ay} Sulmp vurvs oy} AyVuessea sy sdnoa3 OM} oY} UL UOJSSY JO 93a BY} UT VoUBIBYIP II} IVY) VION ‘quoultiedxe ay} Jo suluu[seq oy} WOI Sskup sussfosqv ‘uoIssy JO ƏJVI Åep oSvavav endə saleurpag ‘sfkep vAsseoons JAY 107 dnoas yous uy seul, Əy} Jo Je Jo UOIsSSY JO 0} vA OY} Supsusovav Aq pəupezqo svar SdAAIND oY} up JUIOd yvy “Jueujsfoue 1əjze suopvriouo3 F zer pue uoNRsnfuoce ase suon -yaoues + IZL [UNPTATpuy a[suls e woay Zupeupo wnyapiqd Jo seuly zo sdnoag om} up selva uorssy Əy} Fupjueseadea SƏAIMO ‘P OIA AANA k > > ) \ \ 5 Tal NAE TA S ma z $ ) f s | WAV) MEAN | No. 606] MUTATION IN DIDINIUM NASUTUM 357 close of the experiment, however, the rate of fission was not as high as the temperature at this time would lead one to expect. The didinia in both groups appeared to be in poor condition. There were numerous very small in- dividuals produced and an unusually large number of monsters. Conjugation was prevalent, but it was almost impossible to induce encystment. The death-rate was, however, not abnormally high. Whether or not the lines would have recovered from this depression if the ex- periment had been continued, is a question which can not be definitely answered. Before the experiment was closed some cysts were se- cured in both groups. These were kept in a damp cham- ber as usual until the following year. Then they were put into culture fluids of various sorts containing paramecia; but only a few developed, all of which belonged to the more rapidly dividing lines. From these, five new lines were started and carried on for 40 days. During all this time the condition of the individuals was much as it had been immediately before encystment. Throughout the entire experiment the didinia were cul- tivated in rectangular watch-glasses having a depression with a curved bottom. These dishes were piled one upon the other and kept in a damp chamber. - All of the didinia were fed with paramecia from the same cultures. At each feeding an equal amount of solution was taken from two of the most vigorous of four pint cultures which were continuously kept in as flourishing conditions as possible by adding fresh water and a little timothy hay from time to time. The two equal quantities of solution were then thoroughly mixed and two drops of this mixture contain- ing numerous paramecia were put into each of as many watch-glasses as there were didinia cultures. One drop of solution containing one didinium was then taken from each of the didinia cultures and added to’each of the watch-glasses containing the paramecia. The remaining didinia, after recording the number of generations pro- duced, were destroyed or used in studying conjugation 358 THE AMERICAN NATURALIST [Vou. LI and encystment. During the coldest weather it was suffi- cient to transfer every other day, but during the warmest weather it was found advantageous to transfer twice a day. Nothing was sterilized in these experiments, but the same pipet was used in all transfers and the watch-glasses not in use were exposed to the air and allowed to dry. Moreover, from time to time the didinia in each line in either group were transferred directly to the watch- glasses from which the didinia of the other group had just been taken. In these dishes there always remained considerable solution, in some instances a drop or more. Furthermore, in a few cases didinia from the more rapidly dividing lines were transferred directly without the addi- tion of fresh food to dishes in which more slowly dividing lines had died or from which all of the didinia had been removed. Such treatment had no appreciable effect on the relative rate of fission in the two races. It is obviously evident, therefore, that the difference in the rate observed was not due to difference in the bacterial contents of the solution if there really was any such difference, nor was it de- pendent upon selection, natural or otherwise, for mem- bers of the more rapidly dividing pairs were always trans- ferred in all lines. And the number of lines lost by death and encystment was essentially the same in both. In the one 30 were lost by death and 8 by encystment, in the other 33 by death and 3 by encystment. Assuming that the weaker lines died out in every case, it is evident that in this respect both races were subjected to practically the same sort of selection. And since all of the cultures were subjected to the same conditions otherwise, it is clear that the difference in the rate of fission in the two races must have been due to the constitution of the organisms. We have consequently demonstrated that marked vari- ations in the rate of fission may appear quite suddenly in the progeny of a single individual without conjugation or encystment, that some of these variations are heritable, and that they can probably be produced by subjecting the individuals to abnormally high temperature. - No. 606] MUTATION IN DIDINIUM NASUTUM 359 By referring to the table it will be seen that the mu- tation investigated originated, as previously stated, at the close of a period of extraordinarily high rate of fission and immediately after a short period of very high death- rate in which all but one of the lines died out. At the beginning of this period and at the close of the preceding period the individuals were very small and showed all the characteristics in behavior common to individuals about to conjugate. Whether or not anything in the nature of a nuclear reorganization in preparation for con- jugation occurred in these didinia is not known, but or- dinarily such phenomena do not begin until some time after union takes place in conjugating specimens. More- over, the period between fissions was not long enough to admit of much in the way of reorganization aside from what ordinarily occurs during the process of fission. Whether or not the ancestors of the mutants were actually homozygous is not known. If they were not the mutation may possibly have been due to a rearrangement of unit characters represented in the chromosomes during fission resulting in a change in dominance. However, if this did actually take place it is not in accord with the results ob- tained in very extensive investigations, all of which seem to show that changes in dominance do not occur in asexual reproduction. It is probable, therefore, that the mutation was due to a direct effect of the environment on the physi- ological processes in the organism and not to inherited nuclear phenomena largely independent of the immediate environment. The mutation theory so ably championed by DeVries has of late lost greatly in prestige, owing largely to the contention that the plants (Gnothera) in which DeVries discovered mutations were hybrids. If the conclusion reached in this work proves to be correct it will strongly support the theory in question. It will demonstrate that marked variations may appear suddenly in organisms re- producing asexually, that such variations may be heritable and that they may have a decided evolutionary value. 360 THE AMERICAN NATURALIST [Vou. LI This conclusion, though in opposition to a great bulk of the experimental evidence gathered by some of the fore- most biologists, Johannsen, Maupas, Morgan, Castle, Jen- nings and many others, is supported by some of the re- sults obtained by Barber (1907), Calkins and Gregory (1913), Middleton (1915) and Jennings (1916). SUMMARY In a race of didinia originating from a single individual there suddenly appeared a heritable variation in the rate of fission. This variation occurred 721 generations after conjugation and 197 generations after encystment. Two strains were isolated from this race and kept under observation for 315 days. During this time the lines in one strain produced an average of 838 + gen- erations (2% per day) and those of the other 634 + gen- erations (2 per day). THE METHOD OF EVOLUTION FROM THE VIEW- POINT OF A GENETICIST" A. FRANKLIN SHULL UNIVERSITY OF MICHIGAN A symposium upon the method of evolution, participated in by students from widely different provinces, ean hardly be expected to develop harmony of opinion. But to be assigned a place in such a discussion as a representative of one field of endeavor does not imply that the conclusion one reaches will necessarily differ from that of his co- symposiasts; for the question of the method of evolution is a question of fact, and when the fact is discovered, it will be recognized as completely by the paleontologist as by the physiologist. And, in particular, to be assigned the final place in the argument does not mean that the doctrine preached will be, or even is, the last word on the subject. In the present state of knowledge of biology the most that can be expected of an address on this subject is a statement of principles which should guide us in a search for the facts. Beyond these principles there are justifiable suspicions, and there may even be militant con- jecture, but little else. ` The first fundamental prindiple for the guidance of one who would find the method of evolution is a principle common to all sciences which seek to explain the occur- rences of a remote past. No agencies may be assumed to have operated fifty million years ago of a different order from those that operate to-day. If the phenomena of the present afford a plausible, or even possible, expla- nation of the past, there must be no appeal to other phe- nomena, the like of which do not now exist. Just asa geologist mentally constructs the rock strata a thousand feet below the surface, and the glacial drift of regions 1 Concluding paper of a symposium on the method of evolution before the zoölogical section of the Michigan Academy of Science, March 28, 1917. 361 362 -~ THE AMERICAN NATURALIST [Vou. LI now temperate, on the basis of processes now going on in certain parts of the earth; and just as the astronomer creates the planetary systems in his mind by forces that still govern, so the biologist must conceive evolution in the past to have been the result of agencies that are still causing change to-day. Whatever causes evolution now may conceivably have caused it during the early history of living things, and there are no circumstances which compel one to devise other causes for past change. Adherence to this principle automatically removes the . first solution of the problem of the method of evolution from the realm of the investigator who deals only with past events or the results of past occurrences, and places it in the hands of him who studies present-day. phe- nomena. Conclusions based on statistics have repeatedly shown how dangerous it is to argue from end results to causes. The compiler who finds that among the poorer classes of a population the ratio of male children to female is higher than in the well-to-do classes, and coneludes that deficient nutrition causes the high male-production, might also have discovered, had his investigations borne upon that point, that the poorer classes lived in houses pro- tected with a cheaper grade of paint or even without this protection. It would have been ridiculous to conclude that cheap paint favored boys, but- that conclusion would have been as nearly proven by the data collected as was the more plausible conclusion involving nutrition. Causes are not safely to be judged from results. In discovering the method of evolution, the initiative is denied the paleon- tologist, zoogeographer and the morphologist. No doc- trine of scientific cloture is here advocated, however, for the right of debate and even of veto is still theirs. The experimentalist alone may propose, but his colleagues employing the older forms of investigation may, and doubtless will, dispose. The experimentalist accepts his burden cheerfully. He knows that he may be unable to create a correct theory, but he prefers dispensing with a theory to adopting the wrong one. No. 606] METHOD OF GENETICIST 363 If you grant that evolution in the past was caused by the same agents as cause evolution to-day, to what phe- nomena of living things will you apply this principle? Evolution requires two things, namely, modification and inheritance. Given these two things, the occurrence of a new characteristic and the inheritance (or even only the heritability) of the new characteristic, evolution has occurred. It matters little now what becomes of the indi- vidual or individuals possessing a new heritable char- acter. They may even perish before they leave offspring, yet evolution has occurred. What became of these incipient races was the theme of the evolutionists of the past half century, who devised many and fanciful theories to account for their preservation or their destruction. To-day we are concerned less with the fitness of the new characteristic for the environment; we demand rather to know how the new feature arose and why it was inherited. Fortunately there is no fundamental disagreement with regard to inheritance. Too much is known of the mechan- ism of inheritance to allow of dispute. The chromosomes have been saddled with the main responsibility. There has never been any general attempt to refer inheritance to the environment. No one has supposed that a goose egg laid in the sand would produce a turtle. It is true, the cytoplasm has a share in determining what shall develop from an egg; so does oxygen, and so do other components of the medium, as can be readily shown by altering those components. What develops in the pres- ence of this cytoplasm, and out of this cytoplasm, depends specifically, however, upon the chromosomes. Disputes regarding this fact have seldom been dragged into argu- ments over the method of evolution. With the primary requirement of evolution, the pro- duction of new characters, matters have been otherwise. The mode of origin of modifications has not shared the good fortune of the mechanism of heredity. There is no need to cite the hosts of opinions that have been held by reputable scientists regarding the inception of evolu- 364 THE AMERICAN NATURALIST [Vou. LI tionary change. They have ranged from those who would make the living world of to-day wholly the product of the environment, to those that deny any participation of external factors in the course of evolution. Where in this array of opinion is the probable truth? To answer this question will be to express another opinion; but it is possible to formulate an opinion which is based upon principle, and which will therefore be more inviting than mere conjecture. With the aid of Sir Charles Lyell, who more than any one else has taught us to seek the explanation of past events in present processes, let us look about us for the cause of diversity among individuals. We may ignore differences which, from their fundamental nature, are not permanent, that is, modifications which are not heritable; for of such as these evolution is not made. Since inheri- tance depends upon the continuity of material of the chromosomes of the germ cells, changes in adult struc- ture or function can only be permanent when they follow a corresponding change in one or more chromosomes. These chromosomal changes may conceivably arise from within, or be impressed from without. Much of the mod- ern investigation which has a bearing upon the method of evolution is concerned with the question whether the modifications of chromosomes are caused by internal or external agencies. Two of the most striking cases of the origin of new heritable characteristics are those of the fruit fly Droso- phila and the evening primrose (Enothera. Scores of permanent changes in these organisms have appeared within the last decade. These new features have appeared in one individual among a hundred in the same bottle, or among a thousand in the same field. Environmental dif- ferences seem excluded in these cases. For, if one attrib- utes these changes to invisible and unsuspected variations of the environment in circumscribed regions in a bottle or field, there is no need to appeal, as has usually been done, to the grosser elements like climate and medium, and No. 606] METHOD OF GENETICIST 365 evolution is once more made wholly speculative. Inas- much as the very instability of protoplasm, which accounts for the manifold metabolic processes that char- acterize living things, makes not only possible but highly probable alterations of the chromosomes which, in our present state of knowledge, must be regarded as of internal origin, the discovery of cases like those of the fruit fly and the evening primrose, in. which environ- mental agency is apparently inadmissible, should leave no doubt that evolution can occur without reférence to- specific elements of the outer world. What the actual method of producing changes in the chromosomes may be can only be conjectured. Morgan and his students have abundantly demonstrated that the continuous identity of chromosomes is, in at least one animal, an invention based on appearances; that the chromosomes of one individual are often not identical with those of its parents. The crossing over which they postulate is an even exchange of corresponding parts of two chromosomes. How this exchange is brought about, whether through the twisting of the chromosomes as the students of Drosophila have assumed, or because of the variability of the forces that hold the chromosomes together, as Goldschmidt (1917) suggests, is immaterial. If, occasionally, this exchange between the chromosomes is not equal, an occurrence that is not inconceivable, a chromosome might be produced unlike any that ever existed. If the germ cell containing such a chromosome were capable of producing a viable individual, to predict the probable nature of such an organism would be idle speculation. Failure of the chromosomes to divide, and the passage of one or more of them bodily to one end of the spindle, would produce daughter cells with an unequal comple- ment of hereditary material. Hyde (1916) has recently reported a case in Drosophila which is probably of this _ nature; the two X chromosomes appear to have remained undivided, going to opposite daughter cells, resulting in 366 THE AMERICAN NATURALIST [Vou. LI the production of right and left eyes of different sex- linked colors. If, among the descendants of such unusual cells, germ cells should be produced, a new type of organ- ism should result from their development. Although Babcock and Lloyd (1917) reject somatic segregation because one supposed case of that phenomenon reported from the Oregon Agricultural Experiment Station proved, in their opinion, to be something else, there is still some evidence, like that of Hyde’s, that such unequal divisions’ do occur. If they occur in the line of the germ ‘cells, new modifications may thereby be produced in the next generation. Failure of the maternal and paternal chromosomes to separate in the reduction division of maturation, a phe- nomenon discovered in Drosophila by Bridges (1913) and named by him non-disjunction, may also be the cause of occasional evolutionary changes. The foregoing chromosomal irregularities, which have the appearance of being mechanical rather than chemical . phenomena, are not, however, necessarily the instruments with which permanent changes of organization are wrought. Probably they are not the usual ones. They have been mentioned first because there is evidence that such changes are occurring now. Minute changes far below the present limits of visibility are as conceivable, and in my opinion quite as probable, as the grosser ones named. In what these changes occur no one knows, for no one knows the nature of the hereditary elements. Suggestions involving enzymes and side chains have been made. These are only conjectures, but they reveal a belief that the phenomena of inheritance are chemical phenomena. If we believe that heredity is dependent upon chemical processes, there seems to me no escape from the assumption that evolution is first of all a chem- ical change. What the cause of these changes may be is another question; but if changes in considerable frag- ments of chromosomes, or even in whole chromosomes, ean occur as a result of agencies within the organism, as No. 606] METHOD OF GENETICIST 367 is plainly the case in Drosophila to-day, there is no reason to deny that the invisible modifications of chromosomes, if such oecur, are likewise of internal origin. In suggesting possible sources of internal change result-., ing in evolution I am not blind to the fact that an ultimate explanation of the method of evolution is not thereby offered. The chemical processes which cause these phe- nomena, while they are distinctly within the field of the geneticist, are not within his knowledge. If there were any prospect that an ultimate solution of the problem of the causes of germinal changes could be offered at the present time, invitation to participate in this discussion should have been extended to a physiologist; for it is from him that the eventual explanation of these internal changes must come. In this account of possible ways in which changes in the chromosomes of germ cells arise, I have not forgotten that it is conceivable that the changes are forced by external agents. There are, indeed, biologists who regu- larly attribute such changes to environment. The paleon- tologists not infrequently seem to regard evolution as ordinarily so caused. But with a few exceptions, those who hold these views are not experimentalists. They are not the biologists who are engaged in studying present phenomena. They reason from results to cause. Out of the conceivable causes they have picked on one which has a chance of being the right one, but only a chance. I venture to suggest that the theory of internal origin of modifications will account for all paleontological, morpho- logical, and geographical phenomena, and accord with all evidence from those fields, quite as well as the environ- mental theory. Among the experimentalists, it is to be admitted, there are a few who occasionally proclaim the discovery of a modification produced by the environment and subse- quently inherited. By one or two, not possessed of the still small voice, these proclamations are made repeat- edly. Sometimes the effect of the environment is admit- 368 THE AMERICAN NATURALIST [Vou. LI tedly directly upon the germ cells, and the results are not usually challenged. In other cases it is claimed to be only upon the soma, which then modifies the germ cells. These latter claims, however, meet with singular indif- ference or even distrust on the part of other biologists. Vulnerable places are too easily found, such as the lack of adequate controls. Sometimes the environmental evo- lutionist is charged with unwillingness or inability to show his hand when pressed for further information. Furthermore, it may seem strange that in a world of biol- ogists, all anxious to solve the problem of the method of evolution, and all so far as I am aware willing that that method should be anything whatever, all of the important supposed eases of permanent modification caused by environment should be advanced by a handful of investi- gators. To conclude: We have affirmed our adherence to the principle that evolution in past time is to be explained by phenomena that occur to-day. No processes that do not occur in living things now may be assumed to have occurred in living things formerly, unless there is plain evidence that events not explainable in terms of modern metabolism once occurred. Applying this principle only to the origin of modifications, not to their preservation, we have shown that animals are evolving now through agencies within themselves, independent of the environ- ment. Whether environment also produces permanent modifications is questionable, with the burden of proof still resting upon those who hold that it does. All of the known steps of evolution may be explained as originating from within the animals’ organization. There is no neces- sity of appealing to any other mode of origin, except, perhaps, to satisfy a certain type of imagination. In view of these considerations, it seems not illogical to me to suspect that evolution, at least among all but the very low animals and plants, is usually if not always initiated by a chemical change, either directly or indirectly pro- duced, in the chromosomes of the germ cells; that these No. 606] METHOD OF GENETICIST 369 changes are inherited because they result from changes in the chromosomes, and for no other reason; that such changes are usually, if not always, independent of the environment; that such changes produce unpredictable changes in sdai structure or function; and that these changes have no reference to the usefulness of the change in the evironment in which the animal exists or in any other environment. If one desired to go beyond the first steps of evolution, and discuss the factors that determine the course of evo- lution by effecting the survival or destruction of such new forms, it would not be difficult to maintain that sur- vival is much less dependent upon fitness than is com- monly supposed, and that natural selection probably operates only to eliminate the most unfit. But such a proposition necessarily involves much speculation, with comparatively little information regarding present day phenomena to serve as guide. I shall content myself, therefore, with the above categorical statement of views regarding the origin of permanent modifications, and allow my colleagues to begin the sifting and testing opera- tion which is theirs to perform. LITERATURE CITED. TUR E. B., and Lloyd, F. E. : Somatic Segregation. Journal of Heredity, Vol. 8, No. 2, Feb- ruary, pp. 82-89. Bridges, C. B. 1913. Non-disjunction of the Sex Chromosomes of Drosophila. Jour- nal of Experimental Zoology, Vol. 15, pp. 587-606. Goldschmidt, R. 1917. Crossing over ohne Chiasmatypie. Genetics, Vol. 2, No. 1, January, pp. 82-95. Hyde, R. R. 1916. Two New Members of a Sex-linked Multiple (Sextuple) Allelo- morph System. Genetics, Vol. 1, No. 6, November, pp. 535-580. SHORTER ARTICLES AND DISCUSSION AN INTRINSIC DIFFICULTY FOR THE VARIABLE FORCE HYPOTHESIS OF CROSSING OVER THE assumption of a ‘‘variable specific force,’’ made by Gold- schmidt,* may seem to account for the frequency of the crossovers occurring in a given simple case of linkage; but when this ex- planation is extended to the results which such crossovers give when bred, it creates a difficulty of the same type and magnitude as the original problem of crossing over, for which, therefore, it is not a satisfactory solution. Briefly put, the explanation advanced by Goldschmidt as- sumes that the genes are carried by the chromosomes, and that each gene is incorporated in its characteristic locus by virtue of a force residing in the chromosome and possessing properties specifically related to the properties of the genes of that locus. In the heterozygote Gg (see accompanying figure, line 1), the two forces F, and F, residing in the homologous chromosomes C and C’ possess not only a locus specificity but also an allelomorphic specificity corresponding to the allelomorphs G and g. When the chromosomes of the Gg heterozygote go into a resting stage, these forces F, and F, relax, so that the genes G and g become freed. When the chromosomes are reassembled preparatory to division these forces again come into play with the result that gene G is again incorporated into the chromosome in which F, resides, while gene g is likewise reincorporated into the homologous chromosome characterized by the presence of F,. In order that crossing over may occur, the allelomorphie specificities of forces F, and F, must, in the first place, be commensurable variables; i. e., forces F, and F, must vary in that property which con- stitutes their essential difference, and in such a manner that when all the values of force F, are represented by a character- istic frequency distribution and likewise all the values of F, are represented by a second specifice distribution, these two distri- butions will have a common base (see diagram, line 1). In the second place, these two distributions must overlap on the common base line so that a value chosen from the lower range of one may 1Dr. R. Goldschmidt, ‘‘Crossing over ohne Chiasmatypie?’’ Genetics, 2: 82-95, 370 No. 606] SHORTER ARTICLES AND DISCUSSION 871 be of the same magnitude as a value taken from the upper range of the allelomorphic distribution, though the two forces thus chosen are, of course, no more identical than are two of Johann- sen’s beans which are of the same size but belong to different pure lines. It is then assumed that in those cells in which the values of F, and F, are equal, the chromosome carrying F, should incorporate gene g as frequently as gene G, and in those cells in which their normal order of magnitude is inverted, the crossover incorporation should occur more frequently, depending on the amount of inversion. Let us assume that in a given case this overlap is of such a per cent. that one per cent. of the gametes are crossovers (%1 of the diagram). Now, in order to present the crux of the matter, let us proceed with the analysis of the behavior of the crossovers produced in the above experiment (see diagram, line 2). Let us mate a cross- Heterozygote heterozygote New heterozygote obtained from two crossovers Gametes of heterozygote 372 THE AMERICAN NATURALIST [Vou. LI over individual in which gene G is held incorporated by force F, with the converse crossover individual in which gene g is held incorporated by force F, (see diagram, line 3). As soon as the ch1omosomes of the resulting heterozygote enter the proper resting stage the forces F, and F, relax, freeing the genes G and g. It must now be recalled that every value of force F, is a member of a specific frequency distribution representing the entire behavior of F,, and that any particular value of force F, should give in succeeding generations the same result as every other value of F,. Each value of F,, whether chosen from the extreme upper range, the extreme lower range, or the mid-region, should give rise among its descendent cells to a series of variates which reproduce the original distribution F, and no other. That is, the two distributions which describe the variates of F, and of F, in the cells of the new heterozygote, being specific, overlap in exactly the same fashion and to the same extent as did the distributions of the forces F, and F, in the original hetero- zygote (see diagram, line 3). Consequently, when the chromo- somes are reassembled force F, will, as before, incorporate gene G in 99 per cent. of cases and gene gœ in 1 per cent. of cases (see diagram, line 4). But gene G entered the heterozygote as part of the chromosome possessing force Fy, hence the 99 per cent. of emerging offspring in which gene G is incorporated by the chro- mosome bearing F, or gene g by the chromosome bearing F are crossovers. As everyone acquainted with linkage knows, the crossovers given by the heterozygote from the mating of two crossovers are of the same frequency as in the original exper- iment. The intensities of coupling and of repulsion are equal and not complementary. Goldschmidt’s machine which at the first revolution turned out a mere driblet of crossovers, should overwhelm the operator with a deluge of crossovers at the next turn of the crank. The whole explanation fails unless some added agency be devised to take over the duty which the specifie allelo- ` morphic forces abandon after the occurrence of crossing over. The original problem was to secure the replacement of gene G in chromosome C by gene g, and at the same time the replace- ment of gene g in chromosome C’ by gene G. Having assumed the machinery of specific variable forces to accomplish this inter- change, we find that the products of the interchange are not stable, and furthermore they give a result the opposite of that demanded by the well known facts of linkage. In order that No. 606] SHORTER ARTICLES AND DISCUSSION 373 gene G should be stably related to its new position in chromo- some C’ it must be held incorporated by force F, and not by force F, as is the case. Added on to the original problem of the interchange of the genes is now the second and equally imposing problem of the interchange of the forces subsequent to the inter- change of the genes. An actual bodily interchange of the forces seems impossible in view of the assumptions we have had to make as to their nature and action. The transformations would then have to be accomplished by some transmutation in situ. It is evident that no internal autonomous change short of a complete and absolute mutation of force F¥ in chromosome C into F, and simultaneously of F, in C’ into F, would suffice. But we have no precedents for assuming such reciprocal mutations, and if we had, we could have sidetracked this whole machinery by applying this reciprocal transmutation idea to the genes and thereby solved the first problem in such a way that the second could not arise. Instead of localizing the cause of the reciprocal transfor- mations of the forces in the forces themselves, one might transfer it to the genes; i. e., one might endow the genes with the power _ of causing reciprocal transformations of the forces rather than empower the forces to transmutate of their own accord. While this form of the transmutation idea carries something of an air of plausibility, it can not be taken as more than an attempt at formal escape from the difficulty—a lifting of one’s self by one’s boot straps that makes more demand on credulity than, for ex- ample, one would in assuming crossing over offhand as a specific property of genes which needs, as support, only such formal ex- planation. ` CALVIN B. BRIDGES COLUMBIA UNIVERSITY ON THE PROBABLE ERROR OF MENDELIAN CLASS FREQUENCIES Ax old friend of geneticists who dislike excessive calculation has recently been attacked by Pearl,’ viz., the familiar formula, ¢==\/npq for the standard deviation of a Mendelian class fre- quency. He proposes to substitute a more refined but much more complicated method, originated by Pearson. In a Mendelian illustration he obtains a result which differs by over 40 per cent. 1 Pearl, R., ‘‘The Probable Error of a Mendelian Class Frequency,’’ AMERICAN NATURALIST, Vol. LI, pp. 144-156, 1917. 374 THE AMERICAN NATURALIST [Vou. LI from the usual. This seems to indicate that the old method is wholly inadequate, but further examination shows that the differ- ence is not due so much to method as to the fact that Pearl has calculated something with a different significance from the usual probable error. A cross of Mendelian heterozygotes (Blue Anda- lusian fowls) gave three classes of young in the numbers 14:33:11. Expectation is 14.5:29:14.5. Pearl assumes that a first sample of 58 has given exactly expectation and then cal. culates the quartile deviations for each class in a second sample of 58. The results are given as 3.13 for the heterozygous classes, 3.55 for the homozygotes which indicate an excellent fit of ob- servation to expectation. By the usual method, if a first sample of 58 had given exactly 14.5 black chicks and nothing were known of any theoretical expectation, the probable error in a second sample of 58 is measured by the probable error of differences. The probable error of either sample as given by the formula .6745\/npq is 2.22. The ee error of differences by the usual formula .6745\/c,? + o, is 3.15. This does not differ ap- preciably from Pearl’s quartile of 3.13. Neither of these methods, however, gives what we really wish to know, the close- ness of fit to Mendelian expectation. We have a theoretical ex- - pectation which is not based merely on a particular sample of 58, but which should hold with increasing accuracy the larger the first sample taken. With an infinite first sample, the formula given by Pearl reduces to the usual one, .6745\/npq giving a quartile of 2.22. This is less lenient to the discrepancy between expectation and observation than the first result, but the fit is still not bad. In a second illustration which is given, we do have two samples and no theoretical expectation suggested. The usual method of comparing samples of different sizes would be to find the standard deviation of differences on a percentage basis. The percentage standard deviation for a sample of n individuals is \/pq/n, for a sample of m individuals is \/p’q’/m and for differences is V (pq/n) + (p’q'/m). The expected standard deviation of a sample of m individuals is, however, my (pq/n) + (pq/m) if p and q are based merely on the first sample as in Pearl’s illustration. The formula given by Pearl for the standard deviation rapidly approaches this form for large values of m and n. Following are the results given by the long method, by an approximation given by Pearl and by the usual one just cited. No. 606] SHORTER ARTICLES AND DISCUSSION 375 | | Long Method | Approximate Method | Usual Method | | Median | 83.53 | 83.95 83,71 ower quartile 75.61 75.84 75.64 Upper quartile L TOLE a dooi OAO 91.79 The usual method gives substantially the same result as the long one and a better result than the approximate method. From the nature of experimental work, great refinement in statistical treatment is often a waste of effort, and without questioning the value of Dr. Pearl’s suggestion in cases in which the greatest accuracy is warranted it appears that the simple formula is still adequate for most practical purposes. SEWALL WRIGHT BUREAU OF ANIMAL INDUSTRY, March 12, 1917 CHARACTERS INDICATIVE OF THE NUMBER OF SOMATIC CHROMOSOMES PRESENT IN (ENOTHERA MUTANTS AND HYBRIDS THE pollen grains of 28-chromosome @nothera Lamarckiana igas de Vries were long ago shown (Lutz, ’09)* to be character- istically 3 + -lobed (chiefly 4-), instead of 3-lobed, as in O. La- marckiana and other diploid forms. Gates has since contributed much to our knowledge of this subject. Recently Bartlett (’15)?, in discussing the 3 + -lobed condition of the pollen of 28-chromo- some O. stenomeres mut. gigas, stated that these 3 + -lobed grains ‘fare larger than the triangular grains of the type’’ (O. stenom- eres). It may be added that the largest, best-appearing 3- of tetraploid forms in general, is larger than the typical, best-ap- pearing 3- of diploid, and the largest, best-appearing 3 -+ - of the former, larger than the typical, occasional 3-+-- of the latter. Smaller 3- and 3 + - grains are found in the pollen of both, but they are rarely perfect-appearing, and it is doubtful whether slightly imperfect-appearing grains are capable of functioning. A careful examination of the adult characters of a form, together 1 Notes on the first generation hybrids of (@nothera lata X O. gigas, Science, N. S., 29: 263-267. Gates (Pollen development in hybrids of Ginothera lata X O. Lamarckiana, Bot. Gaz., XLIII, 81-115, Feb., a had earlier observed 3 + -lobed grains in the pollen of a triploid form 2‘‘The Mutations of (nothera a . Amer. Jour. Bot., k 100- -109. 376 THE AMERICAN NATURALIST [Vou LI with a microscopical examination of the pollen of 10-15 buds from different parts of the plant will enable one to estimate its probable somatic chromosome number; this estimate becomes more trustworthy when one considers also (using Lamarckiana as the standard for comparison) the number of seeds produced per fruit by selfed flowers, the percentage of seeds which germi- nate, and the hereditary behavior of the plant. I All parts, or most parts, slender; pollen consisting of small, 3-lobed grains - with an occasional 3 + -. Plant will probably be found to be dip- loid, o r A, PERL of pollen, 50 per cent. or more of the grains good-ap- pearing. Abundance of seed secured from selfed oe germina- ting about as for O. Lamarckiana when sown in soil; spring, with the exception of a few mutants, Vig Westie the aN of the parent. Plant will probably have 14 chrom B. (a) Anthers barren as for Lamarckiana idia, or Oy Vanikie amounts of pollen produced in different buds of same plant; entirely ab- sent in some, present in small quantities or in moderate amounts in ‘others. In best buds, imperfect-appearing grains considerably in excess of good-appearing; in others, greatly i in excess of them and Lamarckiana flowers; lower eip capable of germinating, when sown in soil, than is usual for Lamarckiana seeds. Plant will probably have 15, possibly 14", or even 16, chromosomes, II Plants intermediate between O. Lamarckiana and Q. oe gig stoutness of all parts; pollen absent, present in sm eaka or in moderate amounts, much as for I B (a) and (b). Pollen contain- ing a mixture of 3- and 3 + -lobed grains, the former exceeding the latter in number. Largest grains larger than typical, best-ap- pearin ins of diploid forms; relatively few of the grains good- preng; even in best buds. Selfed flowers produce no seeds or very when s are sown in soil, very small percentage germinate Plant probably triploid, or approxima at 80; probably 21- pasas selfed 14-chromosome plant, or of 14 x 28; possibly 20 or 22 chromosomes. If paige by a selfed 15-chromo- some form, or by 15x28, the chances of its having 22 are greatly enhanced. és III All oa? stouter than for triploid forms; pollen grains characteristically 3 + -lobed, with relatively few 3-lobed (typical grain 4-). Larg- est and best-appearing 3- and 3 + -lobed grains larger than typ- No. 606] SHORTER ARTICLES AND DISCUSSION 377 ical, best-appearing 3- and occasional 3 + -lobed grains of diploid forms; 40 per cent. or fewer, good-appearing. Moderate amount of seed obtained from selfed flowers. Seeds large, germinate quickly. Plant will probably have 28 chromosomes, particularly if an offspring of a 14- or a 28-chromosome form, selfed, or of a 14 x 28 If the product of 15x28, it may have 28, 29, or even 30, chromo- somes, I. Forms which are approximately, but not precisely, tetra- ploid, may be wholly male-sterile. These statements are not intended to imply that all diploid, triploid and tetraploid forms have the characters enumerated above, but merely that forms displaying certain pollen conditions and vegetative characters will probably (by no means certainly) have the number of chromosomes specified. ANNE M. Lutz LAFAYETTE, INDIANA ON THE PERIODIC SHOREWARD MIGRATIONS OF TROPICAL NUDIBRANCHS' Many northern gastropods, including nudibranchs, are well known to exhibit the habit of congregating in shallow water along the shore at their time of breeding. This has been com- monly interpreted as the result of migration from deeper water at the approach of the egg-laying season. Certain species, at any event, are from time to time found in great quantity at shore stations which they do not frequent at other periods, and field observations have apparently established beyond a doubt that this inshore appearance is closely connected with mating and ovi- © position. The migration into shallow water, or other means which accomplishes the shallow-water flocking in these cases, may be regarded as a device which insures the concentration of individuals within a relatively small area, thus tending to make more certain the chances of pairing in a large number of in- stances, as well as a method of determining favorable conditions for larval development. Collectors of nudibranchs who have worked in tropical waters have also reported cases which at first sight seem to afford addi- tional examples of the coincidence of the spawning period with appearance in great numbers in the littoral zone (e. g., Cross- land, quoted by Eliot, 1904, p. 87). While engaged in working ong the shore during a period of some days or weeks, it is 1 Contributions from the Bermuda Biological Station for Research, No. 59. 378 THE AMERICAN NATURALIST [Von L- noticed that a certain species of nudibranch, until then found sparsely, if at all, suddenly begins to occur in abundance. It is also observed that at this time these nudibranchs are depositing eggs in the field, or that they pair readily and lay egg strings when kept in aquaria. The inference which has been drawn in such cases, namely, that the appearance in shoal water is in some way intimately related to the mating process, seems legitimate enough. But I have observed at Bermuda certain facts regarding the normal migrations of a member of the typically tropical genus Chromodoris which, it seems to me, cast considerable doubt on the theory that this species, C. zebra Heilprin, moves into shal- low water for the purposes of mating and egg deposition. The facts in this case, so far as they have been observed, are briefly as follows: It was necessary to obtain considerable numbers of C. zebra for use in experimental work (Crozier, 1916, 1916”); conse- quently collections were made at short intervals (every day during some months) over the period from August, 1915, to October, 1916. I had had occasion, also, to note the occurrence of this species in the summers of 1913 and 1914. In June, and during the early part of July, Chromodoris was found in great abundance upon the ‘‘eel grass’’ in certain tidal ‘‘creeks’’ (Fairyland Creek, Millbrook Creek). Subsequently, in the last two weeks of July and in August, they became very scarce in such places, although a few could almost always be discovered by careful searching. At other times of the year a supply of the animals was obtained on hard, open bottoms in somewhat deeper water (1 to 2 fathoms, at low tide), in places where, I am certain, they would never have been seen during ordinary shore collecting. Occasionally, however, as was noted particu- larly in December, 1915, Chromodoris was abundant along the rocky shores of smaller islands, ranging well up to low-water level. It seems clear enough that in Chromodoris zebra there un- doubtedly does occur from time to time a movement of numbers of individuals toward the shore. But there are several facts which sharply contradict the view that this migration is con- nected with reproduction. The nudibranchs pair in the lab- oratory and lay strings of fertile eggs at all seasons of the year (ef. also Smallwood, 1910), and not merely at the times when No. 606] SHORTER ARTICLES AND DISCUSSION 379 they are abundant near low-water level. Moreover, I have ob- tained the egg masses in dredgings at every season of the year; hence we may regard the fact of egg laying at all seasons under laboratory conditions as of significance in this connection. The eggs, which are quite characteristic in appearance, and hence easily identified, have been collected in depths of eight fathoms and more. Large individuals of C. zebra are likewise not un- -common at these depths; in fact, the first ones to be described were dredged from ten fathoms in Harrington Sound (Heil- prin, 1889, p. 187). A further point of considerable significance is found in the fact that these nudibranchs, unlike Elysia and certain other species, do not appear to deposit any egg masses upon the ‘‘eel grass’’ on which the animals occur in such great numbers throughout the early summer. The egg strings found in the field are invariably attached to rocks, or to the shells of Arca noe, the ‘‘mussel’’ with which the adults are frequently associated. The gelatinous egg-ribbons (cf. Smallwood, 1910) are quite large, measuring usually 120 to 150 mm. long by 15 to 17 mm. broad, and are much too heavy to be supported by a blade of ‘‘eel grass,’’ as can readily be deterthined by trial. It is only rarely that an egg mass has been obtained in shore col- lecting. The migration of C. zebra into shallow water cannot, then, be directly connected with reproductive activities. Since, in the laboratory at least, they deposit eggs usually within twenty-four hours after pairing, it does not seem to me probable that these nudibranchs pair to any great extent during the time which they appear to spend in the tidal ‘‘creeks’’—no eggs, as stated, having been collected from among the ‘‘eel grass,’’ nor were any ever obtained on the muddy bottoms of these ‘‘ereeks.’’ Chro- modoris seems to require a firm, hard substratum for the attach- ment of its egg-ribbon. If individuals obtained in quite shallow situations are kept singly in aquaria they sometimes deposit after several days fragments of egg-jelly containing several dozen unfertilized eggs, while they almost invariably pair readily when given the opportunity. Nevertheless, it should be stated that the nudibranchs usually do not occur singly, two or three being commonly found within a space of several square meters even when the total number of individuals in a given area is small; and I am well aware that laboratory findings with re to breeding habits are liable to be misleading. The established 380 THE AMERICAN NATURALIST [Vou. LI fact of egg production throughout the year in deep water is, however, good evidence that the periodic (or intermittent) abundance of this nudibranch in shoal situations can have little if any relation to oviposition. It might at first be suspected that the periodic shoreward movement represents the phylogenetic persistence of a well- defined habit possessed by not distantly related northern species. From this standpoint, reproductive functions in C. zebra might be conceived of as having become dissociated from the habit of migratory periodicity, since in warmer seas, where the seasonal alteration in physical conditions is reduced to a minimum, it is well known (ef., for example, Semper, 1881, p. 135) that many forms have no specially restricted time for breeding. How- ever attractive such a speculation may appear, it is eminently more satisfactory to regard these periodic littoral appearances of tropical nudibranchs as being controlled by definite physical influences in each individual case. Such directing causes would not necessarily be always the same for each periodic occurrence of the animals in shallow water. Although shoreward migra- tion and egg laying are closely connected in northern forms, it is still probable that physical circumstances in the sea imme- diately control the migrations even in this instance also. I have purposely refrained until now from discussing certain minor fluctuations in the littoral abundance of Chromodoris which are, nevertheless, important in connection with the idea that the supposed ‘‘migration’’ at certain times into very shallow water is, after all, only the unrestricted expression of a tendency to upward movement—negative geotropism. It has been men- tioned that during the greater part of the year Chromodoris was collected in 1 to 2 fathoms. But after storms of some severity they were to be had only in much deeper water. The nudibranchs undoubtedly move into deep, qufet places when the surface is greatly disturbed. Just what their behavior is under these circumstances can not be stated from direct observation, for obvious reasons; and for several days, or even for a week after a severe blow, the water in the sounds and bays remains so roily that it is impossible to see the bottom. But I have fre- quently observed individuals creeping up from deep water after the sea has become quiet and transparent. As regards the bear- ing of these facts upon the major flocking into the littoral zone, which occurs in early summer, it is to be noted that the mere No. 606] SHORTER ARTICLES AND DISCUSSION 381 continuance of quiet, still weather is not enough to determine the abundance of Chromodoris in the tidal ‘‘creeks,’’ since they disappear for the most part before the calm summer season is half over. The occurrence of individuals in deep water, together with field observations of specimens which were engaged in creep- ing downward on the sloping sides of rocks and reefs, leads me to doubt very much that any form of geotropic irritability exerts a preponderant control over the normal behavior of these ani- mals. My observations strongly suggest, however, that there does occur to some extent (in appropriate places) a diurnal ver- tical movement of Chromodoris, which is directly determined by the positive phototropism of these nudibranchs.* Specimens of the species known as Chromodoris roseapicta Verrill (there is some doubt that it is really a Chromodoris) have been found in littoral locations, only in the summer time, but this type is not sufficiently abundant to make possible a test- ing out of ideas concerning its migratory movements. The point which I wish to emphasize most is the uncertain nature of conclusions having reference to the normal behavior of animals inhabiting the warmer seas on the basis of compari- sons with superficial features of the movements of their rela- tives in colder waters. In the case of Chromodoris zebra, it seems to me definitely established that the periodic flocking of individuals into very shallow water has no immediate connec- tion with reproduction. On Jan. 10th, 1917, I found that C. zebra was crowding in great numbers into the entrance of Fairyland Creek. During the next few days they became very abundant indeed, so that on one occasion 230 of them were picked up in less than an hour’s collecting. On Jan. 12 I began to find egg masses attached to certain sponges, matted alge, mangrove roots, and sundry moor- ing stakes in the ‘‘ereek.’’ I had not before found any in this place, as stated above. The nudibranchs were observed in copu- lation, and great numbers of egg-masses were found. The at- tachment of the egg-masses was most frequently to some firm object. Within the week Jan. 10-17 they began to disappear, and after a fairly severe storm which came at that time very few were obtainable in the ‘‘ereek.’’? This occurrence seems to form a good 2I am anticipating here the statement of certain facts regarding the re- sponses of C. zebra which were established in this laboratory several years ago by Dr. L. B. Arey (ef. also Crozier, 1916°). 382 THE AMERICAN NATURALIST [Vou. LI instance of shoreward movement coupled with reproductive ac- tivity, but the fact remains that the nudibranchs do breed abundantly at other times and in much deeper water. REFERENCES Crozier, W. J. 1916*. Cell Penetration by Acids. Jour. Biol. Chem., Vol. 24, pp. 255-279 1916". Cell Penetration by Acids. III. Data on Some Additional Acids. Ibid., Vol. 26, pp. 225-230. 1916. On the Immunity Coloration of Some Nudibranehs. Proc. Nat. Acad. Sei., Vol. 2, pp. 672-675 Eliot, C. 1904. On Some Nudibranchs from East Africa and Zanzibar. V. Proc. Zodl. Soc. Lond., 1904 (2), pp. 83-105, pl. Heilprin, A 1889. Part The Bermuda Islands. Philad., vi+ 231 pp., 17 pl. Semper, K. 1881. Animal Life as affected by the Natural Conditions of Existence. Internat. Sci. Series, Vol. 30, xvi + 472 pp., New York. Smallwood, 1910. Mitek: on the Hydroids and ae of Bermuda. Proc. Zool. Soe. Lond., 1910 (1), pp. 137- AGAR’s ISLAND, BERMUDA. W. J. Crozier NOTES AND LITERATURE DEAN AND EASTMAN’S BIBLIOGRAPHY OF FISHES In order that the production and diffusion of knowledge may but be promoted, knowledge gained must be published in some permanent form. But when the publications become numerous and scattered throughout many journals, and in various lan- guages, it becomes at length difficult, or even impossible, for any human being to retain in mind all that others have discovered and written. The literature must be organized in such a way that the seeker after knowledge and the producer of knowledge may be enabled to determine easily what has been published on any particular subject. Hence the need for bibliographies and bibliographies of bibliographies, for the Zoological Record, and the International Catalogue of Science. Hence the justification, the necessity, for Dean and Eastman’s Bibliography of Fishes. Dean tells us in the preface that in this work there are listed more than 40,000 titles. How small a number of these could any man command were it not for some such collection! One volume only of the work has yet appeared. This is a book of 718 octavo pages of small print; and this is occupied simply by the authors’ titles of papers alphabetically arranged. And only those authors have been reached whose names begin with the letters A-K. A second volume is to follow which is to include the others. The time, the patience and the labor which the ac- cumulation of such a list demands may be surmised by the reader of the preface; it can only be realized by one who has tried his hand at something of the kind himself. As the work will then stand, the student of fishes can deter- mine readily all the papers that any author, as Agassiz or Baird for example, has written; or he can glance over all the 40,000 titles and pick out those which seem to have a bearing on his subject. To obviate the latter necessity, a third volume is to follow which is to be an index to the preceding volumes. In the two volumes of authors’ titles each paper is followed by the year of publication and a serial number, as ‘‘ Jordan, 1891, 4’’; and in the index each paper is to be referred to briefly by the author’s name, the year and the serial number. Economy of labor and 383 384 THE AMERICAN NATURALIST [Vou. LI expense is thus effected. The index will certainly be classified in such a way as to make it reasonably easy to arrive at the papers desired. In estimating, therefore, the work that Dr. Dean and his editor and assistants have done we must consider not only the collection and preparation of the titles, but likewise the analysis of these papers and the recording of the contents under their proper heads. There is a need for more yet to be done. The author tells us ' that the index does not include detailed references to species, genera, or ever, in many cases, families of fishes. ‘‘This would entail many years’ additional listing, but should unquestionably next be done.’’ The busy student may want to know what has been written on the Centrarchids, or the genus Lepisosteus, or the rainbow trout; and he ought to find all of the papers re- corded under each head. May the good men who have worked on this Bibliography of Fishes retain their powers and live long enough to accomplish the work. However, it will be open to any one to go through those 40,000 books and papers and cull out the things bearing on the subject he has chosen and to publish a little bibliography of his own. The present writer has not undertaken to discover omissions of papers or errors in quoting them. Certainly omissions and errors occur, as in any human production ; but doubtless all pos- sibla care has been taken to avoid them. Two omissions have incidentally been brought to notice. The first of these is a paper by Eigenmann on a fossil species of Sebastodes, in Zoe, Volume I, 1890, page 17; although another paper cited ends with page 15. The other paper omitted is B. K. Emerson’s ‘‘Geology of Old Hampshire County, Massachusetts,’’ in which there is a list of the Triassic fishes found in the state mentioned. Dr. Dean must have had a record of this paper.‘ A paper by E. W. Clay- pole? is quoted from the American Geologist, Volume XXIX, p. 44; but the paper is not found as cited; nor elsewhere, so far as the present writer knows. Ouver P. Hay 1 Science, Vol. oe 1902, p. 701. 2 Claypole, 189 THE AMERICAN NATURALIST Vot. LI. July, 1917 No. 607 RATS AND EVOLUTION A. C. HAGEDOORN, Mep.Arts, anp A. L. HAGEDOORN, Px.D. In treating a large group of animals from the stand- point of a systematical zoologist, it makes a very great amount of difference whether one does the work in the region inhabited by the animals, or somewhere else with the aid of collections in a museum. A real systematist, of the museum kind, does not come into touch with a number of very real problems which present themselves to field workers, and when he does, he has every induce- ment to brush them aside with an authoritative gesture, as he is not in a position to valuate their importance. He takes for granted that two similar skins with similar skulls which he receives from the same place, correspond to a multitude of individuals, all with these same char- acters; that they are a sample of a multitude of animals all exactly alike, and when he finds that animals of such a description have not hitherto been named, he can invent a well-sounding name for the two skins, and publish a description, and henceforth this description of the type specimen and this species name are welded together. If it so happens that an animal is never again collected which corresponds to the published description, the species becomes known as very rare. There exist conventional rules, which, in the descrip- tions of species in certain groups, ascribe more value to certain characters than to others. In the systematic clas- sification of rats, the points which are specially noted 385 386 THE AMERICAN NATURALIST [Vou. LI in this connection are the shape of certain ridges on the skull, pads on the soles of the hind feet, the relative length of the tail, the length of molar complexes, and the length of the ears. It is significant to. observe, how every field worker who occupies himself specially with rats has his own opinion about the relative importance of these different points for the systematic classification of the animals, and discovers very soon that the work done in museums does not materially help him in his quest. In 1915 one of us was commissioned by the government of the Netherlands to make a biological and zoological study of the rat population of the Dutch East-Indian col- onies, more especially of the island of Java, with the ultimate object to find out what measures could be taken to prevent the exceptionally serious damage to public health and to agriculture caused by rats. Some pre- liminary work on the subject had been done by medical investigators and by a systematist working with pre- served specimens in Holland. The systematic-zoological work in Java was begun some years previously by Maj. G. Ouwens, who is continuing the work after we were obliged, for reasons of personal health, to leave the tropics. Very soon after arrival we discovered how very little the work done in European museums was to help us out in the field. We are not systematic zoologists, and our rea- sons for accepting the task lay in the promise the material gave of throwing light on the question of species (in which it has not disappointed us). Therefore the only group of animals with which we have at all deeply con- cerned ourselves with systematics is the rat, and we would not be prepared to maintain that for other groups the ordinary museum-zoology has so little value in giving a conception of the relationship between species in nature. Still, the study of rats from a semi-economical point of view has certain advantages over purely scientific col- lecting, as the material studied is very plentiful, and an No. 607] RATS AND EVOLUTION 387 extraordinarily great number of keen-eyed persons, public health officials, anxious owners of coffee-plantations, managers of sugar factories, native officials in rice-grow- ing centers, are continually observing the animals, and are more than willing to collect extraordinary large numbers on request. Itis not uncommon for any one studying rats to see several hundred animals brought together for him to look over, and one of us has had the occasion to observe a batch of ten thousand rats in one day within the grounds of a sugar factory where between nine and twelve thou- sand rats were killed daily for several years. The study of rats has set several authors to speculate as to the nature and the origin of species. Very prominent amongst these is Lloyd (The growth of groups in the animal kingdom). Our conclusions differ materially from those of Lloyd, however. The reason for this dif- ference, we venture to think, lies chiefly in the fact that whereas Lloyd studied dead rats, and speculated upon the origin of his animals, more especially of aberrant types, we have been breeding rats for some six years, and have witnessed the origin of aberrant types. The examples in this paper will be found to be nearly all taken from rats. When it is found in field work, that two species-names, each given to a skin in a museum drawer, in reality corre- spond to two real groups in nature, of which they are representative, we may be dealing with one of two dif- ferent possibilities. It may be that the variability within the first group is not so great that individuals belonging to it fall within the limits of variability of the second group, or it may happen that two different skins in a museum belong to one highly variable group of animals, in which it is difficult to establish dividing lines. If, for instance, two skins with different names in a museum differ considerably in size, it may happen that even the largest animals of the group to which the smallest skin belongs are still very much smaller than the smallest adult individuals of the group which corresponds to the bigger skin. It may happen that two skins are consider- 388 THE AMERICAN NATURALIST [ Vou. LI ably different in a preserved state in respect to some salient character, whereas in nature this very character may be found to be so variable even within a small, - closely related family of animals, that it has no value whatever for distinguishing two species. A case in point is the presence or otherwise of flattened hairs, or spines in the coat of rats. On the other hand, it may happen that two species, if once they are dried and preserved in a museum, present no, or, no appreciable, differences, whereas in reality, these two species may be found to differ very definitely biologically. As an illustration we may cite the case of the field-rat and the tree-rat in Java. The easiest way out of the difficulty is the one taken by a great many zoologists, working through large collections of animals in museums namely, to give a new species-name to every animal which differs markedly from other described species, and which as yet goes with- out a name. But if one wants to go deeper into the subject, if one wants to know whether these species of the drawer have their counterpart in as many species in the forest and field, the task becomes more difficult and even hopeless for a great many investigators. As soon as specialists. take in hand some group or other, it is very soon obvious that the task of finding out just how many species they are dealing with and how they differ is very much more com- plicated than it looked when studying the collections in a museum, however well stocked. In treating rat material from a zoological-systematical standpoint, a number of problems confront the investigator from the very outset, and he must try to find his own solutions. Every inves- tigator treats the material in his own way, and where one man makes fifty species, some other man will make two species out of the same material. It is evident, that if the term ‘‘species’?’ means anything at all, it must hypo- thetically be possible to divide the material into a fixed number of species, neither more nor less. The vague way in which the term ‘‘species’’ is applied, must be chiefly No. 607] RATS AND EVOLUTION 389 responsible for the unrestricted feeling of personal lib- erty which systematists undoubtedly have about the way in which they divide a number of dried animals into -= species. It is for this reason that it here becomes neces- sary first of all to give our definition of the term ‘species. For numerous systematists, a ‘‘species’’ is the descrip- tion of a skin and a skull deposited in a museum— the type- - specimen—and to this species belong all the animals which have just such a skin and skull. Some few botanists are just now trying to reserve the term for a group of animals or plants which have the same genotype, the same set of inherited factors of development. As long as we concern ourselves with autogamous plants, such a definition might pass, we might, at least hypothetically, divide a popula- tion of such plants into a number of species and a few hybrid individuals. Itis very obvious that this definition of ‘‘species’’ falls short, as soon as we concern ourselves with animals, or with allogamous plants. In such groups, according to | this definition, there would be no species. Even the geno- typically purest group of animals would in every instance still be composed of two species, the males and the females, for we now know that the sex difference is caused by a difference in genotype! Therefore, such a definition of the term although very concise and very short, is prac- tically untenable. When we say: Species are those groups of individuals, which have a common genotype, and which are pure for that genotype, we can most certainly concede to Lotsy that species are not variable,? but if we do so, we limit the use of the old word ‘‘species’’ to those groups of plants which really are pure and therefore invariable, so that they can not be changed by selection, natural or artificial. -If we solemnly state that dogs have short twisted tails, 1 ‘í Mendelian Inheritance of Sex,’’ A. L. Hagedoorn, Archiv fiir Ent- wicklungsmechanik, 1909. . Lotsy, Handelingen van het Natuur en Geneeskundig Congres te Delft, 1912. 390 THE AMERICAN NATURALIST [Von. LI we are perfectly within our rights when we use the term ‘dog’ for bulldogs only. But such a statement brings us no insight in the shape or the length and variability of — the tail in the big group of animals which everybody, ex- cepting breeders of bulldogs, knows under the name of «d ogs.”? We can say: ‘‘Carriages have small wire rubber-banded wheels’’ and if so we are within our rights if we limit the . term carriage to baby-carriages, but all such and similar statements of wheat-growers, breeders of bulldogs and manufacturers of baby-carriages, no matter how plausible they may look to the people under consideration, have this one thing in common, that they may not be general- ized. Breeders of New Foundland dogs have as much right to reserve the name dog for their animals, and to say that dogs have long bushy tails, as the breeders of bulldogs did, and if we permit the manufacturers of gocarts to reserve the term ‘‘carriage’’ for their product, and if we allow the breeders of autogamous plants to limit the term species to species of wheat and barley and peas, manufacturers of Pullman carriages certainly have the right to state ‘‘Carriages are ninety-five feet long and are entered by steps four feet from the ground”’ and the breeders of sugarbeets or rye, and the zoologists will have the right to state that species are variable. When we want to make a definition of the term ‘species’? we must make it so that it fits rat-species as well as wheat-species, and in such a way that the gene- ticians as well as the systematicians can apply it to the things they are wont to call by the name. We know that all the different genes, all the different inherited factors whose cooperation or non-cooperation to the development of the most diverse organisms pro- duces the hereditary differences among them, are each in themselves invariable. We have called this invariabil- ity of the genes Johannsen’s law.* Only in this way can 3A. L. Hagedoorn, ‘‘Wetten en Regels in Genetica en Eugenetica,’’ Handelingen van het Genootschap van Natuur, Genees-en Heelkunde, 1913. No. 607] RATS AND EVOLUTION : 391 one explain that those groups of plants, which are so constituted that they become automatically pure in a short number of generation—the autogamous plants—con- sist in the main of pure and invariable species, which can not be changed by any amount of selection. Selection within a group of plants which descends from one indi- vidual, homozygous for all its genes by a continued auto- fecundation, is ineffective. As we have the name ‘‘pure line’’ for these groups of plants, there is no good reason to limit the use of the term ‘‘species’’ to these groups exclusively. Liability to change by selection is synonymous with genotypic variability, and this true variability is synony- mous with impurity. Those species which do not exist exclusively of individuals which are all mutually identical in respect to all their genes, are variable and therefore liable to change by selection. One single, genotypically pure species as a rule can not give rise to new species. There have become known a few cases® of real spon- taneous genovariation, mutation, in which every known cause for change in genotype was excluded (one of us has noted three such instances in the mouse) ; but as in every instance we have been concerned with a dropping out of one gene we can practically leave them out of account here. There exist pure species, but there certainly also exist variable species, species which are certainly liable to change by selection. In evolution we are certainly concerned with two dif- ferent sets of processes, on the one hand with the causes of variability, and on the other hand with the processes which limit variability. Throughout this paper we will call total potential variability the quantity of genes which not all the mem- oorn and A. L. Hagedoorn, ‘‘ Studies on Variation and Se- lection,’’ Zeitschr. fiir Induktive Abstammungs- und Vererbungslehre, 1914. A. C. Hagedoorn and A. L. Hagedoorn, ‘‘ Can Selection improve the ity of a Pure Strain of Plants?’’ Journal of the Board of Agriculture, 1914. 5A. L. Hagedoorn, ‘‘The Genetic Factors in the Development of the fouse-mouse,’’ Zeitschrift fiir Induktive Abstammungs- und Vererbungs- lehre, 1911. 392 ` THE AMERICAN NATURALIST (Von. LL bers of a group have in common, or for which they are not pure (homozygous), and the variability which this impurity makes possible in the descendants. At least ideally, we can express the potential variability of a group of individuals in a number. There certainly exist species with a total potential variability of zero; these are, for instance, the pure lines of certain autogam- ous plants, those species for which Lotsy would like to re- serve the term species altogether. We will now try, by the aid of this new term, total poten- tial variability, to give such a definition of the word ‘t species” that it comprises everything which zoologists and botanists, geneticians and systematists, have vaguely meant by it. Our definition is as follows: A species is a group of individuals which is so consti- tuted genotypically and which is so situated, that it auto- matically tends to restrict its total potential variability. Every group of individuals which is closed to the ad- mixture of individuals from without, such as the de- seendants of an autogamous plant, the dogs or cattle in an exclusive stud, a ‘‘Paarungsgenossenschaft’’ of ani- mals or plants bound by a peculiar habitat, has the tend- ency to become purer and purer automatically, and to reduce its variability continually. Species originate, given a certain variability of a group of individuals, through all those agencies separately or in combination which bring a group of individuals (not necessarily a small group) into such conditions that the new group has a tendency to become pure for its own genotype. We can not say in general that species are produced by inbreed- ing, or by isolation, or by a change of habitat, or by colonization, or by selection exclusively. An individual or a group must have a certain amount of potential vari- ability to be able to produce a species, different from the one to which it belongs. We know now that the genes themselves are invariable. There remain only very few authors who still believe in the variability of the genes. It is therefore necessary to No. 607] RATS AND EVOLUTION 393 find out the causes for genovariability. Real mutation, as far as we know, exclusively consists of an occasional loss of a gene without visible cause. Mutation therefore can at the utmost heighten the potential variability by one. De Vries’s conception of periods of mutation is at present only of historical interest. In our opinion, crossing, recombination of genes by mating of individuals of unequal genotype, is to be re- garded as the only real cause of variability. There is no good reason to change the opinion of one of us, namely, that there exist three different kinds of variability.® A. Modification, the non-inheritable effect of the non- genetic developmental factors. B. Real inheritable variation caused by mutation, loss of genes. C. Real inheritable variation by recombination of genes. Lotsy has subscribed to our statement (loc. cit.) with the exception that he denies the existence of loss-mutations. We can no more say that species originate by crossing, than, that they originate by isolation. New pure lines of autogamous plants, the kind of species for which Lotsy wants to reserve the term, can of course originate in the descendance of one hybrid plant. There is no funda- mental difference between evolution in these plants in a state of cultivation and what it must be in nature. But in allogamous organisms, we will only in exceptional cases meet in nature the same course of evolution as in our cages or experimental plots. Even if crossing in the widest sense is the sole cause of variabilty, we must not suppose that, as a rule, new species come into being in the F, or F, generation from a cross. If we make a hybrid between species, this hybrid indi- vidual will have a total potential variability which is at least as great as the number of genes which were not common property of both the forms crossed. If we com- oorn, ‘‘ Autokatalitical Substances the Determinants for 6A. L. H the Inheritable Characters,’ Roux’ serie Vorträge und Aufsätze über Ent- wicklungsmechanik, Leipzig, 1911. 394 THE AMERICAN NATURALIST [Von. LI pose a group of nothing but such hybrid individuals we will get an enormous amount of variability in succeeding generations, and when the group gradually becomes more and more pure for an own genotype this may be a com- pletely new one. A species may have been produced with totally new characters, possibly intermediary between the parent species in some of them. The chance that hybrids of allogamous organisms, even if they are viable and perfectly fertile, will inter se produce a new species is exceedingly small in nature. It is much more probable that the process of species formation after crossing is as follows: There exists a species A, with a restricted potential variability, a set of habits and mode of living all of its own, adapted to a certain environment. As a general rule, individuals of this species A mate exclusively with members of their own species. Once in a while, small groups may split themselves off from the multitude by colonization, and each of these groups will have its own potential variability, and each will gradually become pure for its own genotype, and will be less variable than the multitude. In the same country there exists a species B, with a slightly different genotype, a different potential variabil- ity. Species B is somewhat differently built, somewhat differently coated, compared with A, and therefore fits into a somewhat different environment. As a rule, indi- viduals of the two species do not come into touch. Let us take as examples the grey-bellied Mus alexandrinum which lives in houses and on roofs in northern Africa, and the white-bellied Mus tectorum, which lives in trees in the same countries. The same holds true for the house- rat and the field-rat in Java, likewise for the house-rat and the tree-rat. Even if matings between the two species furnish hybrids which are completely fertile, even in localities where two species overlap and are plentiful, the occasional hybrids will be far in the minority compared to individuals pro- No. 607] RATS AND EVOLUTION 395 duced by matings between house-rats and house-rats or tree-rats inter se. If the occasional hybrids grow up, they will either become house-rats or tree-rats, biolog- ically speaking. In the first case they will mate with in- dividuals of the house-rat population, in the other case with tree-rats. A new group, so situated that its potential variability is bound to be reduced to produce a genotype of its own or a new species, these, a few hybrid rats will certainly not produce. A single mating of a house-rat female with a tree-rat male may be the cause for a height- ening of the potential variability of the house-rat popula- tion into which the hybrids merge. Eventually this higher potential variability will be reduced again. And re- versely, an occasional mating of tree-rat females with- house-rat males may be the cause for a greater potential variability of the group of tree-rats to which the females belong. If it so happens that a few animals colonize out of such a population at the time when the potential variability is still higher than ordinarily, such a colony, which will have a potential variability smaller than that of the mul- titude, will have a chance of having a range of variability differing from that of the multitude. Such a group may become pure in respect to a somewhat longer tail, a some- what darker belly or a somewhat greater size, as com- pared to the population from which it ultimately was de- rived. Very good examples of such a process can easily be found by observing the evolution of certain species of dogs or poultry under domestication. For instance, the species Airedale terrier has become variable, and there- fore liable to the influence of selection in different direc- tions, because of the fact that hybrids with Dobermann pincher in Germany, and with the Gordon setter in Eng- land, have been taken up into the species, the stud not having been closed rigorously, such as the Sloughi stud, or the Jersey cattle stud. But it must not be thought that a new, improved species of Airedale terrier has been 396 THE AMERICAN NATURALIST [ Von. LI bred from such hybrids inter se. The potential variabil- ity of the collie was very small, a few years ago. Hybrids with the Russian wolfdog have been taken up into the species. For this reason the variability has been very extensive during a number of years. And at present this variability is reduced again, by selection. In the mean- time the species has been changed as a whole, the fashion having changed and having made much of the possi- bilities afforded by the cross. The collie, which formerly was an intelligent, affectionate dog, with a head shaped like a fox; : inclined to be noisy, and to run after every- thing ; with long straight, outstanding hair, and with color ranging between black and tan, and sable, with a variable -amount of white, has changed completely. The show collie now is rather a treacherous and surly dog, with a head shaped like that of a llama, silent and lazy, with hair which inclines to be soft and wavy. The color is much more variable and now includes white, slaty, creamy, and generally fade tints. In chickens, crossing is a common way of ‘‘improving’’ a species, and in all those instances we happen to know, the hybrids were made by using an individual presenting a character which it was thought desirable to fix into the breed, or a certain degree of development of a character, not reached by even the best individuals. Such an indi- vidual used for crossing is sometimes a pure-bred animal of another species, but much more often a mongrel of un- known extraction, happening to be somewhat like the breed to be improved, with the exception noted above. It is the practice to breed the hybrids obtained from the cross back to the species, and their offspring again, always selecting those individuals which come nearest to the general conformation of the species, but which have the character to be fixed into the breed. For instance one may want to breed blue Wyandottes. The breeder will then take any blue fowl which happens to look some- what like a Wyandotte, mate it into a strain of first-class white or black Wyandottes, breed the hybrids back into No. 607] RATS AND EVOLUTION 397 his species, selecting from among their blue offspring those which are the most like a good exhibition Wyan- dotte, and so on, for a number of generations. It is easy to see that in such a case the general potential variability of the whole group is very much increased. It diminishes automatically again, because of the fact that in every generation a few animals produce a great number of off- spring. If ten young cockerells of a new species of fowl were habitually derived from ten fathers, the progress in the direction of purification, ‘‘fixing’’ the breed, would be almost nothing. But as ten young males habitually have only one or on the average less than one father, in other words, as only a very small percentage of males in every generation is used for breeding purposes, auto- matic purification, automatic diminishing of the total potential variability of the group, is very rapid. It is to be noted that in the absence of selection, the group may become pure for almost any conceivable genotype given in the potential variability, the genotypic diversity of the first animals. Therefore any character which has re- ceived no or small attention from the breeder may turn out to be different from what it was in the species to be altered into a new breed. It is for this reason, very com- mon to observe that a number of apparently closely re- lated species in the common fowl, or in domestic pigeons, differ, not only in the points which are obvious to every observer, but in other minor points as well, points which need not be in any way correlative to the obvious dif- ferences. A few examples. The different species of Leg- horn resemble each other very closely, differing to a casual observer in color only. But the comb of black Leghorns is noticeably larger than that in white and brown Leghorns and the ears of the black species are larger than in the brown and the white. The white Leg- horn has a lesser tendency to become broody than the buff. The hens lay more eggs than those of the buff or the black breed. The plumage is generally looser and longer in buff Leghorns than in blacks. 398 THE AMERICAN NATURALIST [ Von. LI In the Wyandotte group of species, the texture of the comb is very different in the blue kind from what it is in the silver Wyandotte. The length of the tail feathers differs very much in the different breeds. The white Wyandotte lays dark brown eggs, the silver Wyandotte lays salmon-colored eggs with minute white spots, the black Wyandotte lays white eggs. : It is very rare for new species in chickens and pigeons to come up to the quality of old established ones, unless the fashion or standard happens to change. The shape and carriage of the tail, and the general shape of the body are very much better in white Fantail pigeons than in the newer black-tailed whites or white-tailed blacks. And the shape of the new blue Wyandotte is not yet what it is in the white and the silver. We know of only a few instances of new dog or poultry species being bred from hybrids inter se. In those cases the breeders had no very definite object in view to start with. This mode of origin of species under domestication is certainly not the common one. Species of cultivated animals are commonly being changed by a very notice- able conscious selection. The variability necessary for improvement is continually kept up, sometimes by de- liberate, but mostly by a kind of unintentional crossing, that is to say by admitting exceptionally fine offspring produced by matings of hybrids back to the species, into the registry. On the other hand, automatic purification, automatic reduction of the heightened potential variabil- ity, is the necessary outcome of the fact that only very few and very exceptionally ‘‘good’’ males are used as breeders. Species of tame animals, especially fertile ones as chickens, are easily kept apart so that excessive splitting up into secondary species is possible and even the rule. For instance, in those species in which a certain much sought quality is influenced by the internal secretion of the sex-glands, it is obviously impossible to make a pure strain in which both males and females come up to one No. 607] RATS AND EVOLUTION 399 standard. In silver Wyandottes, the standard calls for white feathers, which are bordered by a line of black. Now the males are very much lighter than the females, so that in a pure species, in which the males are correctly marked, the hens are too dark and in a strain in which the hens are good, the males are too light. The only way out of the difficulty has been the establishment of two dif- ferent species, one which produces correct males and the other which produces exhibition hens. This splitting up of a species into two is very common in chickens. Such pairs of two species are kept as far apart by careful breeders as Wyandottes and Leghorns. In the natural state, two species, even when hybrids between them are perfectly fertile, and when the indi- viduals exhibit no preference for mating with their kind, may keep apart, if only each group is specially adapted to an own environment, so that the bulk of the animals of each species has no chance of mating with anything but their kind. Even if there is no sdaplation to an environment to keep the multitude of the individuals of a species in their place, two species may keep apart when only the animal’s habits keep them from wandering very far. In such a case the borderline, where outposts of both species mix, will present a highly variable population of hybrids of all grades, the variation becoming less and less the more we look.for the animals in the exclusive territory of each species. A case in point is, we think, the case of the two woodpeckers cited by William Bateson.” When the spe- cies differ in only one salient characteristic, the difference between them being in the main due to the presence or absence of only one gene, intermediates must be absent. In such a case the two species may be present in more or less extensive patches, separated from each other by narrow strips of territory having a mixed population. Such seems to be the case of the black and the hooded crows in Europe. Here adaptation plays no rôle ap- parently. 7 William Bateson, ¢ Third, small letters affixed to the series number may be used to designate individuals of the first (or P,) gener- ation of any series, whether hybrid or not. The definition of individual would necessarily be determined and re- corded for each type of work; where desired, the ‘‘indi- vidual’’ may be a branch, or even a single bud or fruit. Thus, in my work with Citrus, a letter is assigned to each - self-pollinated branch giving seedlings. For example, series 28 consists of descendants of selfed Valencia orange; ‘‘28a,’’ then, indicates one bagged branch of Valencia, ‘‘28b’’ another branch, ete. The F, progeny will be 28a—11, 28a—12, 28b-32, ete. For a hybrid series, on this plan, two letters are used to designate the two parents, the letters being independent of those used in the corresponding non-hybrid series. For example, my Citrus series 27 represents Valencia orange 9 xX Imperial pomelo g, and ‘‘27aa’’ designates a cross between two particular branches. 27aa—21, then, will be an F; hybrid. Similarly, in my work with Matthiola mutants, series 23 indicates the ‘cross Snowflake (‘‘normal’’) type Ẹ X Slender type £. ‘‘23ea’’ designates the combination of 5 The complete pedigree numbers might be 28a—32—51a—a and 28a-32-5la—b below m gakk where the initial parents of a series are selected from larger groups requiring designation of or the plants of these larger groups may be given temporary numbers. For example, my Raphanus series 14 began with plants from one lot of commercial seed of one variety, numbered 14, 1; 14, 2; 14, 3; ete. In the next generation, 14, 3 became 14a, and 14, 16 became lb. No. 607] NUMBERING PLANTS IN PEDIGREE CULTURES 433 two particular individual plants, not of two particular branches as with Citrus. Thus with Matthiola a series starts with the crossing, in a given direction, of two apparently uniform seed- reproduced types, while with Citrus a series starts with the crossing of two clons. In another case it might be de- sirable to ignore some evident genetic variation in fea- tures not being studied, and to base the delimitation of the series on genetic differences of seed-reproduced P, individuals in some particular feature under consideration (e. g., in color characters). Fourth, in cases where polyembryony is to be expected nine numbers (1 to 9), instead of one, may be given, so far as needed, to the product of each seed. Thus, with the Citrus cross mentioned above, the first seed will be as- signed the numbers 27aa—11 to 27aa—19, the second seed the numbers 27aa—21 to 27aa—29, ete.” The actual plants from the first three seeds are therefore designated as follows: Seed No. 1 (2 plants), 27aa—11 and 27aa—12. Seed No. 2 (2 plants), 27aa—21 and 27aa-22. Seed No. 3 (1 plant), 27aa—31. If, then, two trees are budded from tree 27aa-31, these will be 27aa—31—a and 27aa—31-b. It may be objected that this treatment of polyembryony gives misleading numbers to plants from adventitious em- bryos. It may, however, be impossible in many cases to differentiate positively these asexually produced progeny before maturity, if at all without progeny tests, and the probable inapplicability of some of the numbers must be kept in mind. Where, as with crosses of Citrus trifoliata with Eucitrus species, the strictly maternal individuals can be separated at an early stage, they can be discarded; or, if such plants are to be kept, and their exact origin in- dicated, letters may be used, thus: 43aa—11 (true hybrid). 43aa—la (for 43aa—12)) (plants from adventitious or 43aa—lb (for 43aa—13) asexually produced embryos). 7 The seeds themselves might be called 27aa—10, 27aa—20, ete. 434 THE AMERICAN NATURALIST (Vou. LI By thus inserting letters in the hybrid number, to desig- nate the second and third plants, we indicate the vegeta- tive origin of these plants, and also their quasi-sib rela- tion to the true hybrid. Open pollination, or lack of protection of flowers in gen- eral, can be indicated in the numbers, if this seems de- sirable, by underscoring the proper letter or number,— the use of two letters with the series number for hybrids making it possible to do this with the pollen parent as well as with the seed parent. An incidental advantage over the original plan results from the use of letters with the series number in the way here indicated. This consists in the fact that the numbers for hybrid and non-hybrid series are thus made sym- metrical in relation to the number of generations involved. As an illustration we may take the case of the Matthiola cross mentioned above. This cross was made in 1914, and the hybrid seeds were planted, together with seeds from selfed parents of the same types, in 1915 and 1916. Corre- sponding to the F, hybrids 23ea—1, etc., we have the F, selfed progeny 19a-1, etc., from a Snowflake parent, and 25c-1, etc., from a Slender parent, while by the original method these three numbers might be 23G1, 19-1-1, and 25-3-1. As will readily be seen, this scheme is elastic; if the five main features stated above are adopted, different workers may add, modify, or omit various details, and still use numbers intelligible to each other with little or no special explanation. Further, the method of designating series and initial individuals is essentially independent of the other features suggested, so that any of the latter features may be adopted without the former. All the schemes so far discussed give cumulative num- bers, which include the whole pedigree from the P, indi- viduals down. Shull (1908, pp. 59-64) has described a non-cumulative system, in which individuals have only temporary numbers (in the field depending on row and position in row) until selected as parents. Each parent No. 607] NUMBERING PLANTS IN PEDIGREE CULTURES 435 is given the number of the notebook page on which its progeny are to be described, preceded by two figures in- dicating the year in which those progeny are grown; e. g., 06230 is a parent whose progeny are grown in 1906 and described on p. 230 of the 1906 notebook. Only parental and grandparental numbers (e. g., 0557.230 for the case just mentioned) are shown at the head of each notebook page, but these numbers permit ready reconstruction of pedigrees from the pages indicated. The labels. bear only the parental number. In Shull’s scheme, then, there is no cumulation beyond the second generation, even in the numbers as written in the notebook. The numbers would include 4 to 8 figures (e. g., 083.7, or 05157.230) in the notebook and 3 to 5 fig- ures (e. g., 097, or 06230) on labels. In the scheme here suggested there is a continuous cumulation; a plant of the second or F, generation is usually represented by 3 to 7 letters and figures (e. g., 5b-8, or 32ba—251), and after from 2 to 4 more generations the numbers begin to be de- cidedly unwieldy. The greater inconvenience and danger of error in copying these larger numbers would seem, however, to be largely offset by the growing familiarity to the worker of the earlier part of a pedigree, and by the identity of the temporary and permanent designations of individuals. Further, the page-to-family feature would be inconvenient in some cases, where several or many actual pages are given to one progeny lot and so would require the same number. With a cumulative system, the simple device of using temporary yearly lot numbers on individual labels will obviate the necessity of writing a long pedigree number for each plant. If the full number is written in the note- book and on each lot label, together with the lot number (e. g., L1 = 16aa—-6-3-44-18-3), and the latter alone on the individual labels if these are used, much work can be saved. In an extreme case, the parents belonging to a given series which are included in a given culture can be 436 THE AMERICAN NATURALIST [ Vou. LI arbitrarily made the initial (P,) individuals of a new series. The system here presented has been used for two or three seasons’ work with the two types of hybridization mentioned above, with Raphanus hybridization beginning with highly heterozygous material, and with pure-line breeding of the tepary bean (Phaseolus). It appears, so far, to be adequate and generally satisfactory for all these types of work. : SUMMARY ` This paper describes a system of pedigree numbering adapted to various types of genetical work with the higher plants. Other similar systems secure greater brevity or clear- ness in certain cases, but are usually of less general ap- plicability ; the main differences relate to the designation of series and their initial individuals. This system provides for polyembryony and somatic variation, and permits of the addition of various other re- finements in cases where they may be needed. The basic scheme is perhaps as simple and convenient as is con- sistent with use for all purposes without change in essen- tial features. In this and similar systems, the ibra are cumu- lative; Shull’s non-cumulative system has both advan- tages and disadvantages in comparison. To summarize the most essential features of the pro- posed method: (1) the series for a given plant (genus or species) are numbered consecutively; (2) the initial ‘‘in- dividuals,’’ as defined in the records, are denoted by small letters affixed to the series numbers; (3) in each following generation the individuals are numbered (if sexually pro- duced) or lettered (if asexually produced), an affixed letter indicating a particular part of an individual; (4) reproduction or propagation is always indicated by a No. 607] NUMBERING PLANTS IN PEDIGREE CULTURES 437 dash; (5) capital letters are employed for miscellaneous special uses. BIBLIOGRAPHY Belling, John. 1914. The mode of inheritance of semisterility in the offspring = certain hybrid plants. Zeitschr. f. indukt. Abstam.- Ve i Hayes, H. K., and East, Edward M 1915. Parther experiments on inheritance in maize. Conn, Agr. Exp. Sta. Bull. 188: 1-31 1. Jennings, Herbert S. 1916. eae Pg variation, and the results of selection in the uni- parental reproduction of Difflugia corona. Genetics, 1: 407- 19 fig. Pearl, Raymond. 1915. A system of Seem types of — in experimental breed- ing operations. Science, n. s., 42: 383-386. Shull, George H. 190 The pedigree culture: its aims and methods. Plant World, 11: 21-28, 55-64. Webber, Herbert J. 1 Ferne ee ae in the Dene breeding work of the Depart Agriculture. In: L, H. Bailey. Plant brooding. re rt i p. MASAA New York 8 Of course, a period may be used, as is done by Jennings (1916) with Difflugia, with a saving of space but possibly a greater liability to error in copying. SYNCHRONISM AND SYNCHRONIC RHYTHM IN THE BEHAVIOR OF CERTAIN CREATURES H. A. ALLARD Wasuineton, D. C. Ir is a matter of common experience to observe in- stances of synchronous behavior and expression among creatures. In such instances an entire group of creatures may react simultaneously to the same external stimulus. A flock of birds will arise from the ground and dash away at the first signal of danger, or a school of fish will swerve as a unit from a stick pushed toward them. In the same way certain frogs in a pool may be started into a brief, explosive chorus of simultaneous croaking by the notes of a single individual. After a brief period during which all have expressed themselves, silence ensues until the next singing-reflex is unlocked by the croaking of another individual. In New England I have heard the wood frog, or so-called clucking frog (Rana sylvatica) give rise to just such outbursts of simultaneous clucking, started either by the frogs themselves or by my giving an imita- tion of their notes during an interval of silence. This habit of singing in concert is not unusual among certain species of frogs, and is mentioned by D. D. Cunningham (1903) in his excellent book, ‘‘Some Indian Friends. and Acquaintances.” He says: Such utterances recur several times in succession; a short pause fol- lows and then the conversation begins again. The curious thing is that all the performers seated in one patch of swamp should have such a tendency to synchronous action that periods of total silence alternate with those of general uproar. The phenomenon is parallel to that of the synchronous mewn: that sometimes occurs so markedly in groups of fireflies In these scala synchronism is aiakohta at inter- vals, but there is no regular rhythmic expression within the group itself. Each frog croaks in its own way, until a perfect babel of noise is produced. Similar synchronous outbursts of sound also occur among the musical insects. The tendency to respond to the notes of their kind is very strong, and a single singer 438 No. 607] RHYTHM IN BEHAVIOR OF CREATURES 439 may be followed by thousands. Although the seventeen- year cicada rarely sings at night, Hopkins’ noted a most remarkable nocturnal concert which began with a single singer. He says: I was fortunate enough to hear the starting of one of these concerts on a clear, moonlight night in June. One male in an apple tree near the house suddenly called out as if disturbed or frightened. Hus neighbors in the same tree were thus apparently awakened. One started the familiar song note, which was at once taken up by numbers of other males, and, like the waves from a pebble dropped into still water, the music rapidly spread until it reached the edge of the thick woods, where it was taken up by thousands of singers, and the concert was in -as full blast as if it had been the previous day. This continued a few minutes, until all had apparently taken part and the song had reached its highest pitch, when it began to gradually subside, and in a short time silence again prevailed. In his book, ‘‘ Bolivia’’ (1914), Paul Walle (p. 268) speaks of the sudden outbursts of noise in the tropical wilderness as follows: As the darkness grows deeper, the silence of the forest is broken uproar lasts for a moment, and all relapses into a silence in which one still hears, more or less sensibly, the murmur of a million insects. In such instances where a group of creatures respond simultaneously to the same initial stimulus, we have the simplest case of synchronic behavior. I once witnessed an instance of a similar synchronic behavior in the move- ments of a colony of plant lice which thickly covered the tip of a twig. While I was watching them, a tiny, para- sitic wasp suddenly approached and hovered over the colony preparatory to attacking them. The plant lice became aware of its presence and in an instant the entire colony raised the hind portion of their bodies simul- taneously into the air at an angle of about 45° and began waving their hind legs about. This behavior was prob- ably more or less a protective response to the primary stimulus afforded by the presence of the wasp. It is also possible that the reaction, once started in a few individ- 1‘‘The Periodical Cicada,’’ by C. L. rage Bull. No. 14, Div. of En- tomology, U. S. Dept. of Agr., 1898, p 440 THE AMERICAN NATURALIST [Vou. LI uals, would tend to be transmitted to the entire colony by the mere contact of their bodies. The same sort of synchronous movement may sometimes be noted in the reactions of the leaves and leaflets of the sensitive plant (Mimosa pudica). A sudden shock may cause every leaf of the plant to react synchronously. Sometimes, when the stimulus is applied only to one or two pairs of the basal leaflets, the closing of these and their contact with those beneath them produce a pro- gressive closing movement in all the pairs of leaflets until the terminal ones are reached. In this instance we have a progressive transmission of stimuli quite similar to the transmission of certain automatic reactions which sometimes take place in groups of insects, birds, or ani- mals. In other instances the reaction to a certain stimulus may involve several similar, instinctive movements producing the simplest form of synchronic rhythm. One interesting instance of this sort of synchronous action has been ad- duced by H. H. Newman, in Science, N. S., for January 12, 1917. Newman found that an enormous colony of ‘‘harvestmen,’’ of the genus Liobunum, resting beneath an overhanging rock, when disturbed began a rhythmic body movement, raising their bodies up and down at a rate of about three times a second, for a brief period. If the stimulus was set up in a few individuals of the colony, the synchronous body movements spread rapidly over the entire colony. After a time it was found that the reactions became weaker and finally ceased. In this instance it ap- pears that there was a secondary transmission of the stimuli from individual to individual by means of their closely interlocked legs. The writer has noted a similar behavior among the individuals of certain caterpillars which were arranged close together on a twig. These caterpillars, when resting, had the habit of keeping the anterior portion of their bodies raised in the air. If the colony was disturbed, each individual began a synchronous swinging of the free portion of the body from side to side, violently, for a brief period. The primary stimulus could affect all the caterpillars of a group, or a secondary trans- No.607] RHYTHM IN BEHAVIOR OF CREATURES 441 mission could take place down the twig. Although the swinging was not always in the same direction, a remark- able degree of synchronic movement was brought about. In those instances where the rhythmic expression of each individual, when set up, continues for long periods of time, as in the case of certain crickets, we have the de- velopment of a more complex synchronic rhythm having its origin in the instinctive habit of response, and prob- ably also built up and maintained by the unconscious in- fluences of the regular sounds of the crickets upon one another. I do not feel the slightest hesitancy in affirming that certain crickets which have the intermittent habit of chirping may build up and maintain a synchronic rhythm under favorable conditions. I refer to the well-known snowy tree cricket (Œ canthus niveus), whose synchronous music has been noted and described by many well-known observers and naturalists in this country. I have many times heard the remarkable synchronous chirping of these crickets in New England, and under exceptionally favor- able conditions the synchronism has been so marked that waves of solemn, rhythmical music have been produced for long periods of time. This rhythmic music was spoken of as a ‘‘slumbrous breathing’’ by Thoreau. Hawthorne called it an ‘‘audible stillness” which, ‘‘if. moonlight could be heard, it would sound like that.’ Bur- roughs called it a ‘‘rhythmic beat.’’ McNeill has said of the music of this cricket: It is heard only at night and occasionally on cloudy days, but in the latter case it is only an isolated song, and never the full chorus of the night song produced by many wings whose vibrations, in exact unison produce that characteristic “ rhythmic beat,” as Burroughs has happily phrased it. Dolbear, in the AMrertcan Naturauist for November, 1897, refers to this cricket when he says: i At night, when great numbers are chirping, the regularity is aston- ishing, for one may hear all the crickets in a field chirping synchro- nously, keeping time as if led by the wand of a conductor. Although the facts may have been slightly overdrawn when Dolbear stated that all the crickets in a field were | 442 THE AMERICAN NATURALIST [Vou. LI chirping in unison, Shull was not justified in concluding that the synchronism observed by Dolbear was merely an illusion. If such were the case, it is surprising that so many other excellent observers have also been misled by the same illusion in reporting the music of the snowy tree cricket. For many years I have made a very close study of the stridulations of insects, yet I have noted this syn- chronous chirping in but one other species of those crickets which possess the intermittent habit of trilling. One may frequently hear great numbers of the common field cricket (Gryllus pennsylvanicus) chirping in the fields, but these crickets never show the least tendency to chirp in unison as do the snowy tree crickets. The jump- ing tree cricket (Orocharis saltator) is also an intermit- tent triller and may sometimes be heard chirping in great numbers in certain copses, yet the ‘‘illusion”’ of syn- chronous trilling one somehow never experiences. Like- wise, there appears to be no tendency whatever for the ground crickets, Miogryllus saussurei, Nemobius fas- ciatus, or Nemobius ambitiosus, to chirp in unison. Many other locusts and katydids, such as Conocephalus ensiger, Conocephalus exiliscanornus, and the common katydid (Cyrtophyllus perspicillatus) produce regular, intermit- tent notes and stridulate in well-defined colonies, yet so far as I have observed the individuals of a species never show the least tendency to stridulate in unison. In the south, however, I have heard the tiny tree ceils (Cyrtoxipha columbiana) chirping in unison with re markable precision, producing waves of shrill, rhythmic sound, as in the case of the snowy tree cricket. In this in- stance great numbers of these crickets were located in the branches of a sarge evergreen holly tree. The shrill notes of this little cricket are delivered with great regularity, as are the low, solemn chirping notes of the snowy tree cricket. This regularity in the delivery of the chivping of these two crickets is especially striking when compared with the musical efforts of other chirping crickets, such as Ecanthus angustipennis, Orocharis saltator; or Gryllus pennsylvanicus. This sustained regularity in the rate of chirping of the snowy tree cricket has been noted by No. 607] RHYTHM IN BEHAVIOR OF CREATURES 443 various observers. On the supposition that the rate of chirping of these crickets is entirely a function of tem- perature, it has even been considered that the crickets may serve as an accurate thermometer. As temperature indicators, however, the crickets can not always be relied upon under all conditions, since wind, humidity, and elec- trical conditions of the atmosphere preceding thunder- storms, also appear to influence the activities of these in- sects. It is now a question as to whether these crickets per- ceive the rhythm which is so pronounced in the regular sequence of their chirpings. I believe they must, for it is quite evident that they hear and respond to the peculiar rhythmical ‘chirpings of their kind, which have become the common language of the species. If they are able to recognize the notes of their kind, it is reasonable to be- lieve that the rhythmic character, as well as pitch, manner of delivery, and even more subtle tonal differences enter into the recognition. The rhythmic chirping in unison which oftentimes be- comes such a pronounced feature in the music of these crickets takes place only in the evening and appears to depend upon the nice adjustment of certain nocturnal atmospheric relations—moonlight, temperature, humid- ity, and stillness of the air. Early in the evening, per- haps, a single cricket begins its stridulations which stimu- late others to respond, and by degrees the great chirping chorus is augmeiited. There may be no noticeable syn- chronism in the chirping at first, but if conditions are favorable, the crickets gradually build up a synchronic rhythm until waves of solemn music are — by a certain colony. There seems to be a marked tendency for the indi- viduals of each colony to adopt the rhythmic beat of their particular colony, so that not infrequently a neighboring colony may establish an antiphonal rhythm, with the re- sult that waves of quavering sound swing backward and forward between two neighboring colonies. How is this synchronal chirping built up and main- tained? It is probable that the instinctive habit of re- 444 THE AMERICAN NATURALIST [Vou. LI sponse tends to bring the chirping of many individuals into a regular synchronism, but I feel, also, that the crickets are somehow unconsciously influenced by one an- other in their chirpings so that they tend gradually to build up a common synchronic rhythm as the night ad- vances. Once a synchronic rhythm is established, I am inclined to feel that it would be natural for those indi- viduals which began chirping, at first asynchronously, to chirp, sooner or later, in unison with their fellows. From the tone of certain discussions which have taken place in Science for the past several months, it is clearly indicated that the synchronous flashing of fireflies as a reality is somewhat doubted. There is no reason to con- sider such synchronism ‘‘contrary to all natural laws,” as Laurent as recently stated in Science for January 12, 1917. In the tropics it appears that great numbers of fireflies frequently establish themselves in the crowns of certain trees, and I am not at all ready to deny that under certain conditions such colonies may not flash synchro- nously, as reported by Cunningham (loc. cit.), Shelford, and others. It is unfortunate that in most instances these reports are accompanied by very meager details as to the general behavior of the fireflies in these arboreal colonies. Cuthbert Collingwood, however, in his book, ‘‘Rambles of a Naturalist,’’ (1868), pp. 254-255, has given a very careful account of his observations of the synchronous flashing of fireflies, as follows: ‘‘ At Singapore, and also at Labuan, the little luminous beetle commonly known as the firefly (Lampyris Sp. Ign.) is common. When flying singly it shines with an intermittent light which alternates with darkness; but on fine evenings and in favorable (i. e., damp and swampy) locations, they present a very re- markable appearance. Clustered in the foliage of the trees, instead of keeping up an irregular twinkle, every individual shines simultaneously at regular intervals, as though by common impulse; so that their light pulsates, as it were, and the tree is for one moment illuminated by a hundred brilliant points and the next is in almost total darkness. ‘The intervals have about the duration of a second, and during the intermission only one or two re- No.607] RHYTHM IN BEHAVIOR OF CREATURES 445 main illuminated. To all appearances they are not on the wing at the time but settled upon the tree; for I was able to recognize certain points of light which I especially no- ticed, and which remained in the same situation with each successive flash. When I disturbed them under such cir- cumstances they flew about at random, each giving out a more rapidly intermittent light. At Labuan, however, I have frequently seen them shine with a steady light as they flew along looking like little, flying stars of the second, or even first, magnitude.’’ If there are periods of repose or inactivity during which the insects cease their flashing, it is easy to see how a part of the col- ony or even all the insects could react simultaneously to the same visual stimulus, such as the sudden flash- ing of a single individual. If it is the habit of the in- sects, before again becoming quiescent, to flash several times in succession, following an appropriate stimulus, it is very easy to see how distinct synchronous flashes of light would now and then illuminate a portion of the tree, or even the entire tree. Such synchronic rhythm in the flashing of fireflies would be similar in every way to those instinctive, automatic body movements observed by Newman in the case of the ‘‘harvesters,’’ Liobunum—the swinging of the pendulum, so to speak, two, three, or more times, as the case may be, following the initial stimulus. As Newman well suggests, it is possible that a transmis- sion of stimuli could even build up and maintain for some time a synchronous flashing in colonies of fireflies in a tree or field. This flashing in unison would parallel the syn- chronous trilling of the snowy tree crickets, and would not necessarily violate known natural laws governing the instinctive synchronic activities of various creatures. By selecting two individuals which were evidently chirping in unison, Shull? attempted to study the reality of synchronous chirping in a colony of snowy tree crickets by statistical methods. No greater statistical fallacy 2 Shull, A. F., ‘‘The Stridulation of the Snowy Tree-cricket (@canthus niveus),’’ Canadian Entomologist, Vol. 39, 1907. 446 THE AMERICAN NATURALIST [Vou. LI could be adduced, however, than to attempt to determine the reality of synchronism in a colony of chirping tree- crickets from statistical results obtained for only two crickets of the colony. That synchronous chirping may be a reality, it is not necessary that 100 per cent. of the individuals of a colony chirp exactly in unison. A few crickets chirping asynchronously would not necessarily prove that a synchronism did not exist. As a matter of fact, some individuals always chirp asynchronously, even when the synchronic rhythm is most pronounced. Taking any two crickets from such a colony, the results would depend entirely upon which two crickets were chosen, and either perfect synchronism or absolutte lack of synchro- nism, would be established. X It is obviously impossible to subject all the individuals of a colony of chirping tree-crickets or a group of flashing fireflies to statistical analysis. Because this is impossible, however, one is not justified in concluding that judgment based upon careful observation is of no value in deter- mining the reality of synchronism. If one hundred men were marching down a street and 75 were in step while 25 were not, judgment alone would establish the fact that a - synchronic rhythm existed. This would be quite as true for the chirping of tree-crickets or the flashing of fireflies. As a matter of fact some of the most marvellous dis- criminations depend upon niceties of judgment alone, and no amount of statistical data would simplify the matter. I can not yet agree with those who are inclined to be- lieve that the snowy tree-crickets never chirp synchro- nously, and that it is impossible for fireflies to flash syn- chronously, especially certain tropical species of fireflies. I will agree, however, that in practically all instances of synchronic rhythm there seems to be no evidence of con- scious, intentional imitation, but merely instinctive, reflex, or automatic reactions to certain stimuli, similar in many respects to the unconscious reactions of the leaves and leaflets of the sensitive plant. SHORTER ARTICLES AND DISCUSSION SOLID MEDIA FOR REARING DROSOPHILA BAUMBERGER and Glaser (1) recently described a method of raising the banana fly on transparent solid media, thus enabling the investigator to observe more accurately the rate of growth and metamorphosis and the larval habits of this insect. The medium was made as follows: Five or six bananas were mashed up in 500 c.c. of water. This was allowed to infuse on ice over night, after which the liquid was passed through cheese cloth. Powdered agar-agar was then added in the pro- portion of 144 gm. to 100 c.c. of the banana infusion. This was then heated until the agar had dissolved. The liquid was then filtered through a thin layer of absorbent cotton into test tubes. The tubes were then plugged, sterilized and slanted in the eustomary manner. As pointed out in the above article, one of us (2) had observed that the bacterial growths which always develop on this medium ‘‘do not Reem to harm the larve’’ and the mold which sometimes appears ‘‘is usually destroyed by the larve just as soon as they hatch.’’ This question was further investigated and it was found (2).that the prinicpal food of Drosophila is yeast and the flies can not develop on banana which is kept free from micro- organisms. Delcourt and Guyenot (3) had previously (un- known to the author) published similar conclusions and Loeb and Northrop (4) had confirmed them shortly before the author’s report (2) was sent to press. It is therefore very well known that the food of Drosophila is yeast and the prime necessity of any medium for rearing this fly must be either abundant food for yeasts to grow upon or the presence of large numbers of yeast cells. If a medium is made of sterile compressed yeast and agar-agar? it serves as a perfect food for flies which have been freed from microorganisms; how- ever, if living yeasts develop young larve are usually killed. As 1 Loeb and Northrop (4) raised a few flies on aseptic banana, but all flies were sexually sterile. kniee (5) also succeeded | in raising a small percentage of flies on steri 2 An excellent sichikiieiin ican eonsists of yeast cake moistened with water. 447 448 THE AMERICAN NATURALIST [Vou. LI adult flies usually carry living yeast cells upon them this medium would be difficult to use for work in heredity. A nutrient medium for yeast would best suit the needs of geneticists. Into such a food the adults or pupæ would carry living yeast cells which would ferment the sugars and produce odorous substances which cause oviposition by the female fly. Larve on hatching would spread the yeasts throughout the medium, thus increasing growth and alcoholic fermentation which may prevent the de- velopment of injurious microorganisms. wo media might be suggested for this purpose, viz.: Fer- mented banana agar and Pasteur’s nutrient agar. The former may be prepared by adding two cakes of bread yeast separated in 100 c.c. of water to one dozen mashed bananas, mixing thor- oughly and allowing to ferment for twenty-four hours. This material should be pressed through a sugar sack and the liquid resulting thickened by heating with 1.5 gm. agar-agar per 100 c.c. and poured into slant culture tubes, plugged and sterilized. A perfect nutrient medium for yeast consists of 10 gm. am- monium tartrate, 10 gm. ashes of yeast, 100 gm. rock candy, 1,000 gm. distilled water. This is called Pasteur’s culture fluid and can be treated in the same manner as the above but should be sterilized in an Arnold sterilizer for three successive days, as it can not be heated to as high a temperature as the banana with- out preventing jellation. Pasteur’s nutrient solution can be ap- proximated by diluting molasses with three parts of water. In the two media described above yeasts develop rapidly and furnish abundant food for Drosophila larvæ and also produce odors of fermentation which cause the female to oviposit readily. A concentrated food for yeast such as banana flour would prob- ably increase the value of the first medium as this substance is now used for raising yeast in the brewing industry (6). ‘REFERENCES 1, Baumberger, J. P., and Glaser, R. W. Science, N. me XLV, pp. 21-22. 2. Baumberger, J. P. Proc. Nat. Ac. Sc., III, pp. 122- 3. Deleourt and Guyénot, Bull. Se. France et ay Pa 45, pp. 249-332. 4. Loeb, J., and Northrop, J. H. Jr. Biol. Chem., XXVII, pp. 309-312, 5. Guyénot, E. Compt. Rend. Soc. de Biol., 1913, LXV, pt. 1, p. 178. 6. Nagel, Z. Spiritusind., 35, p. 185 J. P. BAUMBERGER Bussey INSTITUTION. THE AMERICAN NATURALIST VoL. LI. August, 1917 No. 608 BIOCHARACTERS AS SEPARABLE UNITS OF ORGANIC STRUCTURE HENRY FAIRFIELD OSBORN INTRODUCTION As to the term ‘‘character,’’ in the commonly accepted sense of zoology and botany, it has long been known that different characters, large and small, exhibit different rates of evolution and are in this sense ‘‘separable’’ or ‘‘independent,”” while in another sense it has long been known that every character of an organism is correlated or coordinated with every other character in function and adaptation. Our general knowledge of character separa- bility (for the want of a better English term) has been more than confirmed by researches based upon the great discovery of Mendel, namely, that many large as well as minute characters which are closely associated or even blended in the adult organism are very sharply separated from each other in the germ plasm, in such a manner that they may entirely appear or disappear in the crossing, or hybridization of species, subspecies, varieties, races, no less than of individuals. The theoretic separability in heredity of the germinal ‘‘determiners’’ or ‘‘factors’’ of characters is in full accord with the several aspects of separability in evolution previously observed in paleon- tology, embryology and individual development. The purpose of this synopsis is to review and bring together some of the noteworthy phenomena of character separability as contrasted with those of interdependence, cooperation, correlation and coordination. 449 450 THE AMERICAN NATURALIST [Vou. LI It is probable that this principle of separability of units of function and structure, which has been discovered in so many processes of the organism, will prove to be a universal principle. This principle is more or less a necessary consequence of the unit-cellular structure of the organism; if instead of being composed of cell units, each possessing its peculiar heredity qualities, an organ- ism were composed of fluid or solid masses of different kinds, we can not imagine how it could split up into char- acter units. Yet the cell is probably not the ultimate, or least character unit, which is doubtless a chemicophysical unit. This is extremely suggestive in connection with our conceptions of the nature of the evolution process, for the process more or less clearly resolves itself into three problems, namely: (1) how do these character units arise; (2) how are these character units coordinated into harmo- nious action within the organism; (3) why do these char- acter units independently evolve, progress or retrogress? Since the word ‘‘character’’ is very vague, and since the term ‘‘unit character’? has a special and limited meaning in Mendelian heredity, I propose the new term ‘*hiocharacter’’ as a general designation of the character unit in the organism. | 1. Kryps or BIOCHARACTERS OBSERVED IN PALEONTOLOGY Biocharacters are those characters, large and small, which through the evidence afforded by ontogeny (t. e., zoology, embryology), phylogeny (7. e., paleontology), or heredity (Mendelian inheritance )are found to be separable from or independent of each other as units in the proc- esses of heredity, of evolution, and of individual de- velopment. Paleontology gives evidence no less positive than em- bryology and experimental heredity that each organism is made up of a number of such separable characters, includ- ing those known as the ‘‘unit characters’’ of Mendelian heredity. In paleontology this separability in heredity and evolution is found to apply to structural, or anatom- ical, characters of three principal kinds, namely, I, II, ITI: No. 608] BIOCHARACTERS AND ORGANIC STRUCTURE 451 ‘siojouanya zo sdnois pue siojovinyo Suysa ur Åpsuəgur pue uotjzodoid zo sasuevyo əarndepe ‘yoriq ut í Ajs ypepsyoriq ‘Ajsprpoprop ‘Areqdəs ‘kyeqdookyoesq ‘Areqdəsoyərop mys 30 uon jo suonaodoad pue ozts ur SUOTIULIE A sajdupxy *“wsynau to aajdnpp ‘mbro ur snonury, -u09 paasasgo sp wwf os ‘fijvpasay w Burpuajq sawayawos ‘a)qnvundas hywsauay 'SLIJIDADYI aarsuazw PUD aayopyupnd ‘(usogsQ fo suospaw -0110) saajopanyo pouoywodosd MAN “TIT ‘sqyaud Surjsrxe Jo uorpnururp 10 WOTpRaTdypnU əy} ynog worypnpoag “UMIE jo suonenea IOUT ‘aIngonays Jo suoregyjes eynutme (g) fsedunpeqd peppe ‘y3003 peppe ‘sqrı peppe ‘eiqeyi0A pəppe Áq ‘urou (z) £939 ‘sqr ‘waqas “q499} ‘Syp Arvsoumusdns ‘“quasaad Apear syed jo uorjeorpdnper “9 ~ QeotsojozRie} 10 ‘Ruouqy (1) ¿Spuy 3914} JO ƏB SIƏPPLIVYI Yong saduvri 'UDLULO UL SNONWNUOISIP P480. sp wf os ‘fjyypasay w apqvandas fijajajdwoo ‘aaydvpv-wow 40 am} DSI aouasaad SLIJIDLDYI JRLUNU MIN “TT ‘sas1oy JO 4400} Ə} UL JueTRAerd e10ml ynu puge ‘sa1oyjOURITy UL IBI APAYL[OL 1G SUOTZBpvIS 001 Yong “Ayyiquandes jeuruo3 ‘suoreanyd pue sploy [ejuep mou urepiag “‘Aqyrqeavdos peutuses ‘sdsno pejuep MƏN “AyyIqeavdes [wuruies ‘suro MƏN salur "wbo Ur snonwyu0o paasasgo sv sof os ‘fipypasoy w arqvandas fiyajzajd -W09 ‘}40}Ss IY} wosf UNPINP fiupuoynjora aaydypp up Buany suajonapya *(usogsQ fo suovyoposb -yoas) ssaqgovapyo oyauahoyji10 MIN ʻI , divisibility, independ- lity i The property of separab ence of these biocharacters stands in contrast to that of 452 THE AMERICAN NATURALIST [Vou. LI their interdependence and coordination of function in the organism. As remarked above the discovery of this property of separability of characters was gradually prepared for in various branches of biology. First, in comparative anatomy, through the observations of the laws of independent degeneration, balance and progres- sion of adjoining parts; second, through embryogeny and ontogeny, in the observation of the hastened or retarded development of adjoining parts; third, through heredity, as, for example, in the mosaic inheritance of Galton; fourth, through paleontology and phylogeny, in the law of acceleration and retardation of Hyatt and, through the observations of many other paleontologists, concluding with the detailed observations of Osborn and Gregory. But the property of separability first shone out clearly and most brilliantly in the discoveries of Mendel in hered- ity. This property of structural separability and inde- pendence of evolution is closely connected with separa- bility or independence of function. Thus separability may distinguish a great number of cells and tissues which are united by a single function, or separability may dis- tinguish a single cell, as, for example, the single giant cells of the spinal cord of certain fishes and amphibians. It is important, therefore, to realize that this property of separability, the separability of biocharacters, is now observable under six classes of evidence. 2. MODES or SEPARABILITY OF BIOCHARACTERS IN HEREDITY, In GENESIS, IN Rate or EVOLUTION I. Biocharacter heritage separability, unit or exclusive inheritance. 1. Evidence of Galton, etc., as to mosaic inheri- ance. ; 2. Character heritage separability, under the sepa- rate ‘‘determiner” or ‘‘factor’’ hypothesis of Mendel, Bateson, Morgan, etc., according to ` which each ‘‘biocharacter’’ of the soma may spring from one or more ‘‘determiners’’ or No. 608] 3. BIOCHARACTERS AND ORGANIC STRUCTURE 4538 ‘‘oroups of determiners’’ in the germ. Many characters are not entirely separable, but tend to go in groups or to blend. Character heritage blending, Castle, Osborn, etc., as observed in color, size, proportion, ete. II. Separability in genesis, biocharacter origin and evo- ar bo Oo (P op > lution. . Origin through saltation, large heritable leaps (saltus, a leap, syn. discontinuity, Bateson). . Origin through minute heritable gradations (gradus, a step), Darwin (minute heritable variations), De Vries (mutations). . Origin through continuity (continuatio, an un- broken series; continuous describes that which is absolutely without pause or break, in con- trast to saltation—leaps, gradation = steps). It has been shown (Osborn) that characters continuous in origin may be separable or dis- continuous in heredity, as in both the new rectigradations and new proportion characters of the hybrid of the horse and the ass. Continuous rectigradations, Osborn, 1. e., ortho- genetic or adaptive origin followed by adaptive evolution in a single direction, a principle probably corresponding with the Mutations- richtung of Neumayr. Homomorphic (Fiirbringer) rectigradations, through independent origins, the independent production of similar biocharacters in organ- isms of similar ancestral affinity, the ‘‘poten- tial homology” of Osborn, as distinguished from the true homogeny, homology sensu strictu, of direct descent of characters from each other. Mutations of Waagen, subspecific steps or grada- tions in evolution of characters under the mutationsrichtung, or trend of development i in certain directions, of subspecific value, i. e., intergradations between species. 454 THE AMERICAN NATURALIST ~ [Voi LI II. Separability in velocity, biocharacter motion (i. e., rate of transformation, or relative displace- ment). 1. Ontogenetic acceleration, hastening of develop- ment in ontogeny, resulting in certain biochar- acters appearing at earlier and earlier stages of development. Hyatt’s law. . Ontogenetic retardation, slowing down of bio- characters in ontogeny, resulting in certain biocharacters appearing at later and later stages of development. Hyatt’s law. 1 and, result in heterochrony (Gegenbaur). Ontogenetic balance, resulting in biocharacters appearing at the same stage as in the ances- tral forms. . Phylogenetic acceleration, appearance of certain homomorphie biocharacters at earlier geologic periods in some phyla than in others. . Phylogenetic retardation, whereby some homo- morphic biocharacters appear at later geologic periods in some phyla than others. . Phylogenetic balance. bo 9a N or jor) IV. Separability in grouping, biocharacter cooperation and coordination. A biocharaecter may be closely linked with others in group function or it may detach itself — and connect with another group, e. g., color and speed biocharacters in horses. 1. Biocharacter correlation into similar adaptive character groups. 2. Biocharacter grouping through heredity, attrib- utable to prolonged ancestral grouping. 3. Biocharacter compensation, the gain of one char- acter at the expense of another. The law of compensation of St. Hilaire. 4. Biocharacter sex linkage, the union of groups of characters with the function of sex as family or secondary sexual characters. : No. 608] BIOCHARACTERS AND ORGANIC STRUCTURE 455 V. Separability in proportion characters. Character proportion genesis, separability in the evolu- tion and heritage of proportions. 1. Proportion genesis through fluctuation, theo- retically through continued selection of plus and minus variations in a given direction in eases where such variations of proportion are of adaptive significance, as in fluctuating variations of length of neck in the giraffe. 2. Character proportion continuity in zoologic and paleontologic series, the unbroken succession in the evolution of changes of proportion, as in brachycephaly and dolichocephaly in kindred races of man. 3. Character proportion saltation, in E only, sudden appearance of profound changes of proportion, as in the Moùchamp breed of sheep. 4. Chace proportion heritage separability. Complete separation in heredity of certain proportional characters, as in crosses between dolichocephalic and brachycephalic types in man. 5. Character proportion heritage blending, the origin of blended, or intermediate, forms of certain proportional characters in man and horses. 6. Character proportion genesis, partly at least through interaction of separably specific harmones, chalones, internal messengers of the thyroid and other internal secreting glands. VI. Separability in orthogenesis, rectigradations, gene- sis of new adaptive characters from infini- tesimal beginnings. 1. Independent genesis of similar adaptive charac- ters in different phyla giving rise to homo- morphy. 456 THE AMERICAN NATURALIST [Vor LI 2. Evidence of hereditary predetermination or po- tentiality of the genesis of similar new char- acters in phyla derived from similar ancestors. 3. Continuous evolution of rectigradations m one direction, Mutationsrichtung. 4, Mutations of Waagen, subspecific gradations of character. 5. Complete separability of rectigradations im heredity, as shown in the teeth of the hybrids of the horse and ass, all rectigradations being either present or absent but not blended. 3. SUMMARY It appears from the foregoing classes of evidence that biocharacters are separable in origin,’development, evo- lution, and heredity. -First, biocharacters are separable through their many different modes of origin from the germ, either saltatory, gradational, or continuous. Sec- ond, biocharacters have different rates of motion, or velocity, in individual development (ontogeny), exhibit- ing acceleration or retardation. Third, biocharacters have different rates of evolution in different phyla (phylogeny), again exhibiting acceleration or retardation (phyla). Fourth, all the biocharacters of an organism cooperate through various modes of grouping in func- tional correlation, in compensation, in sex linkage. Fifth, in the hard parts of the body while the biocharacters of form and proportion may originate through continuity, through saltation, or through minute gradations, all the known evolution of proportion biocharacters is contin- uous. Sixth, in the hard parts the biocharacters of recti- gradations have only been observed to originate and de- velop through continuity. EVIDENCE OF MULTIPLE FACTORS IN MICE AND RATS C. C. LITTLE HARVARD MEDICAL SCHOOL Tae object of this paper is to record certain data on the inheritance of two complex characters and analyze these data together with those obtained in certain analogous experiments by other investigators. This is done with a view to ascertaining what they contribute to our knowl- edge of the relative merits of the two more or less con- tradictory hypotheses, multiple segregating factors or a single fluctuating factor which are being advocated by geneticists to explain certain cases of inheritance. It will be useful at the outset to state in a somewhat definite manner how the alternative hypotheses differ from one another. MacDowell (1916) in a recent paper has clearly and precisely defined the two views. From this start the following statement may be made. The first view supposes that variations of the germ plasm are in the nature of fluctuations. The germ plasm is in a continuous state of variation. The hereditary characters all vary under observation and this is taken to mean that the fac- tors in the germ plasm determining them also vary. Fluctuation in the character is measured and used as a means of detecting and recording a similar though not identical variation in the germinal factor underlying and determining the character under observation. To use a concrete example. In a given species the individuals are of various sizes, some larger, some smaller. Some of this variation is considered as in part due to non-heritable environmental influences. There is, however, a distinct ‘*inheritance’’ of size. This is considered the result of a variable germinal factor which as Castle suggests may be ‘‘Perhaps some substance or ferment which varies in 457 458 © THE AMERICAN NATURALIST [Vou. LI amount, larger amounts producing larger results’’ (1916, p. 55). In crosses between animals possessing distinctly different degrees of the size ‘‘factor’’ blending rather than alternative inheritance is supposed to result. This also holds in the case of crosses between races differing only slightly in the degree of the size character. In such cases says Castle ‘‘we do not find it easy to detect segre- gation’’ (1916, p. 55). The ‘‘second’’ view supposes that complex physiolog- ical results such as the ‘‘size’’ of a certain animal depend upon a complex of genetic factors. All these factors are concerned in the growth and therefore the size of the animal. The germinal determiners of these many factors are considered to be units.? As such their behavior in inheritance is definite, non-blending, and essentially Men- delian in nature. The blending, fluctuating nature of the character studied is supposed to be due to environmental factors, to the large number of hereditary factors in- volved, and to absence or incompleteness of dominance. With this rather incomplete statement we may turn to an examination of the experimental data. I. THE INHERITANCE oF SPOTTING In MICE Experiments on the inheritance of spotting in mice have been carried on by the writer since 1909. As they pro- gressed it became evident that spotting in mice offered remarkable material for the investigation of the inheri- tance of a variable character. Spotting as a character is easily measurable and classifiable. It is not affected by changes in the external environment. Variations in 1It is interesting to note in this connection, the quotation from Castle given by MacDowell, 1916, page 741, describing the inheritance of minute quantitative differences in intensity of yellow pigmentation in guinea-pigs. The later work of Wright, 1915, also has a distinct bearing on this par- ticular ease in that it shows four truly allel elomorphie grades of intensity of pigmentation in guinea-pigs, Fluctuation occurs about these four variation aim and Castle’s ‘‘complete series of intermediates’’ between red and in guinea-pigs is broken into three distinct centers of variation with overlapping range. 2 East, 1912, has made clear the fact that variation within the unit, under i certain sondia must of course be supposed to be possible. No. 608] MULTIPLE FACTORS IN MICE AND RATS . 459 degree of spotting are almost certain to be quickly recog- nized. In mice a so-called recessive ‘‘piebald’’ spotting has long been known. In addition a distinctly different hereditary type, commonly dominant, was recorded by Miss Durham (1911), and was further investigated by the writer who found it to be entirely independent in inheri- tance from the ‘‘piebald’’ type. This ‘‘dominant’’ spot- ting I have called the ‘‘black-eyed white’ type of spot- ting. Further in 1914 a third type of spotting known as ‘‘blaze,”” consisting primarily of a white forehead spot, was reported on by the writer. This type has since been found to be independent of the other two in inheritance. The existence of at least three genetically distinct types of spotting in mice is therefore proven. It is not the purpose of this paper to generalize from this fact, but- to attempt to show that various grades of a single type of spotting may be introduced into a cross and may re- appear in the F, or back-cross generation, not rarely, but in a considerable proportion of the animals obtained. ` This is a common breeding test of segregation as com- pared with contamination and of alternative as compared with blending inheritance. The races of mice used offer for the most part extremely homogeneous material. The experimental animals come from closely inbred races which have undoubtedly reached a degree of genetic homogeneity which would lead to defi- nite results in breeding, and a clearer opportunity to observe segregation if it existed. The degree of white spotting in the races of piebald mice used is estimated by recording the approximate per cent. of the dorsal sur- face which is pigmented. This method is subject, as is any system of arbitrary grades, to a certain amount of error, It is, however, reasonably accurate and affords a fair measure of the degree of pigmentation of any individual. The first experiment to be recorded is a cross involving an English piebald race descended from black-eyed white mice (Little, 1915). These piebalds vary in the extent of dorsal pigmentation from 20 to 96 per cent. It 460. THE AMERICAN NATURALIST (Vou. LI will be seen from Table I that there are essentially two modes to the variation curve, one of these occurring at 30 to 44 per cent., and the other at from 80 to 92 per çent. Animals from this race were crossed with dilute brown self mice, and with mice from a yellow and from a mixed black-agouti and black race. The F, generation consisted of solid colored young. When these F, self young were erossed back with animals from the piebald race, a range of variation was obtained which is shown in the second line of Table I TABLE I PERCENTAGE OF DORSAL PIGMENTATION alelelaigiaisieisiaisielelsisisielaisigigis HAE ool Bl bl cl ul al ol bl el ul bl dl esl ot) Al bl ol djs d| ASR A R S 8] HS) S$) 8) S e| S/F) Ke) | S S „Piebald Eng- | Daal | qa | | | lish stock. . —|—|—i— 2 | 2 |11127/18|26|18|.9 |14| 6 |10| 4 112| 2 |7 17 13 16/15) 1 Back-cross | TAR | Lia pieb: 1} 1) 6|13} 6| 7]10! 2111) 2 | 6113/5 6101015 It will be seen that even though the total of the young obtained is less than in the pure piebald race, its range of variability is essentially the same and there are only four animals among those raised from the pure piebald race which fall outside of the range of variation of the back-eross generation. There is, moreover, no evidence of a single mode in the curve of the back-cross generation, but the modal centers of the parent piebald race are each represented by a large number of young in the back-cross generation. Another cross in which larger numbers have been . recorded is one between Japanese spotted mice and two races of self pigmented mice raised at the Bussey Institu- tion. About two years before the self mice were crossed with the Japanese, certain spotted mice of common ances- try with them were separated off as a different stock. The range of variation in the degree of spotting within this last named race may be considered a fair indi- eation of the probable appearance of the self animals used if they had been spotted. It will be seen from No. 608] MULTIPLE FACTORS IN MICE AND RATS 461 the first row of Table II that, in this race, animals varying from 56 to 96 per cent. were obtained. There is a tendency for a mode to be formed between 80 and 92 per cent. The Japanese race used is one which for sev- eral years has beenclosely inbred and which holds remark- ably true to type. Its range of variation is shown in the second row of Table II. It will be seen that the animals possess from 4 to 36 per cent. of the dorsal surface pig- mented. A distinct mode is observed between 13 and 16 per cent. TABLE II PERCENTAGE OF DORSAL PIGMENTATION ae lol «i | lxo] Eef aa +o) 2/3] 84) 9 8/8) 9] 3] 8/8) 312 3/2/88) 3 5 3) e) 2 Jed SIS aR eS a s3S sees ssl ks 2 8's Pe ra preie | | ——— Tame race... —— — 1} 1/1 6 105/131110 aa 9 Japanese ea EIRA | bo | spotted... -|1 -8 |74|92/65 23| 7| 4) -2 — l F: spotted i—i ——i 3} 2) 5 6| 4110/6] 5913 1u 5 6 4 ia 10 8 ra 18 Back-cross | orrie | en. 1....}—/—| 3| 2) 4 4/14/14/14/16| 5 13| 8 |7 41 6 al} 2) il 2 sle Back-eross | hetera | | aa | spotted | ae Wt Ad | | SHIN ihar ay telat obia aia | a aS The F, generation obtained from crossing the Japanese race with the self dilute brown or brown agouti animals above referred to, consists of solid colored animals and is therefore not to be recorded. The F, generation in which 146 piebald animals have been recorded con- tains mice ranging from 20 to 96 per cent. pigmented. Among the 146 animals recorded, 20 or 13.6 per cent. fall in grades characteristic of the Japanese grandparents. 94 or 63.6 per cent. are of grades found in spotted animals of the other grandparental race, see Table II, line 3. Two other generations of great interest have been tabu- ‘lated. The first of these, line 4, Table TI, is the result of back crossing F, generation animals with animals of the Japanese parent race. In this generation there is distinct evidence of segregation. Fifty-five of the 131 animals recorded fall within the limits of the Japanese grand- 408 . THE AMERICAN NATURALIST [Vor LI parental race and 35 within the, variation limits of the other grandparental spotted ancestor, and of these an appreciable number occur at the upper limit of variabil- ity, there being no evidence of a tapering of the curve at this point. This fact would appear to be of marked significance. When first generation back-cross animals are bred inter se, a second back-cross generation is pro- duced, which, though it includes only 38 animals, has given extremely interesting results, see line 5, Table Il. There are five animals which are distinctly Japanese segre- gates, and of these three show a degree of pigmentation which would make them easily mistaken for even the extreme variants in the Japanese race. At the other end of the curve it is interesting to note that six of the thirteen young reproducing the condition of the non- Japanese grandparental stock, fall into the two extreme upper classes and may be considered as true segregates rather than due to any chance occurrence of an abnormal physiological condition. A third cross involving spotted mice has been made. This is between animals showing a small white forehead 1 2 3 3 4 Fic. 1 : spot and a self race. The dilute brown self race used for one parent is the same that has been already recorded above in the second experiment with mice. The range of . variation of white spotting in the blaze or forehead spot race used, is recorded in line 1 of Table III. The grades of spotting designated in Fig. 1, numbers 1 to 4, repre- sent the increasing degrees of white spotting. The ani- mals comprising this spotted race are all of them pure No. 608] MULTIPLE FACTORS IN MICE AND RATS 463 wild mice descended from a few individuals captured at Forest Hills, and at Wenham, Mass. The F, generation obtained by crossing the ‘‘blaze’’ and dilute brown self races is all self in character. The F, generation shows a distribution of young recorded in line 2, Table III. TABLE III GRADE OF SPOTTING Pure blaze animala ...........,:¢.90-7 | 21) \ fF eel 6 | 3 2 Me pioke animale... rt ok. ed. es Pete LBs 31 | | 3 1 F:, Fs, Fs blaze animals ........:...... 25 M36 | 24 (107) 25 | 5 (a ee It will be seen that the extremes of variability in F, are the same as in the pure inbred wild blaze race. F, Fy, and F, animals raised without selection give a result recorded in line 3 of Table I, and again show the same limits of variation. It will be seen, therefore, that the blaze character is segregated from the cross without apparent modification, although the extremely minute degrees of spotting which it includes are those which one might expect would be modified or entirely swamped if contamination between the gametes of the self and blaze races occurred. This cross together with the cross be- tween mutant and wild rats recorded by Castle to be con- sidered later, give conclusive evidence that even a minute quantitative character segregates after crossing and does not afford grounds for the objection raised by Castle that crosses between races differing slightly in size or like fluctuating characters do not readily show segregation. To sum up the results of spotting inheritance in mice, it may be said that all the crosses made show a reappear- ance of grandparental conditions in F,, back-cross, and other advanced hybrid generations. The reappearance of these grandparental types is frequent enough to lead one to conclude that if segregation in a strictly Mendelian sense is not taking place, that at least the outward appear- ances of such a process are all of them present. 464 THE AMERICAN NATURALIST [Vou. LI Il. THE INHERITANCE OF SPOTTING IN Rats Castle’s work on the inheritance of fluctuations in the hooded coat pattern of rats is well known to all geneti- cists, and may now be considered in an attempt to examine the bearing of spotting on the inheritance of fluctuating characters in general. The hooded rats with which -34 -1⁄4 O m + jj In grades higher than +4 the whole dorsal surface is pigmented and the ven- tral surface is becoming progressively more so. FG. 2 Castle worked have been shown clearly by Doncaster, Mudge, and others, as well as by Castle himself, to be recessive to self or solid colored coat in inheritance. The hooded pattern is, however, subject to wide fluctuation producing a series of rats from those with the whole dorsal surface and most of the ventral surface pigmented, No. 608] MULTIPLE FACTORS IN MICE AND RATS 465 to rats in which only small spots around the eyes remain pigmented. The following diagram will show roughly the limits of variation and some of the intermediate types as graded by Castle. Throughout the course of his selec- tion experiments, Castle has at times crossed the plus variants with the minus variants or else crossed selected animals from the plus or minus series with wild or with Irish rats. Both the last mentioned varieties are essen- tially self pigmented animals and are dominant in inheri- tance to the hooded pattern. The race with which Castle started his selection experiments showed within the first two generations of selection rats varying from grade minus two to grade plus three and three quarters. Castle and Phillips, 1914, have recorded the results of two crosses between minus variants and plus variants. The first of these crosses was between females of minus two grade in the sixth selection generation and males of plus three and one half. or three and three quarters grades in the fifth selection generation. It is possible to use the range of variation among the progeny of animals of these grades within the selection generation from which they are chosen to control the results of the cross. In-Table IV? the first line represents the range of varia- tion in the progeny of minus two animals of the sixth selection generation. The second line represents the range of variation in the progeny of plus three and one half and three and three quarters animals of the fifth selection generation. It will be seen that there is a space of seven grades between the two limits of variation of the parent races. The cross between these two types produced in the F, generation a range of variation shown in the third line of Table IV. It will be seen that for the most part, the young fall in the seven classes between the two parent races. In these seven classes will be found sixty-eight or 73.1 per cent. of the ninety-three young. Two or 2.27 per cent. of the young show grades of pig- 3 The references to tables and pages appearing in the first column of Table IV, and ensuing tables, refer to Castle and Phillips, 1914. [Vou. LI THE AMERICAN NATURALIST 466 bie Sha aE bea white E 19 “d ‘og a1qRL ‘eq —|—|—|—|—|—|T/2elti2}2i¢|el¢]9l6ltsisiei rie =|= |== mlm IlI aaa A a ee ee ea es mers me cca bc cae we us A gt TS “d ‘0g QUL “A T |8 |9pipzt golet 6 t| t| — —|—|—|— ` "Og 'd ‘IT pue oT serquy, ‘881 tg+ —|—|—|— |} —}— | — | — | — — — | — | -— | — T/9/IT|G| S|) 1)/ So) T| 1rd ‘9g aqe ‘syer $z— ‘EZ— iP |s te | fe te} e¢ | te] $e |e! ziijn] EL E/E) OE) E) E/T EC /EL ET) z jte fe lee e-te te | uopo A WiIdViL : —|—|-—|-—|— Pac s L as ‘sg A $ FZ | FZ or] 6r og 92| +g zziz 16 IF T |—j—|—li—l ee “Ig d ‘Og oqe L “a oe | S| ee IT/8 |g |gz;etior)s |t |z |—jze |—|—|—}—| |e d ‘og qer ‘ty 9 |0g | €9| FE} 29 “a SI ¢ II z i —|—|—}|—|/—|-—|-|— —|—|]—}|—|—|—| — | —|— — g d'o orqoq, erea bet ‘te+ —|—|—|—|—|—|— |) —| —] — |! — |---| I I I 8 | FI] If] 09 l6TLis9e Fee! cz; £ | —| — “Tp ‘d ‘Zz oqe L ‘sie g— + | te | fe | | Jt | fe | te elujfrlir|r a] tft] o il t tia |r| 1] g | tz! te | te je — uopuuop AT W&Tavi No. 608] MULTIPLE FACTORS IN MICE AND RATS 467 mentation characteristic of the plus parent and twenty- three or 24.75 per cent. of the young show grades char- acteristic of the minus parent race. When the F, young are bred inter se there is a marked increase in the per cent. of young showing grades of pigmentation character- istic of the grandparents. Ninety-three, or 30.5 per cent., of the 305 F, generation young fall in grades character- istic of the minus grandparents. Fifty-two, or 17.0 per cent., fall in grades characteristic of the plus grand- parents. The remainder, or 52.5 per cent., fall in the seven intermediate classes. A second cross made between selected animals is shown in Table V. Females of grade plus three and three quarters of the tenth selection generation were crossed’ with a male of grade minus three and one quarter of the tenth selection generation. As a control for the females it is possible to use the progeny of grade three and three quarter parents in the tenth and eleventh generations. The reason for using two generations instead of one is in order to increase the number of progeny available, since the tenth generation alone shows such a small number of young that they are not valuable as a significant breeding test. The range of variation in the progeny of rats of this grade is shown in the second line of Table V. As controls for the minus parent, rats of grade minus two and three quarters and two and seven eighths of the tenth selection generation have been used. Although the total of the young produced by them is very small (thirty-one), it is, nevertheless, the only critical data available. The range of variation of these young is shown in the first line of the table. The F, generation resulting from the cross of these two diverse selection types is shown in the third line of Table V. The thirteen young comprising this generation all fall in grades between the two parent types. The F, generation shows a distinctly greater varia- bility than the F, generation and includes four young of grades characteristic of the plus grandparental race. This cross is less conclusive than the previous cross, but [ Vou. LI THE AMERICAN NATURALIST 468 —|—|—|—|1 IT eye te? es e Sree BED os He as Aes “gpd ‘ep aqua SA y | 6 | 62) Se| 6I octet Z —!—|/T j—|— |) llle ed et ed es ee ic "EE a VOL ee ES (OS 1a ie tet "a £I i o 9/2) 9T| Z| OT) F |e |F |Z |: can: Sac) Sot eT Ae eS ghd ‘SF QUL A —|—|— 1 |—|I g —}— |) — | — >ii et et el ic cc ge E iiaa in he ee ely ee es Lj/9 | FIP 1S PS Liao 9 ja ee ea “gpd ‘ep aiquy “A —|—}—|—|—|}—|—}—|}-—|-|-—}— —|—|/—|—1/9 |6 | Ig] 92 Seristsigts) 92| 61) 2 l'IP ‘4 ‘pg pur gz soq ‘syuereg Z— a je | a re [ae tt et el a ee ee “Sb dep QUL “IA —|}—|—|—|-—-|-—|—|-—}-|- 1 - I - —~| —| —| =} | & | Tt ee oe ite [ot 1: ‘OF ‘d ‘Og OQU, ‘syuored FT — aS fel Shee egg | cae) Ol ee ee EET E SF “dep Ode “A PITS pE Ges ey, Onin pala y Take ea A SoM Sie Seely MRS EN I Z e G 9T 9T LI eT gz eI En I y Tt Oe SS oe ee Se eae ee id "68 "d ‘6I | | ade ‘OF 4 ‘og aqe, ‘suomeg X FT— fe | te | 8 | de | #2 | tz elarjirlacfr {ele iet}ol a] e |e] ar far] er tI z | te | te jte- uopeuD IIA WIaVL =e ee a ee | | men] ed ed oe ee fe ee eee | a “spd Zp Oqe, ty el-e ietz —{—|—|-—|-—} -—}- | - | “9g A TT te ‘squoang f+ =j igg Ie lat it It Izit ee a Sb d ‘Zp qE L H —| 1 | + lerleslesizglzFlztietie j—| t i—i — | —| MMMM IMIM et eee et ete oe “d ‘4 JqU ‘syueitg fe4+ ‘pe 7 ‘9 qU L ‘sjuorg Fe+ tiz lt le iz le le {1 Lieta ae Se eS ee ee a re en ee a ae a eae 8 ime se SF 'd ‘Zp QUL I —|—|—|1_ ip l6 l6ziggj6tiggizt] 2) t m Tt a a ed ed eed eee te ct Eg `d ‘p afquy, ‘syuereg E+ ti lo | |e] FIT S| 2Z/%|—|Ft1/8)}6/e jort ge je T j— = ‘gp d ‘Zp qBL SA | —|—|—|—|—|z |T lgt jootjtez'sz|2z| + |—| 1 ‘Grd ‘og aqe ‘syuewwg fg— i—i t]slzlrlolel zla EEE a Sl bale a es EE ‘SF d ‘Zh BL AT | | —;|—|— j 8 | FT |TS |09 jettisesee| gz] e impae Th ‘d ‘ZZ IqQUL ‘syuerwg Z— Lae ie ple "ODE Sri EL) O EA e e e e a er a eee SP d ‘gp orged ta z| elore etier la 6 |m] re eee a 6gd | pe ‘gt pue ZT soiquy, ‘sywomg $I- gol te + | bel fe] te! © |tz tetele |tt Frat] te) E [E+] o |-| €l |r| eri atl etl w | Fe) fel tee | te WORE! IA WIdVi No. 608] MULTIPLE FACTORS IN MICE AND RATS 469 does nevertheless show that 5.4 per cent. of the seventy- three F, young reproduce the degree of spotting charac- terizing one of the parental races. The total number of animals observed, however, is con- siderably smaller than in the previous experiment and the results are therefore subject to greater error. We may now consider several crosses made between hooded and wild rats. The F, generation in such crosses consists almost entirely of self pigmented animals. It is, therefore, in a comparison of the range of variability observed in the F, generation with that observed in the pure hooded race from which the hooded grandparent was taken, that we may expect to find evidence for or against segregation. The first cross to be recorded is between a female of grade — 1% in the 24 generation of the minus selection series, and a wild male. As a control for this female, the progeny of — 1? parents in the second and third selection generations may be taken. It will be seen that their progeny varies between grades — 2 and — } (see line 1, Table VI), in a total of 90 animals. The F, generation shows a range of variability between grades — 14 and grade + 34. It is interesting to note, however, that twenty-three or 37 per cent. of the sixty-two animals observed fall within the range of variation of the hooded grandparent. A comparison of grades — 14 and — 2 in Fig. 2 will show how closely F, animals approach the extreme condition of the hooded grandparental race. In the second cross between hooded and wild rats, females of minus two grade of the sixth generation of the minus selection series were used. As controls for these females the progeny of — 2 parents in the sixth generation of the minus selection experiment may be used (see line 3, Table VI). The 969 young so obtained fall between the grades — 24 and — 4. The forty-eight young show a large range of variability from —1} to +23. Of the forty-eight obtained, twelve, or 25 per cent., fall within the range of variation of the progeny of the hooded grand- parental generation. 470 THE AMERICAN NATURALIST [Von LI The third cross was between wild males and females of grade — 2 or 24 in the tenth generation of the minus selection series. As a control for these animals it is pos- sible to use the progeny of 24 parents in the tenth gen- eration of the minus selection series (see line 5, Table VI). The 474 young so obtained vary between grades — 34 and — 1. Among the ninety-one F, young obtained, eighteen, or approximately 20 per cent., fall within the limits of vari- ability of the hooded grandparents. The range of vari- ability in the F, generation is considerable, being from — 2 on one side to + 2 on the other. Three crosses between animals of the plus selection series and wild rats are recorded. In the first cross (see lines 7 and 8, Table VI) females of grade plus three of the third generation of the plus selection series were used. The breeding capacity of these animals may be fairly judged by considering the progeny obtained from plus three parents in the third generation of the plus selection experiment. The 143 young so obtained range from grade + ł to + 33. Only twenty-one F, young are recorded, varying between grades + 13 and + 34. It is interesting to note that all of them fall within the range of variation of the grandparental hooded race. In the second cross females of grade + 31, fifth gen- eration and females of grade + 34, sixth generation, were crossed with wild males. The progeny obtained from animals of similar grade and generation within the selec- tion experiment can serve as controls. The 320 young so obtained vary between grades + 14 and + 41. The thirty- eight F, young obtained vary between grades +14 and + 3%, again, in every case, reduplicating the grand- parental forms. The third cross, which includes only small numbers and is therefore of less relative value than the two previous crosses, was between a female of + 44, tenth generation and a wild male. The control animals gave twenty-five young ranging from grades + 3} to +43. The F, gen- No. 608] MULTIPLE FACTORS IN MICE AND RATS 471 eration, which consists of only sixteen animals, ranges between the grades +2 and +33. Six of the sixteen young fall within grades characteristic of the grand- parental animals (lines 11 and 12, Table VI). Crosses between hooded and Irish rats have given com- parable results (Table VII). Rats of grades —14 gen- eration 34 minus selection, were crossed with Irish rats. As a control for the hooded animals the progeny of rats of grade — 1} in the third and fourth selection generation may be used. The total of the progeny so obtained is 112. It will be noticed from Table VII, line 1, that they vary between grades — 24 and +4. Only eight or 8.8 per cent. of the F, generation vary outside of the grandparental grades. The range of variability in this generation is be- tween — 2 and + 14. The second cross is between females of grade — 14, from the fifth selection generation of the minus series and Irish males (lines 3 and 4, Table VII). As a control for these the progeny of — 1% parents of a similar generation may be used. The 143 progeny so ob- tained fall within the grades — 2} and — $ inclusive. The fifty-three F, young range between grades — 13 and +1. Almost exactly 50 per cent. of them reproduced the grades of the hooded grandparental types. The third cross is between females of minus two grade in generation 73 of the minus selection experiment and Irish males. As a control, the progeny obtained from the females of — 2 grade in the seventh and eighth selection generations may be used. Such animals produced 2,013 young, ranging from — 23 to — 4. Sixty-six F, young ranged from grades —2 to +2. Of these, 75.8 per cent. reproduced the grades characteristic a the hooded grand- parental race. Two experiments are recorded Foie crosses of plus selection animals with Irish males. In the first of these, the females used were of grade + 2} in the second gen- eration of the plus selection experiment. Animals of similar grade and generation may act as controls, although the number of young obtained from them is very small. 472 THE AMERICAN NATURALIST [Vou. LI The eight young so obtained fall between grades + 14 and + 23. From this cross 239 F, young have been ob- tained ranging from grades —1 to +34. 129 or 53.9 per cent. show grades characteristic of the hooded grand- parent. The remaining cross was between females of grade + 3, third selection generation and Irish males. Controls ob- tained by tabulating the progeny of animals of similar grade and generation show in 143 young, a range between grades + ? and + 3%. Only twenty-three F, young were obtained, ranging between — 1 and + 24. Of these, six- teen, or 69.5 per cent., reproduce the grades characterizing the hooded grandparent. One other striking cross is recorded by Castle in the case of the rats. This is a cross between ‘‘mutant’’ rats showing a particularly advanced degree of plus pigmen- tation and pure wild rats. The range of variability of the pure ‘‘mutant’’ race is according to Castle’s state- ment (Castle and Wright, 1916, page 173) between grades +54 and +53, see Table VIII. The 109 F, ‘‘mutant’’ young show a range of variation between the same grades. According to Castle’s statement on page 174, ‘‘their range of variation does not fall beyond that of the uncrossed mutant race.’’ It would appear as though the evidence of segregation in this case, even to an almost exact de- gree, was clear. This case, together with that of the type of spotting in mice known as ‘‘blaze’’ (Little, 1914) al- ready discussed, appears to show that races differing from each other in only a minute degree of a quantitative character may show segregation clearly. ; TABLE VIII Generation eae rosg ojo ay we jorii Pure mutants. ..........- | 2 | 4 | 28 | 17 | oa To sum up the experiments with rats, it may be said that while it can not be claimed that the evidence is final in regard to the unit nature of the factors involved, they No. 608] MULTIPLE FACTORS IN MICE AND RATS 473 offer distinct evidence of segregation. That is to say, the combination of gametes formed by hybrid parents repro- duce the same zygotic types as did combination of gametes of the pure parent races. The bearing of the results on the practical breeding of farm animals seems clear. If a complicated and highly variable character as the hooded pattern in rats may be introduced in a cross with a non-hooded form and may be recovered in a large proportion of the F, generation, we may encourage crossing as a favorable method of pro- ducing new and important breeds. This will be all the more apparent if we agree with the selectionists who hold that the character reappearing in F, will be at once amen- able to selection and improvement in a desired direction. Contamination of genes in breeding experiments which are conducted on a large scale and are followed by rigid selection, need not be considered as a factor of prime im- portance. ILI. PHYSIOLOGICAL Factors UNDERLYING GROWTH OF IMPLANTED TISSUE The study of the inheritance of spotting in mice and rats has served to give a considerable amount of data more or less directly comparable to that obtained in studies of size inheritance; though presenting, as I shall try to show, certain advantages in their freedom from environmental and age influences and in their definiteness. : We may now consider a very different line of werk, which bears a most interesting relation to other studies of the inheritance of complicated morphological and physio- logical characters. The reaction of various closely inbred strains of mice and their hybrids to implants of a single tumor is definite, and characteristic. It is moreover un- doubtedly hereditary as the work of many investigators: has shown. | Tyzzer and the writer (1916) have reported results ob- tained in inoculating Japanese waltzing mice, closely in- bred races of common mice and their hybrids, with an 474 THE AMERICAN NATURALIST [Vou. LI epithelial tumor which originated in the Japanese waltz- ing race. The parent animals all of them came from races which, because of long continued inbreeding, may be considered to have reached a degree of great genetic homogeneity. It will be seen from the table (loc. cit., p. 403) that all the Japanese waltzing mice inoculated, 58 in number, grew the implanted tumor, while none of the common mice showed continuous growth of the tumor. This absolute difference between the parent races is extremely interesting and offers ideal material for the formation of intermediate conditions of susceptibility in hybrid generations. They further offer a test of the relative value of the hypotheses of multiple factor and of blending inheritance. Sixty-two F, hybrids obtained by crossing the Japanese with tame races were inoculated. Of these, sixty-one grew tumors. The tumors grew in most cases fully as rapidly if not more rapidly than in the susceptible Japanese parent races. It is prob- able that the one animal in this generation which failed to grow the tumor is not a true exception, but that its be- havior may be due to poor technique. The second genera- tion hybrids have given an extremely interesting result. Of the 183 inoculated, only three have shown continuous growth of the tumor. This result is surprising in view of the fact that susceptibility appeared to be a dominant character in the F, generation. Only thirty-eight animals of the F, generation were inoculated; none of them grew the tumor. The striking difference between the F, and the F, generations suggests at once, alternative rather than blending inheritance. As we have suggested in our previous paper, the most logical interpretation appears to be that a certain physiological condition on which the growth of the tumor depends, is produced in the animals of the Japanese waltzing race. This condition is not found in the tame mice used. The differences between the races are hereditary, since succeeding generations of the Japanese and tame mice behave like their parents. The fact that susceptibility of the tumor occurs in both F, and No. 608] MULTIPLE FACTORS IN MICE AND RATS 475 F, generation hybrids shows that the conditions on which the growth of the tumor depends are reproduced in hybrids of the two races. The behavior of the F, hybrid genera- tion produced by reciprocal crosses indicates clearly that even when only one of the parents is susceptible and comes from the homogeneous Japanese race, its contribution to its offspring is sufficiently powerful to produce sus- ceptibility in that animal. In other words, we may say that the hereditary factor or factors underlying suscep- tibility are functional even when present in a ‘‘single dose.’’ If there is a single general factor underlying susceptibility we should expect that the F, generation would show a large number of susceptible animals. This is not the case. It is possible, however, to consider the behavior of the F, generation as being largely due to heterozygosis and not to true inheritance. To eliminate this possibility a back-cross generation was made between F, animals and pure Japanese. The sixty-three animals comprising this generation all proved susceptible to the tumor and in a majority of cases grew it as rapidly as did the F, hybrids themselves. On the other hand, the F, generation crossed with tame mice gave seventy-eight young, all of which were non-susceptible. In discussing the results, it was further suggested that the explanation which best fitted the facts, indeed the only explanation which fits all of the facts, is that susceptibility depends for its manifestation upon the simultaneous pres- ence of several factors in either the homozygous or the heterozygous condition. The gametes of the Japanese race possess all or nearly all of these in a homozygous condition and therefore produce susceptible animals. The F, hybrids possess all of these factors in a ‘‘single dose’’; they having been contributed by the Japanese parent, and are therefore susceptible. When, however, the F, animals form gametes, these factors, if they are mendelizing and not blending in nature, will be distributed at random in the gametes. The result will be that the larger the num- ber of factors involved, the rarer will be the inclusion of + 476 THE AMERICAN NATURALIST [Vor LI all of them within a single gamete formed by the F, ani- mals. Since susceptibility in the F, generation will de- pend upon the presence of all of the factors ordinarily found in the Japanese race, it follows that the greater the number of factors involved, the rarer will be the ap- pearance of a susceptible animal in the F, generation. It further follows that the susceptible animals of the F, generation probably will not possess the factors in a homozygous condition, as did the Japanese grandparents, and therefore they will not, in most cases, breed true, as did the Japanese grandparents, to the character of sus- ceptibility. For a more detailed discussion of these results from a genetic point of view, the reader is referred to Tyzzer and the writer’s earlier paper. It will suffice at the pres- ent time to emphasize the fact that it is the inherent nature of the tissue of the host animal that is being studied. The tumor itself is as near a biologic constant as one can ob- tain. Variation in its growth therefore means variation in the attitude towards it, taken by the host tissue. This attitude appears to be dependent upon a complex of dis- tinct factors. If a change or substitution is made in any one of the.members of this complex, a different reaction is obtained. The behavior of the different factors in any such complex is distinetly that of independent units in in- heritance. The fact that the reactions of susceptibility and non-susceptibility are dependent on multiple factors seems established. If the tissue of the adult mouse may be analyzed in this way, the conclusion is far reaching. If the reaction of the tissue depends on its substance, and its substance depends in turn on a certain hereditary com- plex of factors, it is logical to suppose that the rate and extent of development of the tissue as well as other proc- esses of significance to the organism depend, in so far as they are hereditary, on similar complexes of genetic fac- tors. Environment undoubtedly influences certain char- acters in their development far more than others, and in this respect size appears to be one of the most susceptible. No. 608] MULTIPLE FACTORS IN MICE AND RATS 477 On the other hand, rate and extent of growth is undoubt- edly chiefiy dependent on the nature of the tissue involved, and as we have seen there is every reason to believe that this depends on the interactions of a complex of genetic factors which are independent of each other in inheritance. OTHER EXPERIMENTS WITH INOCULATED TUMORS IN MICE Loeb and Fleisher (1912) have reported a series of investigations on the hereditary factors underlying the susceptibility of mice to a transplantable carcinoma. As parent stocks they used three races of mice, one American race, and two European. By breeding tests including several generations the percentage of American mice to show continued growth of the inoculated tumor was found to be eighty-four, while those of the European races I and II were twenty-three and three per cent., respectively. The F, generation between American and European I gave sixty-eight per cent. susceptible, F, from this same cross gave thirty per cent. susceptible. When American mice were crossed with European II an F, generation was obtained in which one hundred per cent. of the animals inoculated were susceptible. Only fourteen animals were tested and the number is too small to establish this as an accurate percentage for this generation. The F, genera- tion of this cross gave twenty-six per cent. susceptible. It is interesting to note that in the F, generation, where sufficient numbers were obtained to afford critical evi- dence, the percentage of susceptibility was intermediate _between those of the parent races. There is a marked decrease in degree of susceptibility in the F, generation. The fact that some of the animals of the American parent race failed to grow the tumors shows that this race is in all probability not homogeneous, and the same is true of the European races since animals within a single race fail to react similarly to pieces of biologically similar tumor. If such is the case, we should expect an intermediate result — in the F, generation, just as we do when we cross two races differing in size. The F, generation should also be 478 THE AMERICAN NATURALIST [Vou. LI intermediate, though on the basis of blending inheritance we should expect the percentage of positive animals to be much closer to that observed in the F, generation than it actually is. On an hypothesis of multiple factors underlying sus- ceptibility or immunity to the inoculated tumor the ex- perimental results may well be explained. The F, gen- erations of Loeb and Fleisher’s work give a result further indicating the possible presence of multiple factors. If a large number of F, animals mated inter se at random are used to produce the F, generation the percentage of susceptible animals in F, should be roughly approached in the F, generation. 122 animals comprising the F, generation show twenty-four per cent. susceptibility as - compared with thirty per cent. in the F, generation of the same cross. In the F, generation of the American times European II cross, sixty-six animals have been inoculated and have given only two per cent. of susceptibility. This difference is possibly due to the fact that F, animals forming gametes each closely resembling those of the grandparent European race were unconsciously chosen as parents for this generation. The interesting part of Loeb’s work is the fact that the relative homogeneity of the races of mice which he used approximate closely the conditions in respect to sucepti- bility and non-susceptibility which one ordinarily is deal- ing with in size crosses as, for example, Castle’s work and also MacDowell’s work with rabbits. In none of these cases has there been excessively close inbreeding in either parent race before crosses were made. There is, ` therefore, no definite complex of factors characterizing the race. As a result the percentage of susceptible animals varies and depends on the character of the par- ticular animals used for breeding. The whole effect produced is to obscure the true nature of the processes involved. In respect to homogeneity the material at the disposal of Tyzzer and the writer possessed a great ad- vantage which became apparent in the definite results produced. No. 608] MULTIPLE FACTORS IN MICE AND RATS 479 Size inheritance studies have not been recently made with mammals of known ancestry and of approximately pure races. This fact greatly diminishes the value of the results obtained even though they represent work of the most painstaking sort. For this reason the writer started last January a series of experiments on size inheritance in pure races of rabbits. Polish rabbits are being used for the small parent and Flemish giant rabbits for the large parent. It is hoped by a careful study of variation within the pure races to understand more clearly the method of inheritance in the hybrid generations to be produced by later experiments. ` To summarize the work with inoculable tumors, one may say that it presents a type of inheritance not expli- cable on an hypothesis of blending inheritance or of a single variable gene. All the results may, on the other hand, be satisfactorily explained by supposing that the nature of the host tissue and its reaction to the implant depend upon a complex of mendelizing factors. CONCLUSION The fact that three genetically distinct types of spot- ting exist in mice; that segregation of the degree of spot- ting occurs in both rats and mice; that segregation of minute quantitative characters like the ‘‘blaze’’ spotting in mice, and the pattern of the ‘‘mutant’’ rats occurs; and finally that the composition and reaction of epithelial tissue in mice depends upon a complex of mendelizing factors, all indicate that in mammals the multiple factor hypothesis is steadily being strengthened as a scientific theory and a practical principle of great interest and im- portance. BIBLIOGRAPHY Castle, W. E. 1909. Studies on inheritance in rabbits. Carnegie Inst. of Wash., Publ, No. 114. Castle, W. E., and John C. Phillips. 1914. Piebald rats and selection. Carnegie Inst. of Wash., Publ. No. ` 195. 480 THE AMERICAN NATURALIST [Vou. LI Castle, W. E., and Sewall Wright. 1916. Studies of ee i in ee pigs and rats. Carnegie Inst. Wash., Publ. No. Doncaster, ti 1909. On the inheritance of coat color in rats. Proc. Camb. Phil. Soc., oe Ole Ae p. 215. Durham, F. M. 1911. Further experiments on the inheritance of coat color in mice. Jour. Genetics, Vol. 1, 9 East, E. M. 1912. The Mendelian notation as a description of physiological facts. AM. Nart., 46: p. 633. Little, C. C. 1914. Dominant and recessive spotting in mice. Am. Nar., Vol. 48, 4 . Little, C. C. 1914. A ee Mendelian explanation of a type of eaii appar- ntly non-Mendelian in nature. Science, n. s., Vol 904. . Little, C. C. 1915. The inheritance of black-eyed white spotting in mice. AM. Nat., Vol. ex bf 727-740. Little, C. C., and E. E. 1916. Further experiment a studies on the inheritance of susceptibility to ransplantable tumor, carcinoma w of the Japanese waltzing mouse. Jour. of Med. Reisen, Vol. 33, pp. 393—453. Loeb, L., and M. S. Fleisher. 9. Untersuchungen über die Vererbung der das Tumor Wachstum bestimmenden Faktoren. Centralb. f. Bakt, u. Parasit., Vol. 67, p. 135. MacDowell, E. c. i . 1914. Size inheritance in rabbits. Carnegie Inst. of Wash., Publ. No. 196. MacDowell, E. C. - 1916. Piebald rats and multiple factors. Am. NAT., Vol. 50, p. 719. Mudge, G. P. 1908a. On the hereditary transmission of certain coat characters in rats. Proc. Roy. Soc., Series B, Vol. 80, p. 97. Mudge, G. P. ; 1908b. Second paper, same title. Proc. Roy. Soc., Series B, Vol. 80, p. 388. Wright, Sewall. 1915. The albino oa of allelomorphs in guinea-pigs. Am. NAT., Vol. 49, p THE INHERITANCE OF THE WEAK AWN IN CERTAIN AVENA CROSSES’ H. H..LOVE AND A. C. FRASER (In Tn with the Office of Cereal Investigations, U. S. Department of Agriculture.) To some time past, the writers have had under observa- tion many different hybrid series of oats. Certain of these offer an excellent opportunity for a study of the inheritance of awns. This paper is a preliminary report aiming to set forth ideas regarding the factor differences between certain types of awns, as a basis for a further study of the relation between awning and other charac- ters of the oat grain. ; Kinps or Awns Practically all of the wild types of oats are character- ized by a very strong awn. This awn is typically long, stiff, and geniculate. The basal portion of the strong awn is twisted in a clockwise fashion and either black or dark brown in color. . Above the twist, the awn is prac- ` tically straight until it reaches the knee, at which point it turns sharply and proceeds almost at right angles to its former course and usually in a different plane. The first step in the modification of this type of awn seems to be the loss of geniculation, together with a reduction of the stiffness. Then a further straightening of the awn occurs, leaving it practically straight from the point of attach- ment to the tip. Such a change is accompanied by a loss of the dark color at the base of the awn. An awn of this last type is usually spoken of as the weak awn. The weak awn may vary greatly in length, thickness, and rigidity. In some cases it becomes a mere hair-like appendage, ` 1Paper No. 62, Department of Plant Breeding, Cornell University, Ithaca, N. Y. 481 482 THE AMERICAN NATURALIST [Vor. LI Fig. 1. Showing gradations between the spikelet possessing two strong awns, and the awnless spikelet. From left to right are shown: 2 strong awns; 1 strong and 1 weak awn; 1 strong awn; 2 weak awns; 1 weak awn; and the awnless type. : extending scarcely beyond the tip of the lemma and only distinguishable by careful examination. As this awn becomes weaker, it is produced nearer to the tip of the kernel; that is, the rib of the lemma which forms the awn adheres to the lemma for a greater distance before aris- ing as an awn. Among the wild and cultivated types of oats the awns are either characteristically strong, weak, or lacking altogether. In hybrids of these, however, the awns may present all gradations between the awnless and the very strongly awned types. It is usually possible, though, to classify the hybrids as having strong, interme- diate, or weak awns. (See Figures I and IT.) MetHops or Srupy The parent plants and first-generation hybrids were grown in the greenhouse, and the second and third gen- eration hybrids were grown in the field. In the case of the first-generation hybrids, all of the spikelets on all of the plants could be studied. With the much larger Fs, however, it was found impracticable to attempt to study all of the spikelets on a plant. The study was limited, No. 608] WEAK AWN IN AVENA CROSSES 483 Fic. 2. Showing apm in the weak-awn series, from a spikelet having two awns down to the awnless type. therefore, to one representative panicle from each plant. The spikelets were picked from such a panicle, examined for awns, and placed in a small seed envelope on which was recorded the data as to total number of spikelets, number with one awn, and number with two awns. MATERIAL The weak-awned varieties reported on here are the Burt and a strain of the Red Texas. Attention was cen- tered upon the variety Burt. The awnless type in all cases was the cultivated variety Sixty-Day. For the strong-awned oat, a strain of Avena fatua was used. All forms had been grown in pure line culture previous to crossing and were uniform within themselves, Series 514al — Burt X Sixty-Day (Weak Awn X Awnless) A cross between the Burt and Sixty-Day gave an F, which was almost awnless. A few plants had awns on some of the spikelets, but the generation could be con- sidered practically awnless. 484 THE AMERICAN NATURALIST [Vou. LI The second generation plants showed all degrees of awning, from the perfectly awnless condition to those which were one hundred per cent. awned like the Burt parent. These F, individuals were first grouped accord- - ing to percentage of awned spikelets, with a class range of ten per cent. (Table I.) TABLE I F, oF 514412 Per Cent. of Awning ft 0 110 1-10 29 11-20 23 21-30 25 31-40 18 41-50 18 51-60 14 61-70 8 71-80 12 81-90 5 91-99 12 00 66 The occurrence of plants having varying percentages of awned spikelets and forming a more or less continuous series between the parental types, at first suggested a multiple factor condition for awning. A study of the frequencies, however, showed that such an assumption was incorrect. The frequencies of the zero and 100 per cent. classes were too high to accord with an hypothesis of this sort. When the F, plants were grouped in the classes—awn- less, partially awned, and fully awned, it was seen that the data approached, in a general way, a ratio of 1:2:1. (See Table II.) The ratio of the first two named classes to the third was found to be 4.15:1, or, on a basis of four, 3.22: 78. It remained for a study of the F, material, however, to throw light upon the number of factors concerned in this cross, and the relation of these factors to each other. 2 It was necessary to reduce the class range in the case of the 91-99 -= in order to provide the 100 per cent. class. This, however, could have effect on the conclusions drawn from this table, No. 608] WEAK AWN IN AVENA CROSSES 485 TABLE II F, or 51441 P is in sch an cess ee 6k 110 } 274 Partially swned CA ARAN 164 FANY AWE a ha. heal pn 66 66 340 340 Behavior of the Fully-awned F, Plants.—Seed of three of the fully-awned F, plants was sown in pedigree culture for an F,. The results obtained are shown in Table III. TABLE III F, FROM FULLY AWNED F, PLANTS—514A1 Pedigree % Awns in Fe | Awnless Partially Awned | Fully Awned 514a1-22 100 | 0 0 47 5l4al-88 | 100 0 0 36 5l4ai-05 | 100 0 0 | 20 | } Total oe | 0 0 | g From this data, it would appear that the fully-awned type is the pure recessive. We have already seen that this type has very little influence on the F, hybrid, and that it appears in F, in only about 25 per cent. of the individuals. Here we find the fully-awned plants breed- ing true. Behavior of the Partially-awned F, Plants.—Twelve plants which showed some awning in the second genera- tion were planted in pedigree rows for F, The per- centage of awning in these plants varied from eleven to eighty-seven. In spite of the wide difference as to per- centage of awning, their behavior was strikingly similar. With but one or two exceptions, the ratio of plants not fully awned to those which were fully awned was close to 3:1. A total of all these F, plants gave the frequencies 419:118, or a ratio of 3.12:.88. In this case the deviation is 2.4 times as great as the probable error, but this can be accounted for by the somewhat wide devia- tions occurring in cultures 106 and 194. With the excep- tion of these pedigrees, the deviations from the expectancy are not more than twice the probable error. It is of 486 THE AMERICAN NATURALIST [Vor. LI interest to note that plants with a high percentage of awned spikelets in F, did not tend to give a correspond- ingly high number of awned plants in F;. Neither did plants with a low percentage of awns in F, tend to pro- duce more of the awnless or partially-awned types in Fs. TABLE IV F, FROM PARTIALLY AWNED F, PLANrs—514Al | Cent. Par- | Fully | Not | Fully | Ratio | Pedigree | Awns Awnless| tally | Awned | Fully | awnea| x4 | D | PE: aer | Bia 65) 77 | 18 | 16") ee [eee Ep PY edo sa86 oe -100} 24 | 36 | 17] 19 | 53 | 19 | 78 | 06 |.137| 44 106) 11 | 81 7 sf 38 | e PAB ior [trbt wos -119; 20 | 17 R ae are nea 20 tee ae -198|-s7 | 24 | 11 | 129 | 36| i oe [ete | 178 ie iss} 32° fia | i5 5 | 28 5 | 33° | (39 | .203'| 1.92 -472| 15 |. 26 el ki B 8 | SN | -185,| 1.08 ~194|14,|,.36 |: 19,}. 21 | ,55.| 1m | *33 | 33 | 144 | 2.20 saol 18 | 24} a2 {ar | deeh an pI .06 | .a7n | 35 sme æ p niw | 8 Jean a| OS | sis {aoe 9 -244, 60 | 13 | 17 7 1 30 CTT ae 192) 1.24 264. 81 | 23 Gj. 101m.) 10 | oo | ,.08 | 187 | 16 | 267 |. 152 | 118 | 419 | 118 | 222 | 42 | .05 | 2.40 Behavior of the Awnless F, Plants.—Eleven of the awnless F, plants were selected for study in F,, and se from them was sown in short pedigree rows. The be- havior of these plants is shown in Table V. Five of these awnless plants bred true to the awnless condition, giving a total of 249 awnless plants in F,. The other six broke up into awnless, partially-awned, and fully-awned plants. In no case did the ratio of these types suggest a 1:2:1 ratio. When we group all of the plants which are not one hundred per cent. awned, however, and compare them No. 608] WEAK AWN IN AVENA CROSSES 487 with the fully-awned plants, we find that the separate ratios closely approximate a 3:1 ratio. The ratio for all six plants is 2.97:1.03, and its deviation is practically the same as its probable error. TABLE V F, FROM AWNLESS F, PLANTS—514a1 | Aa Par- Not | Pedigree | Awns Awnless| tially | Aon, | Fully | rey} Beto | p> | px. | DIP.E. | in Awned Awned 5l4al— dila D 48 0 o |. 48 0 126 0 75 0 0 75 0 -185| D 31 0 0 31 0 -336| 0 32 0 o | 32 0 232) 0 63 0 o | 63 0 Total 249 249 3.24 | 514a1-339 0 34 0 8 | 34 8 76 | 24 | -180 | 1.33 -36 0 60s akaa T 66.1, g oy 14 .140 | 1.00 -176| 0 A E a a E aita 02 178 -221| 0 ite 8 7 | il 7 | 223 | 97 | 249! 1.08 1.27 | Ma L Bir LBT 95 | e Fa as] 2.70 -291| 0 odes i u a a | 222 | 02-|.178) 4a 1.02 | ; : 2.97 To | o a | 67 | 198 | GY | 75, | 08 |072| 43 It is apparent from these data that the F, grouping used here includes in the awnless class certain individuals which are heterozygous for awning, and which really belong with the partially-awned plants. According to Nilsson-Ehle (1914) environmental conditions have an effect upon the production of awns. It is quite possible that the failure of these six plants to produce some awns is due to undetermined environmental factors. A comparison of the relative numbers of awnless and partially-awned plants in Tables IV and V would seem to indicate that awnless F, plants tend to give a higher per- centage of awnless plants in F, than do the partially- awned plants. This may be explained, however, by the 488 THE AMERICAN NATURALIST [Vou LI fact that a high percentage of the awnless plants in the second generation were yellow in color and consequently many of them might well carry a factor which inhibits awn formation. Data will later be presented to show a definite linkage between the awn-inhibiting factor and the factor for yellow color in the Sixty-Day. Series 2501al — Burt X Sixty-Day A second series of hybrids between Burt and Sixty- Day behaved in a manner similar to series 514al. The F, was nearly awnless in both the direct and reciprocal crosses. The F, results are shown in Table VI. TABLE VI F, or 2501 Purcalty | way. | a | Paty | Ratio Pedigree aie aaah kaii P Awat 34 D P.E.. | D/P.E. 2501b1 26 96 43 | 122 | 43 ns 04 |.0909| .44 250lar1 ‘| 17 61 23 | 78 | 28 yS 06 | .1134!| .53 2501ar2 4 32 17 | 86 | 47 mg 28 | 1604 | 1.75 2501ar3 1 11 4 12 4 Aa 00 | .2921 0.00 Total | 48 200 92 | 248 | 92 1e 08 |.0634 | 1.26 2501b1 = Burt X Sixty-Day. ; 250larl =Sixty-Day X Burt—selection of partially awned F, plants. 250lar2 —=Sixty-Day X Burt—selection of awnless F, plants. 250lar3 = Sixty-Day X Burt—unselected F,. It will be seen from Table VI that the partially-awned and awnless types of F, gave practically the same behavior, each throwing about 25 per cent. of fully-awned plants in F,. Series 2401al — Red Texas X Sixty-Day A cross between Red Texas (weakly awned) and Sixty- Day gave an F, showing only 1.3 per cent. of awning. The second generation of this cross has not yet been grown. No. 608] WEAK AWN IN AVENA CROSSES 489 Discussion of Results in Weak-awn X Awniess.—From the data presented above, the following conclusions may be drawn as to the inheritance of awns in crosses between the weak-awned and the awnless types of oats. The awnless type is almost completely dominant in the first generation, only a few of the plants possessing awns and those in small percentages. ; The second generation gives awnless, partially-awned and fully-awned plants in a ratio which approximates 1:2:1. The totals of data from second generation plants of series 2,501 and 514 are reasonably close to this ratio: | Awnless | Partially Awned Fully Awned Série Sh0¥ i AU: 110 | 164 66 Maries: B14 erani i nere | 48 200 92 irinik. sen annoy | 158 | 364 158 tak a a 170 | 340 170 The behavior of the fully-awned plants shows that this type is the pure recessive, for it breeds true in all cases from the second generation. All of the partially-awned F, plants proved to be heterozygous, throwing in the third generation approxi- mately three plants not fully awned to one fully-awned plant. The awnless plants of the second generation were found to comprise both homozygous plants of the parental type and heterozygous intermediates which later broke up in the same manner as the partially-awned F, plants. It might be expected that some of the awnless F, plants would prove to be heterozygous, since awnless plants are found commonly in the first generation. From these results, it is apparent that we cannot cor- rectly speak of the awnless oat as the dominant, type, nor should we restrict the use of the term intermediates to those plants which are partially awned. It seems very probable that the difference between the weak-awned and the awnless varieties of oats, at least in the varieties studied, may be accounted for by the assump- 490 THE AMERICAN NATURALIST [Vou. LI tion of a difference in one pair of genetic factors. It may be that awnlessness is a definite character which is a true allelomorph of the fully-awned condition. Some might prefer, however, to consider awnlessness simply as the absence of awning. In that case we must assume the presence of an inhibitory factor to account for the partial - dominance of the awnless Sixty-Day over the weak-awned Burt. The data at hand seem to point to the presence of an inhibitor to awning in the variety Sixty-Day. A pre- ponderance of awnless yellows in F, and F, suggests a linkage of this inhibitory factor with the factor for yellow color in the Sixty-Day. (See Table VII.) Such a finding would be in agreement with the results of Nilsson-Ehle (1914). A very definite linkage of the inhibitory factor with the factor for yellow color has already been observed in a cross between A. fatua and A. sativa var. Sixty-Day. This will be brought out in a later publication. TABLE VII SHOWING THE DISTRIBUTION OF REDS AND YELLOWS IN SERIES 514, WITH PERCENTAGE OF AWNING AS RELATIVES | 0 | 5 | 15. ir 25 |) 35 | Ane | 55 | 65 | 75 | 85 | 95 | 100 Reds i) 48°)| 120] 40. faBoler® fo 420 fob |, Sopi fy 74-88 Yaw 60 | 15 | 12 | 12 | 4] 10] 4] 2 | 2 /}01| 4 | 21 | Certain other crosses with the Burt show that this variety contains a factor for yellow which does not inhibit awning. In the crosses Burt (red) X Swedish Select (white), and Burt X Early Champion (white), the F, con- tained a certain number of yellow-seeded plants, which in turn gave some yellows in F;. All of these yellows were fully awned. The existence of this yellow factor in the variety Burt has complicated the study of the yellow of the Sixty-Day in these crosses. The fact of the presence of this yellow in the variety Burt should be kept in mind when Table VII is examined. It will be seen in Table VII, that the red grains are nearly as numerous in the 100 per cent. class as in the 3 The classes are as follows: —0, 1-10, 11-20, ...., 91-99, -100. No. 608] WEAK AWN IN AVENA CROSSES 491 awnless class, and that the other classes are represented in practically equal numbers. In the case of the yellows, however, there are about two and one half times as many in the awnless class as in the fully-awned class. Fifty- seven per cent. of the yellows have less than 20 per cent. of awning, and seventy-three per cent. have less than 30 per cent. of awning. Many of the yellows in the 100 per cent. class are doubtless due to the yellow factor con- tained in the Burt parent. This factor does not inhibit awning. Strong Awn X Awnless.—The results of crosses be- tween Avena fatua and the variety Sixty-Day (A. sativa) agree closely with those obtained in the crosses between the weak-awned and awnless types. (See Tables VIII and IX.) TABLE VIII F, TOTAL or SERIES 2516 ' Partly | Fully pateee | amni, | Zae | Zu, [nern Eug | Reo | o | ee, [oee 2516 | 169 | 377 | 201 | 546 | 201 | aa | 08 | 04 | 2.00 TABLE IX F, or Heterozygous F, PLANTS Pedigree ||, Awniom) | Fey | Fuly, [Not Futy| Fuy: |, Ratio | n, | pe | pje 687a1-15 | 61 im | 67") ish | e7. | Re) 06 | 07 E EN aa 53 | 41 | 132 | 41 Toe k- 05. | 988 | 56 687a1-1 | 15 55. | 24 70 ih, Bg ane iaa AT In a similar study on A. fatua X A. sativa var. Kherson, Surface (1916) obtained results which agree closely with those presented above. The F, plants were nearly inter- mediate, although ‘‘The majority of F, spikelets show no awn whatever’’ (p. 265). In the second generation the following types appeared: . : 492 THE AMERICAN NATURALIST [Vou. LI At first Surface assumed that the awnless plants were homozygous and should, therefore, breed true. A test of these plants, however, showed that a certain number were heterozygous. Fifteen out of twenty broke up in the third generation. This might be expected from the fact that some of the heterozygous F, plants were awnless. The failure of these plants to produce a few awns is attrib- uted by Surface either to an undiscovered factor affecting awning, or to an environmental influence. It seems quite ‘probable that the variety Kherson may carry a factor inhibitory to awning, similar to the factor in the Sixty- Day.* OTHER CHARACTERS OF THE GRAIN In connection with the above studies on awning, studies were also made on the presence of basal hairs and the type of articulation of the lower kernel of the spikelet. A strong correlation was found to exist between the fully awned condition and the Burt type (similar to that of A, sterilis) of articulation, and also between the fully- awned condition and the presence of medium-long basal hairs such as are found on the Burt grains. When the spikelets were all awnless, the union of the lower kernel and its rachilla was generally of the type found in Avena sativa and the basal hairs were either short or lacking. It is interesting to note, in the crosses between the weak-awned and awnless types, that in every case where a panicle had two awns on a spikelet, all of the spikelets on the panicle were awned. The irregular occurrence of these two-awned spikelets, and the wide variability in numbers on a panicle, makes it seem probable that there is no definite factor for the two-awned condition. It seems more likely that the occurrence of such spikelets is due to environmental influences upon the factor for com- plete awning. i #In some logalities the names Kherson and Sixty-Day are used synony- mously. No. 608] WEAK AWN IN AVENA CROSSES 493 BIBLIOGRAPHY Nilsson-Ehle, H. 14. Über einen als Hemmungsfaktor der Begrannung auftretenden arbenfaktor beim Hafer. Zeitschr. ind, Abst. u. Vererb., 12: Surface, F. M. 1916. apresa on Oat Breeding. III. On the Inheritance of Certain e Characters in the Cross Avena fatua X A. sativa var. PA Genetics, I: 252-286. SHORTER ARTICLES AND DISCUSSION NOTES ON THE FAUNA OF GREAT SALT LAKE In the years during which the writer was zoologist at the Uni- versity of Utah (1908-15) observations were made on the life of the Great Salt Lake, when time could be spared from multitudi- nous teaching duties. The animals of this brine lake earliest reported, Artemia fertilis Verrill, and the larve of the small Dipteron, Ephydra gracilis Packard, were naturally the first at- traction, since they were abundant, commonly known to science, and readily observable to any one looking for them. A second species of Ephydra, E. hians Say, was reported by Aldrich in 1912.1. A Chironomus has been reported also, according to Tal- mage,” but no reference to the authority is given and his own statement is confusing, as he says he has ‘‘confirmed the pres- ence of . . . the larve of one of the Tipulide, probably Chirono- mus oceanicus Packard’’! He further states that ‘‘ The larvee of the tipula may be taken anywhere near shore during the warm months,’’ but the present writer is compelled to state that neither larve, pupe or adults of either a Tipulid, or of a Chironomus was ever observed by him in the Lake, nor are any such reported by Aldrich, an authority on Diptera, in his reports of collecting about Great Salt Lake. Other forms than Ephydra might well occur in such a portion of the lake as the great Bear River Bay, where the salt content of the water must be much less, owing to a great influx of fresh water from the Bear River, and to the fact that the bay is partially cut off from the main lake by the causeway of the Southern Pacific Railway. Aldrich, however, has certainly been on the Bear River Bay side of the cut-off, as shown by Plate II, Fig. 8.1 In tentatively ‘‘ trying out ’’ to see what might be a profitable line of study several dilutions were made in order to note the 1 Aldrich, J. M., ‘‘The Biology of Some Western Species of the Dip- terous Genus Ephydra,’’ Jour. N. Y. Ent. Soc., XX, 77-99. In this are photographs indicating the enormous numbers of Ephydra in the lake; also first complete description of E. gracilis, 2 Talmage, J. E., 1900, ‘‘ The Great Salt Lake, Present and Past,’’ 67-68. 3 Aldrich, J. M., ‘‘Collecting Notes from the Great Basin and Adjoining Territory (Dipt., Col.) ,’’ Ent. News, XXIV, 214-221, 494 No. 608] SHORTER ARTICLES AND DISCUSSION 495 effect, if any, on Artemia. The brine varied in density in the lake at Saltair during the years of these observations, according to season and rainfall, and besides the annual fluctuations gradu- ally became less dense on account of a cyclical period of heavier precipitation and consequent rise in level of the lake. In Oc- tober, 1909, the density was 1.158, and in April, 1915, 1.136. Dilutions were made by addition of distilled water as nearly as was feasible to the following densities: 1.12, 1.10, 1.08, 1.06, 1.04, 1.02, 1.014 These dilutions were placed in small aquarium jars, filled only 14 to 1% full, and covered to prevent entrance of dust, and undue evaporation. It may be said at once that while Schmankewitsch’s classic observations on Artemia were of course in mind, it was not hoped to repeat them with the small amounts of water used. However, some data obtained, though incom- plete, seems worth recording, since the present writer can scarcely hope ever to have further opportunity for pursuing this line of investigation to the extent it deserves. It is sin- cerely hoped that some one may be able to further investigate the fauna normal to the unusual ecological conditions of brine lakes. Resistance of the adult Ar/emzas to sudden changes in the con- centration of the salts, while greater than anticipated, only ` showed that they could not indefinitely survive too great a change. Plunged into tap water, they appeared ‘‘heavy,’’ sink- ing at once to the bottom, from which with most vigorous swim- ming movements they were barely able to rise. Exact data as to length of life in these solutions is not sufficient to offer, but in general they survive a change to completely fresh water but a few hours, and they do not survive for long periods in water in which the amount of salts has been reduced to less than half the normal. In stronger concentrations they survive for sufficiently long periods that it seems likely they would live therein for a normal life period if other conditions were favorable. ` Kellogg’ had opportunity in the case of A. franciscana Kellogg to note differences in size, color and abundance of individuals which had developed in waters of different densities, and it is interest- ing to note that he found them largest in waters ranging from 1.11 to 1.13 in density. The latter figure is nearly the same as 4See article by Daines preceding this, with exact densities e one series. Part of these observations were made in collaboration with Dain 5 Kellogg, V. L., ‘‘A New Artemia and its Life Conditions,” play N. S., XXIV, pp. 5 594-596. 496 THE AMERICAN NATURALIST [Von. LI that for Great Salt Lake water at the beginning of my observa- tions, at which time the lake was in the rising period of its long cycle of rise and fall, which rise continued at least up to 1915. Some few years prior to 1909 the lake had been much lower and the water at nearly the saturation point for NaCl. I believe 1.13 is somewhere near the mean density for Great Salt Lake. Interesting facts were noted concerning eggs contained in dilutions made in autumn (see annual cycle below). These hatched in a few days or weeks, and they first hatched in the most dilute water, next in the next more concentrated, and so on up the seale of concentration in nearly regular order. The conclusion naturally presents itself that the stimulus to develop- ment lies in the reduction in amount of salts present, but later it appeared (this point was not finally cleared up) that it lay rather in a lack of oxygen resulting from insufficient aeration of the water used. Young thus hatched never reached maturity. Ephydra larve are even more abundant in the lake than Ar- temias. They were found to be remarkably resistant to changes in density of water, as well as to other changes in liquid environ- ment. These larve will live at least for days in tap water, but whether they could be brought to maturity in this or in very dilute lake water was not determined. The fact that the puparia drift up on shore in great ‘‘windrows’’ has already been noted — by Aldrich, and in the Canadian Entomologist for 1891 (orig- inal article not seen). The countless swarms of imagoes may be seen by bathers resting on the surface of the water or flying up at will, and it was found to be an easy matter to obtain the eggs by imprisoning a number of these in a covered erystalliza- tion dish with clean bottom, partly filled with brine, showing, as suspected, that they drop the eggs freely into the water. AS this was not done until near the close of my service in Utah, no experiments were made with the eggs, but attempts to hatch the eggs and rear the insects to the imago stage in dilute lake water and in fresh water should be made. As instances of the resist- ance of these larve may be mentioned the following: In more than one case the larve were observed to live months in brine which had evaporated to saturation, and beyond to the point of containing a heavy deposit of erystals and of being completely encrusted on top, and in one such case practically all of the water had disappeared. Among the salt crystals in the little remaining water the larve were somewhat inactive, but appeared. to- be in good condition when water to about the normal amount a No.608] SHORTER ARTICLES AND DISCUSSION 497 was restored to the jar. Artemia is resistant to concentration, but not to the same degree as Ephydra larve. Again, in an at- tempt to kill larve without distortion some were placed in Perenyi’s fluid and in this were capable of movement after more than twenty-four hours. In Flemming’s fluid they live several times as long as Artemia, but I have no record of the exact length of time. I am able to verify with salah ia PERE s belief that these larvæ do not rise to the surface for a Most important of the incomplete Gaeta were those indi- eating the presence of Protozoa as normal inhabitants of Great Salt Lake. So far as I am aware, no Protozoa have been previ- ously reported from brine lakes. Representatives of this group, notably Amaebe, were first seen in the moderate dilutions after some weeks in the laboratory, which proved to be in a sense cul- tures. In March, 1910, several jars of a series, including one of undiluted lake water, contained an abundance of these forms. The specimens were of two or three varieties or species, by far the most common being very like Ameba limax. I should not have hesitated to call it that in a fresh-water culture. A class of some 15 students was well supplied with Amebe for ‘labora- tory work from one of these jars. Occasionally, in making micro- scopic examinations of the cultures other Protozoa were met with, but never in numbers. In fact only a single specimen at a time was the rule. Specimens of Ciliata were seen, some closely resembling a species of Uroleptus, while at least once a species of Euglena was definitely noted. Chlamydomonas appeared quite regularly and in great quantity in many of the cultures.* I believe several species of Protozoa to be present normally in the Great Salt Lake, but not generally very abundant, as many of my efforts to secure them directly from the lake were failures. However, some were certainly obtained directly from the more or less decayed masses of organic débris which collect in enor- mous quantity in the great stretches of shallow water along the very flat shore, which masses consist mainly of the gelatinous blue-green alga, Aphanothece packardii.* (This is the alga ‘‘of the Nostoc group’’ mentioned by Aldrich.) In this material it was expected there might be found Nematodes, as they are in so many cases adapted to unusual environment, and so commonly present in decaying substance, but none ever came under obser- vation. Perhaps a statement of the annual cycle of life of Artemia and Ephydra may be of interest. For the latter it may be said that 498 THE AMERICAN NATURALIST [Vou. LI larve and pupe are at all times of the year present in the lake, though less abundant in winter. In the winter months there are but few in the open water, but they are common in the débris above mentioned. Dates for first appearance of adults were not secured, though some appear as early as April; they become com- mon by June, and in July and August are so exceedingly numer- ous as to be a serious nuisance at times about Saltair pavilion, wind conditions being apparently a determining factor in their coming in swarms about the bathhouses. Ordinarily they keep below the level of the floors, on the piles and on the water sur- face. Whether any eggs survive the winter can not be stated. There is no evidence of the pupe surviving on shore, where thrown up by the waves. It seems likely that larve and pup which remain submerged are the principal, if not the sole means of surviving the winter period. Adult Artemias, the females with fully developed egg sacs, are very plentiful throughout the summer and fall into October. In this month the temperature of the water falls from the summer temperatures of between 25° and 30° C. (exact summer maxi- mum unknown to writer) to as low as 15°-18° C. In November with the temperature as low as 6° C. there may still be seen some few adults. At a December temperature of 1° C. and lower (doubtless goes lower at times for short periods) no adults can be found, as a rule, though reported by Talmage. An abun- dance of eggs can be secured in fall, winter and spring, especially in the débris near shore. Possibly some may settle into the smooth oolithic sand of the open lake bottom, but I have no evi- dence that such is the case, and the eggs tend rather to float than to sink. Young appear in April and May, abundantly in the latter month. The earliest record secured for young was March 12 (1910), when a number of minute young were taken. The temperature at that time was 9° ©. It will be noticed that Ar- temia differs markedly in its long season of activity from its fresh-water cousin, Branchipus, which is so soon gone from its evanescent breeding pools. Correlated with this long active period is the continued presence of abundant water and f and an entire absence of enemies. Enemies play no part in keep- ing down the numbers of Artemia, or of Ephydra in the larval stage. In the midsummer bathing season both are present in myriads in the open water, but so transparent are they that the average bather, even the native Salt Laker, seldom notes their presence. No. 608] SHORTER ARTICLES AND DISCUSSION 499 The insect fauna of the lake shore presents material for a study in itself, on which nothing has been published save the material on Diptera by Aldrich, already cited. At the Univer- sity of Utah I left the beginning of a collection of insects taken in or on the waters of the lake, and I recall that a small Corisid was several times seen and some specimens of it taken swimming immersed in the brine near shore. The species appeared to be the same as one common in fresh and slightly salt and sulphur impregnated waters in the Salt Lake valley. Probably correlated with the abundance of Ephydra adults as food, may be mentioned a ‘‘plague of spiders’’ with which the resort (Saltair) was troubled during one bathing season, about 910. Several cases of persons being bitten by spiders were re- ported in one of the Salt Lake papers, though I can not vouch for their authenticity. Certain it is that spiders of more than one species were unusually numerous about the pavilion, as I personally observed, and I learned later that the employees went about with brooms every morning before the hour for opening and destroyed as many as possible. The forest of piles and un- derpinning beneath the structure, however, was an inexhaustible reservoir from which the supply was constantly renewed. After the close of the season, no other remedy having been found, some employees were kept busy for weeks in boats beneath the huge structure collecting and destroying the egg cocoons, and the next season there was no serious trouble. Many bushels were thus collected. The second autumn this task was again taken up, and since that time no further plague of spiders has appeared, but whether autumn cocoon collecting is still kept up I do not ow. I have no doubt that the seemingly sudden appearance of the great numbers of spiders was in reality but the time when, owing to the availability of a great food supply and plenty of space for spreading webs, they reached a high point in numbers, the culmination of years of slow inerease. Cuas. T. VORHIES UNIVERSITY OF ARIZONA, Tucson, ARIZONA ON THE FLORA OF GREAT SALT LAKE Very little investigation has been made of the plant life of Great Salt Lake, either of a scientific nature or otherwise. So far as the author of this paper knows, but one attempt has been made in the past so scientifically classify the flora of the lake, and 500 THE AMERICAN NATURALIST (Von. LI that attempt was interrupted before it had reached a successful conclusion, so that no publication of the work was made. Brief mention of what literature we have on this subject seems not to be out of place here. Professor Farlow (1879)! published a description of a blue- green alga, Polycystis packardi. Dr. A. Rothpletz (1892)? makes mention of the presence of certain genera of blue-green alge, connecting them with the for- mation of peculiar ooliths on shore. Dr. Rothpletz did his work as a geologist, from a geological point of view. He made no systematic study of the lake from a botanical standpoint. The genera of alge he mentions—Gleothece and Gleocystis—we have been unable to find in the part of the lake studied, and it might be said, too, that the connection between these and the ooliths has not been generally accepted, even by geologists. H. F. Moore (1899)*, in reporting on the feasibility of intro- ducing useful marine animals into the waters of the lake, makes mention, briefly, of the presence of diatoms. As diatoms con- stitute the chief food of the oyster, their presence was of con- siderable importance in the investigation, and especially since they are found in greatest abundance at the mouths of rivers where the density of the water is more favorable for the develop- ment of the oyster. Talmage (1900)* speaks of the presence of at least three species of alge—not naming them—and, besides these, he calls attention to the presence of diatoms beds off from shore, as well as living diatoms in the lake. Tilden published in her distribution entitled ‘‘ American Alge,’’ several species from Great Salt Lake. This distribution has not been available, therefore, more definite mention of it can not be made. They are as follows: Aphanothece utahensis, no. 297, Polycystis packardii, no. 298, Dichothriz utahensis, no. 288, Enteromorpha marginata, no. 266, Enteromorpha tubulosa, no, 262, Chara contraria, no. 255. No other proof of the presence of abundant plant life in the 1 This paper was not available. 2 Rothpletz, = a, Centr., p. 3 Moore, H. F., ‘‘ The Feasibility of Introducing Useful Marine Animals into the ae “of Great Salt Lake. No. 608] SHORTER ARTICLES AND DISCUSSION 501 lake is needed than the presence of a fauna, abundant in indi- viduals, if not in species. And no further demonstration of the presence of this fauna is required than for one to visit the lake and see, with his own eyes, the water literally teeming with animal life. The presence of plants is not so evident to the casual observer, although, at certain times of the year, clumps of greenish ma- terial, which must at least suggest a vegetable growth, are very plentiful. Areas of a green scum on the surface of the water in more or less protected places also give evidence, directly, of the presence of plants. The original purpose of this paper was to determine, if pos- sible, the effect as to size of cells and rapidity of growth of dif- ferent densities of Great Salt Lake water on a species of Chlamy- domonas which is foûnd there. The problem, then, was to have been purely a physiological one. During the course of investi- gations along this line, however, other interesting things pre- sented themselves, and a deviation was made from the first plan, so that finally observations were extended to include every species of plant found in the part of the lake investigated. The observations made covered a small portion of the southern end of the lake at what is known as Saltair Beach. This place is easily accessible, and is at such a distance from any stream entering the lake, that the density of the water there is not af- fected to any degree. The following plants are found PNAN i in the water at that place: A green alga, Chlamydomonas sp., which has been examined by Dr. N. L. Gardner. Dr. Gardner believes it to be a new species —he has not yet published a decription—near to Chlamydomonas glæocystiformis Dill, and Chlamydomonas apiocystiformis ar- tari. It has a rich green color, and occurs, during the warmer weather, on the surface of the water in many more protected places. It is found in less numbers in whatever decaying plant or animal material may be present. The indications are that this is one of the means the plant has of surviving the winter; since such material brought into the laboratory in the very cold- est weather has later developed a rich green growth of the alga. A blue-green alga, determination of which has been made by Professor W. A. Setchell. He says that it certainly is an Aphanotheca, and is undoubtedly the same plant as the one named Polycystis packardii by Farlow, and probably also the 502 THE AMERICAN NATURALIST [Vou LI same as the one distributed by Miss Tilden from the Great Salt Lake and named by her Aphanothece utahensis. On the au- thority of Professor Setchell, we shall designate it Aphanothece packardii. The plant occurs in small masses, irregular in size, floating in the water and piled up by the waves on shore. These masses show a gradation in color from a deep blue-green, to light brown, and some were colorless; this depending, no doubt, on the con- ditions of the plants in the individual clumps, and not, as has been suggested,* on a variety of species in the clumps. Micro- scopic examination of this material shows the individual plants in the mature condition, and also in various stages of division by fission. Great numbers of the cells are held together by their gelatinous secretions. The individual plants average about two micra in diameter. Microscopie examination of the lake Waie reveals at least two species of diatoms. They probably belong to the genera Navi- cula and Cymbella. These plants do not occur in sufficient num- bers, in the denser water about Saltair, to be seen with the naked eye. The fact that putrefaction and decay are taking place in the water, especially near to shore, where organic material is abun- dant, shows conclusively that bacteria are present. Here it may be well to suggest that at least some of the plants distributed by Miss Tilden as Great Salt Lake plants, in all prob- ability came from the fresher waters at and near the mouths of rivers, or in the bays formed by the rivers at their place of en- trance into the lake. As the present observations were confined to the denser waters, even an indication of the plants referred to—with the one exception noted—was not found. For the physiological work, water was transferred from the lake to the laboratory, in sufficient amount to make a number of series of dilutions in glass aquaria. These series included solu- tions of different density, varying in specific gravity from 1.0115 to 1.222, a saturated solution. Masses—large in some series, but — in one—of Aphano- thece packard were placed in these solutions, and, in every case, enough Chlamydomonas was thus introduced to start a more or less flourishing growth. From time to time, measurements were made of the Chlamy- domonas present in the solutions, and during the first few 4 Talmage, J. E., ‘‘ The Great Salt Lake—Present and Past,’’ p. 76. Salt Lake City. : No. 608] SHORTER ARTICLES AND DISCUSSION 503 months indications pointed very strongly to the fact that a re- duction in size followed transplanting into less dense solutions. A table below shows the results obtained with the first series. No. 1 contains the water as obtained from the lake, analysis of a sample of which (1910)° gives the following: Constituents Grams per Liter % of Sample Taken | % of ‘Total Solids ns solids 242.25 20.887 — lorine 126.35 10.91 52.23 a dical 16.00 1.38 .65 Bait ite) bt Peers epee pein 5.18 0.45 14 Calcium 0.98 0.08 | .39 Sodium 85.10 7.25 34.68 Potassium | 8.82 0.76 .66 Rotak, of constituents........... 242.45 20.83 | 99.75 Salin kes. 213.32 18.39 88.09 Solutions 0 and 00 were allowed to become further concen- trated by evaporation ih the laboratory. Nos. 2 to 8, inclusive, were diluted with distilled water. The first measurements were made some time after the series was started to allow the plants to become accustomed to the new conditions, only, indeed, after multiplication had begun. Blank spaces in the table indicate that no motile zoospores were present at that time in the solu- tion. SERIES No. 1 STARTED OCTOBER 8 Measurements o. Density | Dec. 15 | Jan. 13 | Feb. 14 |. Mar, 10 | Apr.16 | June 15 Average Solution |. | | 00 1.222 11.76.7| | 15X10 |13.38.3 0 1.1825 | 13X7 ; | 1510 |13.7X7.8 1 1.1580 | 13x7.25| 12X .5|}136.5)12%5.4 |12xX6 112X 7/12.2Xx6.1 2 -1239 | 13X5 10X 5/112. 3 .1088 | 12X5 11x5 11.2x4.5 11X6 9x 5/10.8X5.1 4 .0822 | 12.5X5 12>¢6.5'11.255.5/10 5.5 11.45.6 5 1.0613 11x43 105.5 | 10X 5 10.3X4.9 6 1.0400 | 94.5 9x5 | 9Xx5.5 {105 9.255 T 1.0190 8.55 94.5 (8.5*4.5 8.75 8 -0115 ; 11x6.5 |10x5 |10X 6|10.3X5.8 This table seems to show a slight diminution in size as we pass from the more dense to the less dense solutions, with the excep- tion of the last and least dense of the solutions. It must be said that it is very difficult in measuring Chlamydomonas zoospores 5 McFarlane, Wallace, ‘‘The Water of the Great Salt Lake.’’ (Read be- fore the summer meeting (1910) of the Am, Chem. Soc. at San Francisco by Professor W. C. Ebaugh.) 504 THE AMERICAN NATURALIST [ Vou. LI to make definite comparisons as to size. The size of the indi- vidual cells even in one solution varies so greatly that one can only obtain an average of the size and then very roughly. The measurements recorded in the table, and all others made, repre- sent the average size of the larger cells in the solutions as far as it was possible under the circumstances to measure them. The results from the other series did not corroborate definitely the results shown for the first series. Therefore, the only con- clusion which can be drawn is, that so far as the present work has shown, variations in density of the water of Great Salt Lake cause no corresponding variations in size of Chlamydomonas cells. In every series but one, decided growth of the Chlamydomonas began first in the dilutions about No. 5, and appeared then in order up to No. 1, No. 0, and No. 00, and then down from No. 6 to No. 8. Sbiaeiae No. 4, No. 3, and No. 2, as a rule, showed a greater abundance of the zoospores, judging from depth of the green color given to the solutions by them. 7 The indication is, that water somewhat less dense than that normally present in the lake at its present level is most favorable to development of Chlamydomonas sp. phanothece packardii does not grow well in the laboratory cul- tures. It was interesting to note that they lost their blue-green color and died in the weakest solutions first ; this condition follow- ing regularly up the series to the most dense solutions. This species gave us no further results. Whether this failure was due to the weak solutions being particularly unfavorable to the alga, or whether it merely indicates that this form is difficult to keep under laboratory conditions, is not certain. The latter seems the more likely conclusion. The diatoms recovered from the dense waters, on being trans- ferred to the weaker solutions in the laboratory, multiply read- ily and actually thrive, giving large masses of the characteristic brown growth. In every series, after about a month in the labo- ratory, solutions No. 1 and No. 2 show a very few live forms which soon die. In No. 3 a few persist; but in No. 4, No. 5, No. 6, and No. 7, they appear abundantly and continue to multiply indefinitely. In No. 8 the live plants are again not very numer- ous. These observations are in complete harmony with the state- ments? that the diatoms are found in great abundance in the shoaler, fresher waters near to the mouths of the rivers emptying into the lake. They are reported to be especially numerous on No. 608] SHORTER ARTICLES AND DISCUSSION 505 the alluvial fans at the mouths of both the Bear and the Jordan rivers.® The results seem to indicate that the diatoms obtained are true Salt Lake forms, but have become adapted to less severe condi- tions than prevail in the denser waters. That they are not fresh- water forms which have accidentally found their way into the- lake, is suggested by the fact that they do not thrive in the least dense of the solutions of any of the series. In every series, a cloudiness in the solution appeared as a re- sult of bacterial growth, but the order of appearance in every case was from the least dense solutions up to the most dense. This cloudiness soon disappeared, to reappear at irregular in- tervals. These facts led to an attempt to determine at least the number of species of bacteria which may be found in the part of the lake studied. So far as we can determine, no attention what- ever has been given this phase of the question in the past. Five distinct organisms, which have adapted themselves to conditions there, were isolated in pure cultures. No detailed study was made of them to determine their species, but enough was done to leave no doubt as to their being at least separate varieties, if one may judge from distinct differences in cultural and morphological characteristics. Water obtained from the lake under the strictest precautions, was at first plated on phosphorescent, or salt agar, which con- sists of 40 cc. of normal sodium hydroxide and 25 grams sodium chloride, to 1,000 ce. of plain agar. Later samples of the water were plated on gelatins containing different amounts of the normal NaOH, and NaCl. Better results were obtained with the salt agar than with the gelatin. Later, plain agar was used with good results. The number of bacteria per c.c. varies between 200 and 625, counts having been made from a number of samples taken in the coldest weather—water 33° F.—as well as in the warmer weather. A very interesting fact developed; that of the five micro- organisms isolated, three are decided chromogens, each produc- ing abundant pigment. Of the five, one is a diplococeus, which appears sometimes in tetrads and singly. It forms large white colonies on the media used. The other four are bacilli. The éne producing no pigment, forms delicate white colonies on the solid media. Of the chromogens, one produces a lemon-yellow; a second produces a bright orange; and the third produces a violet pigment. 506 THE AMERICAN NATURALIST [Vou. LI CONCLUSIONS 1. Variations in density of the water of Great Salt Lake, cause no corresponding variations in the size of Chlamydomonas sp. cells. 2. The indication is, that water somewhat less dense than that normally present in the lake, at its present level, is most favor- able to the development of Chlamydomonas sp. 3. The diatoms present in the lake multiply best in water much less dense than the dense water at Saltair. 4. At least four species of alge are to be found in the part of the lake investigated. 5. At least five varieties—possibly species—of bacteria have adapted themselves to the severe conditions in the lake. In conclusion, I wish to heartily thank Professor C. T. Vorhies for the suggestions he has given me in the preparation of this paper. L. L. DANIELS UNIVERSITY OF UTAH, SALT LAKE CITY NOTES AND LITERATURE BIOMETRIC STUDIES ON THE SOMATIC AND dire i PHYSIOLOGY OF THE SUGAR BEE THE beet sugar industry, amounting to hundreds of millions of dol- lars every year, is the direct result of scientific breeding. A biologist is loath to demur at any statement which attaches economic importance to scientific work of the kind in which he is interested. The statement may be true. Certainly no one can deny that far greater system and standardization of routine has obtained in the beet sugar industry than in many other branches of agriculture. But the trained scientific man who conscientiously works through some thousands of pages of the literature of sugar beet breeding and cultivation must hesitate before regarding it as a triumph of scientific method. He will rather, I think, feel that science has fallen woefully short of its possibilities in dealing with many problems of great theoretical interest and economic significance. In no field of agricultural work is the failure of scientifie and practical men to cooperate less excusable than in that of sugar- beet breeding. In the routine operations of sugar-beet produc- tion chemical data of a relatively satisfactory degree of trust- worthiness are obtained for great numbers of individuals. It is not unconservative to say that millions of individual weighings, polarizations or analyses of various degrees of completeness have been made. For two decades the biometric formule which might have given meaning to some of these masses of data have been available. Yet the problems which might have been solved have remained unelucidated, to the material loss of both biology and industry. It seems worth while to illustrate the truth of these state- ments by some of the advances in our knowledge of the genetic and somatic physiology of the sugar beet which have been made possible by the application of the biometric formule. Consider first of all one of the simplest problems—that of the relationship between the weight of the root and the sugar con- tent of its juice. Notwithstanding considerable discussion this 508 THE AMERICAN NATURALIST [Vou LI very simple problem was not definitely solved until 1913 when actual correlations’ were available. Coefficients ranging from — .224 to —.756 for the relationship between weight of root and sugar content of juice in various short series of data were found. The results were published only after failure in a con- scientious and systematic effort to obtain from the agricultural experiment stations really adequate series of data for detailed biometric analysis. Fortunately the conclusions have since been fully confirmed. Pritchard,? in dealing with samples of 250 to 400 beets grown at Fairfield, Washington, found constants ranging from — .2 to —.499. Working with larger samples (n==3.784) from Brookings, S. D., he found r——.258, while Harris and Hogensen,* who pad a sample of nearly 7,000 beetst from Utah cultures, found r= — .288. The splendid work of these investigators leaves no doubt that the percentage sugar content decreases, and, as Harris and Gortner indicated on their limited series of data, in a sensibly linear manner, with increase in weight of root. These studies, based as they are in some of these series, at least, upon closely bred material, fully justify the criticisms of the conclusions of Andrlik, Bartoš and Urban® by Hee and ` Gortner.® Harris and Gortner have also found negative correlations between weight of root and total solids and coefficient of purity. hus the larger roots have a smaller quantity of total solids, a lower percentage sugar content and a lower coefficient of purity. 1 Harris, J. Arthur, and R. A. Gortner, ‘‘On the Relationship between the Weight of ‘the Sugar Beet and the Composition of its Juice,’’ Jour, Ind. an ‘Eng. Chem., 5, March, 1913. 2 Pritchard, P. J., ** R between Morphological ert and the Saccharine Content of Sugar Beets,’’ Amer. Jour. Bot., 3: 361-376, 1916. 3 Harris, F. S., and J. - Hogensen, ‘‘Some Correlations in Sugar pas Genetics, 1: 334-347, 191 4 Unfortunately the aes is not altogether trustworthy because the largest and the smallest ma were excluded. 5 Andrlik, K., Cha wae N rban, ‘(Uber die Variabilität des Gewichtes und des reret der PAAA ATETEA rzeln, und über die AEE Beziehungen cage beiden Merkmale,’’ Zeitschr. f. Zucker- industrie in Böhmen, 36: 193, 6 Harris, J. Arthur, and R. A bik ‘‘ Further Notes on the Relation- ship between the Weight of the Sugar Beet and the Brg Ayr aps of its Juice,” Biochem. Bull., 2: 524-529, 1913. No. 608] NOTES AND LITERATURE 509 That large roots yield an actually larger amount of sugar is to be expected, and Pritchard’s coefficient for the correlations be- tween weight of root and total sugar content is high. Such results are obviously of great practical significance. Laying aside the possible desirability of modifying planting or cultivation in such a manner as to influence root size, the ques- tion of the selection of roots for sampling is one of real im- portance. Much of the early Acme tieas work on the sugar beet was devoted to determining where the crop will give a yield per acre and a sugar content and coefficient of purity satisfactory for economic work. Roots were sent by farmers to the Agricultural Experi- ment Stations, analyzed, and the results published in a great series of bulletins. But since size and percentage of sugar are correlated, and selection for size in the submitting of samples was rarely guarded against, the great mass of figures have little significance as measures of the merit of the cultures from which the samples were drawn. If physical characters of the root be associated with sugar content or with purity of the juice, which is technically a highly important factor, physical characters may serve as a guide to selection. Pritchard has devoted great care to the problem of the cor- relation between a number of the morphological features of the root and leaf and sugar content, and has determined the average percentage of sugar and average sugar content in synthetic types, i. e., those embodying the most desirable of the morpho- logical characteristics. The results are of technical rather than of general biological interest. The conceptions of a synthetic type—a conception that has already been emphasized in a quite different way by Raymond Pearl’—is well worth careful con- ` sideration by all those who have to do with breeding problems. Both Pritchard’ and Harris and Hogensen have extended their studies of the correlation of characters in the root to that of the interrelationship of the characters of the root and those of the fruiting shoots. 7 Pearl, R., and F. M. Surface, ‘‘Selection Index Numbers and their Use in Heed, AMER. NAT., 43: 385-400, 1909. Also, R. Pearl, ‘‘ Further Notes Regarding Selection Index Numbers,’’ AMER. NAT., 46: 302-307, 1912. : 8 Pritchard, F. J., ‘‘Some Recent Investigations in Sugar Beet Breed- ` ing,’’ Bot. Gaz., 62: 425-465, 1916. 510 THE AMERICAN NATURALIST [Vou. LI They agree that there is no correlation between the sugar content of a beet and the quantity of seed which is produced, but Harris and Hogensen find a correlation of .308 + .013 for the relationship between the weight of the seed beet planted and the weight of seed produced, whereas Pritchard, from a number of determinations, concludes that for beets of ordinary size such as are grown for factory use the correlation between root weight, percentage of sugar in roots and quantity of sugar in the seed root on the one hand and the number of grams of seed produced by the seed root is sensibly zero. Harris and Hogensen find a correlation of about + .399 be- tween height of plant and amount of seed produced, about + .277 between number of stems and weight of seed produced, and about -+ .122 between number of leaves and weight of seed produced. Pritchard has shown that there is no correlation between the amount of seed which a beet root yields and the sugar content of its progeny. ‘‘The application of this fact to sugar breed- ing,’’ says Pritchard, ‘‘is obvious, as extensive selection may be made for freer seed production without danger of sugar de- terioration. Moreover, it affords an opportunity to reverse the order of selection by making the chief elimination in the seed generation and thus greatly reduce the amount of chemical work and increase the effectiveness of the working funds.’’ The physiological character time required for maturing seed has received some attention by Harris and Hogensen, who find a greater height and a higher production of seed in beets re- quiring a longer period for maturity. The coefficients are, how- ever, low, r—=+.175 + .016 for height and days required for maturing seed and r= + .195+.016 for days required for maturity and quantity. of seed produced. The correlation be- tween the percentage of sugar in the mother beet and the number of days required for maturing seed is negative, r==— .129 + .014, i. e., the beets with higher sugar content mature their seed more rapidly. All these coefficients are very low. The experienced statis- tician will be cautious in regarding them as significant, remem- bering that when constants reach minimum values probable errors can not be given their normal weight. Those who have „had personal experience in the biological phases of such work will realize its diffieulties, and allow the questions of the signifi- eance of these correlations to remain open until more extensive ~~ No. 608] NOTES AND LITERATURE 511 data are available. There are, furthermore, internal evidences of serious heterogeneity in the materials upon which these con- stants are based. Such irregularities as those seen in the fre- quency distributions of number of days required for maturing require See before coefficients based upon them can be given much weight. The result of Pritchard’s experiments which will arouse the widest interest is the conclusion that with due regard to the probable errors of random sampling, there is no correlation be- tween the weight of the mother roots and the average weight, the average percentage sugar content or the average total sugar content of the progeny roots, that there is no correlation between the percentage of sugar in the mother beets and the average percentage of sugar in their progeny, between the actual amount of sugar in the mother beets and the actual amount of sugar in the progeny roots. Thus in dealing with our long selected varieties of sugar beets the author is faced to the conclusion: Differences in the size and sugar content of individual beet roots show no evidence of inheritance. They are fluctuations, therefore, and apparently play no part in beet improvement. The practical consequences of such a conclusion should be self evident. One European firm is said to carry out 300,000 analyses annually in the selection of roots for seed production. If the conclusion reached by Pritchard be of final significance, it justifies the assertion that ‘‘the cost of analyzing mother beets is an absolute waste of money.’’ Space precludes a discussion of the data given by Pritchard on the average composition of progeny rows and on the influence of environmental factors in observing genetic differences. From this side his paper must be read, and will later be reviewed in connection with one on the technical features of progeny tests.’ His studies show how small are the real genetic differences which may appear, how deeply these differences may be buried under those due to environmental factors, and how difficult in consequence must be the attainment of real progress in the further improvement of so highly selected an agricultural plant as the sugar beet. ° Pritchard, F. T., ‘‘The Use of Checks and ce e Singhal in Ye riety Tests,’’ Jour. Hadi Soc. Agron., 8: 65-81, 1916. 512 THE AMERICAN NATURALIST (Vou. LI Pritchard is a mutationist rather than a selectionist. The selection of choice roots by chemical and physical means has probably played no part in sugar beet improvement except where an occasional root has mutated and thus given rise to a superior physio- logical species. One does not need to agree with the form of Pritchard’s con- clusion to recognize the great value of such studies as those which he has carried out. Full knowledge of the difficulties surrounding a task is one of the essentials to its accomplishment. When all the variables that enter into the problem of sugar- beet production and sugar-beet breeding are known in quanti- tative terms, it will be possible for the practical man to decide on the basis of the cost of labor and other economic considera- tions what operations can be dispensed with and what other changes in routine can be profitably made. Operations can then be more properly designated scientific. J. ARTHUR Harris THE AMERICAN NATURALIST Vou. LI. September, 1917 No. 609 THE THEORY OF THE GENE. PROFESSOR T. H. MORGAN COLUMBIA UNIVERSITY Ir is unfortunate that the method of analysis of the problems of Mendelian heredity that has been adopted in one form or another by those who work in this field, has aroused a certain amount of antagonism on the part of those whose work lies in other directions. In the following pages I shall attempt to explain what the genetic factor means to those who use it, and then try to answer certain specifice criticisms of this form of hypothesis, in a hope that a mutual understanding will remove many of the objections that have been made to this method of handling genetic problems. The objections have taken various forms. It has been said, for instance, that the factorial interpretation is not physiological but only ‘‘static,’’? whereas all really scien- tific explanations are ‘‘dynamic.’’ It has been said that since the hypothesis does not deal with known chemical substances, it has no future before it, that it is merely a kind of symbolism. It has been said that it is not a real scientific hypothesis for it merely restates its facts as factors, and then by juggling with numbers pretends that it has explained something. It has been said that the or- ganism is a Whole and that to treat it as made up of little pieces is to miss the entire problem of ‘‘Organization.’’ It has been seriously argued that Mendelian phenomena are ‘‘unnatural,’’ and that they have nothing to do with 513 514 THE AMERICAN NATURALIST [ Vou. LI the normal process of heredity in evolution as exhibited by the bones of defunct mammals.’ It has been said that the hypothesis rests on discontinuous variation of char- acters, which does not exist. It is objected that the hypothesis assumes that genetic factors are fixed and stable in the same sense that atoms are stable, and that even a slight familiarity with living things shows that no such hard and fast lines exist in the organic world. These and other things have been said about the attempts that the students of Mendel’s law have made to work out their’ problems. I think, however, that while a few of these charges may appear to be serious, some of them rest on a misunder- standing of what numerical treatment of any problem in science means, and others are due to differences of def- inition. But the most common misunderstanding arises, I venture to think, from a confusion of the problem con- cerned with the sorting out of the hereditary materials (the genes) to the eggs and sperms, with the problems concerning the subsequent action of these genes in the development of the embryo. What genes stand for can be most easily shown by means of a few familiar illustrations. Mendel’s cross with yellow and green peas (or any similar case in which two characters are contrasted with each other as a pair) will serve as an example. In the second generation from such a cross the numerical results, viz., three yellow to one green, find their explanation on the assumption that the two original germ plasms (briefly the yellow and the green) or some element or elements in them separate cleanly in the germ cells of the hybrid of the first gen- eration. This cross does not tell us whether the two germ plasms separate as wholes—one from the other— or whether only some part or parts behave in this way. But the situation changes when two or more pairs of contrasted characters are involved in the same cross. 1 This objection is not further considered here since it has been dealt with elsewhere (‘‘ A Critique of the Theory of Evolution,’’ 1916, p. 84). No. 609] THE THEORY OF THE GENE 515 For example, when peas that are both yellow and round are crossed to peas that are both green and wrinkled, there appear in the second (F,) generation not only the original combinations, yellow round and green wrinkled, but also the recombinations yellow wrinkled and green round. Here also the numerical results, 9:3:3:1, can be explained by two assumptions, viz., that, as before, each pair of characters (or their representatives) are separated in the germ cells of the hybrid (F,) and that each pair ‘‘assorts’’ independently of the other pair. Ob- viously, here, it can no longer be the wholes of the original germ plasms that separate, for the two pairs of char- acters behave independently of each other; but there must be separate pairs of elements in the germ plasm that assort independently of one another. As a matter of fact it has been found that the many pairs of characters that follow Mendel’s law are inde- pendent of each other in inheritance. The only restriction that this statement calls for is in the case of linked pairs of characters of which I shall speak later. The germ plasm must, therefore, be made up of inde- pendent elements of some kind. It is these elements that we call genetic factors or more briefly genes. This evidence teaches us nothing further about the nature of the postulated genes, or of their location in the germ plasm. However, even if we postulated nothing more about them than their independence of each other and their distribution in the germ cells, we could still handle the Mendelian results on a purely mathematical -basis that would enable us to predict what new combina- tions should give. This possibility alone would entirely justify the hypothesis as a scientific procedure, whatever carping critics may say to the contrary. In fact Mendel himself did not carry his analysis beyond this point, for he assumed only that definite paired elements that stand in some way for the characters of the finished plant exist in the germ plasm, and that the pairs assort independently 516 THE AMERICAN NATURALIST [ Von. LI of each other at the time when the members of each pair separate (segregate). But between the year 1866, when Mendel published his paper, and the present year, 1917—an interval of fifty- one years—much water has run under the Mendelian mill. In consequence we can now add certain further attributes to the rather formal characterization of the gene as de- ducible from Mendel’s law alone. But before I discuss the evidence for these postulated attributes, I must pause for a moment to call attention to a movement that was in certain important respects a forerunner of our present standpoint. I refer to the views of Roux and of Weismann, both of whom assumed that the germ plasm is made up of par- ticles or determiners, as Bonnet, Spencer, Darwin and others had done before them. Their argument was largely speculative, and not of the same kind as the more recent evidence derived from Mendelian analysis. Moreover in all of Weismann’s earlier and best known writings his idea of-the units in heredity was more involved than are our present ideas. He thought that whole germ plasms were the units that segregated, germ plasms that dif- fered in one or many determiners, whereas the factorial view that we follow since Mendelism came to the front assumes that the units that segregate are themselves only parts of a whole which is the sum total of all the units. In his latest book, however, Weismann accepted the evi- dence from Mendelism and modified his ideas accordingly. We owe to Weismann the popularization of the view hat the hereditary material is carried by the chromatin, ut especially we owe to Weismann the development of the idea that the sorting out of the hereditary materials takes place at the time of the maturation process in the gg and sperm. On the other hand, it must be emphatically pointed out that the earlier idea of Roux, adopted by Weismann, that one of the hereditary complexes is sorted out during the No. 609] THE THEORY OF THE GENE 517 cleavage process of the egg, is no longer acceptable; for there is direct evidence to show that the whole hereditary complex goes to every cell in the body. This conclusion has the most. far- reaching consequences for our present views as to how factors produce their effects in the de- veloping organism, for it follows that the machinery that separates the inherited material into its component elements is not the same mechano-chemical process that brings about differentiation in the embryo. GENE AND CHARACTER So far I have spoken of the genetic factor as a unit in the germ plasm whose presence there is inferred from the character itself. Why, it may be asked, is it not simpler to deal with the characters themselves, as in fact Mendel did, rather than introduce an imaginary entity, the gene. There are several reasons why we need the conception of gene. Let me illustrate by examples :— 1. The Manifold Effects of Each Gene If we take almost any mutant race, such as white eyes. in Drosophila, we find that the white eye is only one of the characteristics that such a mutant race shows. In the present case the solubility of the yellow pigment of the body is also affected; the productivity of the indi- vidual also; and the viability is lower than in the wild fly. All of these peculiarities are found whenever the white eye emerges from a cross, and are not separable from the white eye condition. It follows that whatever it is in the germ plasm that produces white eyes, it also produces these other modifications as well, and modifies not only such ‘‘superficial’’ things as color, but also such ‘*fundamental’’ things as productivity and viability. Many examples of this manifold effect are known to stu- dents of heredity. It is perhaps not going too far to say that any change in the germ plasm may produce many kinds of effects on 518 THE AMERICAN NATURALIST [ Vou. LI the body. Clearly then the character that we choose to follow in any case is only the most conspicuous or (for us) the most convenient modification that is produced. Since, however, these effects always go together and can be explained by the assumption of a single unit difference in the germ plasm, this particular element or gene in the germ plasm is more significant than the character chosen as an index for one only of the effects. 2. The Variability of the Character is not due to the Corresponding Variability of the Gene All characters are variable, but there is at present abun- dant evidence to show that much of this variability is due to the external conditions that the embryo encounters during its development. Such differences as these are not transmitted in kind—they remain only so long as the environment that produces them remains. By inference the gene itself is stable, although the character varies; yet this point is very difficult to establish. The evidente is becoming stronger nevertheless that the germ plasm is relatively constant, while the character is variable. I shall consider this evidence in another connection. Here I wish merely to register some of the reasons why the idea of the gene is useful. 3. Characters that are Indistinguishable may be the Product of Different Genes We find, in experience, that we can not safely infer from the appearance of the character what gene is pro- ducing it. There are at least three white races of fowls produced by different genes. We can synthesize white- eyed flies that are somatically indistinguishable from the ordinary white-eyed race, yet they are the combined prod- uct of several known genes. The purple eye color of Drosophila is practically indistinguishable from the eye colors maroon and garnet. In a word we are led again to units in the germ plasm in our final analysis rather than to the appearance of a character. No. 609] THE THEORY OF THE GENE 519 4, Inference that Each Character is the Product of Many Genes We find that any one organ of the body (such as an eye, leg, wing) may appear under many forms in different mutant races as a result of changes of genes in the germ plasm. It is a fair inference, I think, that the normal units—the allelomorphs of the mutant genes—also affect the same part. By way of illustration I may state that we have found about 50 eye-color factors, 15 body-color factors, and at least 10 factors for length of wing in Drosophila. : If then, as I have said, it is a fair inference that the units in the wild fly that behave as Mendelian mates to the mutant genes also‘affect the organ in question, it fol- lows that many and perhaps a very large number of genes are involved in the production of each organ of the body. It might perhaps not be a very great exaggeration to say that every gene in the germ plasm affects every part of the body, or, in other words, that the whole germ plasm is instrumental in producing each and every part dy. Such a statement may seem at first hearing to amount almost to an abandonment of the particulate conception of heredity. But in reality it is only a conclusion based on fact. The essential point here is that even although each of the organs of the body may be largely dependent on the entire germ plasm for its development, yet this germ plasm is made up of independent pairs of units. 5. Evidence that Genes have a Real Basis in the Germ asm In 1906 Bateson, Saunders and Punnett found that cer- tain pairs of characters in sweet peas did not behave in- dependently of each other, but tended to stay together, or to keep apart, in succeeding generations according to the way they entered the cross. Every year more cases of linkage are found, so that there can be little doubt 520 THE AMERICAN NATURALIST [ Vor. Li that this phenomenon is one of the fundamental attributes of Mendelian inheritance While the linkage relations of genes do not at present have any immediate bearing on our conception of the nature of genes, they havea very important bearing on the problem of the localization of genes in the germ plasm. The original evidence that Weismann accepted to show that determiners are carried in the chromosomes, Viz., . the evidence based on transmission through the proto- plasm-free head of the spermatozoon, was made much stronger from Boveri’s evidence derived from experi- mental embryology. The argument became still more convincing when the facts of sex-linked inheritance and non-disjunction were established. For, it was found that certain characters have the same distribution as do the sex chromosomes, and secondly by the actual cytological demonstration that the rare exceptions to the rule are due to irregularities in the distribution of the sex chromosomes. _All of this evidence has played a rôle in persuading us that the genes postulated for Mendelian inheritance have a real basis and that they are located in the chromosomes. Finally, in Drosophila, where there are four pairs of chromosomes, there are also four great groups of linked genes. This coincidence adds one more link to the chain of evidence convincing a few of us that the gene in Men- delian inheritance has a real existence. CONSIDERATION OF CRITICISMS I have tried to make clear how the genetic evidence has necessitated the assumption of genes in heredity, and I have pointed out what seem to me to be some of the at- tributes that it has been desirable to add to the earlier conception of the gene as our knowledge has increased. Now that the ground is cleared, let me try to answer the objections or criticisms which I mentioned at the start, that have been advanced against this kind of hypothesis. No. 609] THE THEORY OF THE GENE 521 (A) Assumption of Genetic Factors is Arbitrary It hasbeen said that by assuming enough genetic factors you can explain anything. This is true; and it is the greatest danger of the factorial procedure. If, for example, whenever one fails to account for a result he introduces another factor to take care of what he can not explain he is not proving anything except that he is in- genious or only naive. To make good the introduction of another gene in Mendelian work, its presence must be established by the same kind of evidence as that on which the existence of the original factors was established. For example. Bridges found that after eosin eye color had been crossed to a certain red-eyed stock, there appeared in later generations a new class of eye color (Cream II) that was far lighter than eosin. He isolated this new | character and showed that the difference between it and eosin was due to a specifie gene that in inheritance behaves like other genes, although its action is not apparent on the normal red eye, but is evident on the eye color eosin. Here, then, through experimental tests, the actual demon- stration was made that the change in color of the eosin was due to another gene hidden in the normal stock. (B) Stability of Genes versus Instability of Characters It has been objected that it is unreasonable to assume that genes are relatively stable. This objection is based largely on the fact that characters are notoriously fluctu- ating, and since characters form the basis of our numer- jeal data from which the idea of the gene is derived, it is supposed that genes too must be variable. This is by all odds the most common criticism that has been brought against the idea of genetic factors and the most difficult one to disprove. There are five answers, however, to this s . Pii 8 objection. ; ` In the first place it has been shown in a number of instances that the variability of the character is due to a mixed or composite population in which there are sey- 522 THE AMERICAN NATURALIST [ Vou. LI -eral genotypes present. In other words, it was because most material is itself not uniform that an exaggerated idea arose concerning the nature of the variability of the character. In the second place, Johannsen’s experiments with Princess beans have shown that when the material is homogeneous in successive generations the variability of the character is due to the environment and is not due to changes in the genotype. In the third place, any pure stock (and especially one that has been made homozygous by inbreeding), so long as it does not vary, is an argument for the stability of the factorial basis. When changes occur in it as they are pretty certain to do, the fact does not in itself prove that the gene under observation has changed, for other genes that affect the character may have mutated. Jen- nings has recently said? that we maintain the constancy of a given gene by assuming that other genes, rather than the original gene itself, have changed. This would be of course on our part a straight evasion of the issue. The criticism would hold if the question involved were a purely philosophical one, as Jennings might unintentionally lead the reader to believe. Fortunately it is becoming more and more possible to demonstrate that changes of this latter kind do take place; for it is possible with suitable material to show in such cases the exact ndture of the change. Wherever it has been possible to do this it has been found that a definite mutation in some gene has taken place, or has been introduced into the culture through crossing. In the fourth place it has been found that more than one mutant gene may be the mate (allelomorph) of the same normal gene. Since no more than two of them may exist at the same time in a given individual, and since linkage experiments have shown in Drosophila that these multiple allelomorphs have the same linkage relations to all other genes (i. e., as we interpret the result, each such 2 Jour. Washington Acad. Science, VII, 1917. No. 609] THE THEORY OF THE GENE 523 set of allelomorphs has the same locus in the same chromo- some) this experimental evidence shows that several allelomorphs of the same gene may exist. An interesting relation in regard to these multiple allelomorphs is that they affect chiefly the same part of the body in the same general way. They may give a series of types that is discontinuous, such as the quadruple mouse series: yellow, gray white-belly, gray, black; or the more nearly continuous septuple Drosophila series: red, blood, cherry, eosin, buff, tinged, white. Whether such a series of characters is large enough to appear continuous or not is a matter of trivial importance in comparison with the es- tablished fact that the genes behind such a series arose in the same way as do other mutant genes, and after they have appeared, are as constant as are other genes. There is no experimental evidence to show that the multiple mutant alleélomorphs are more likely to arise from each other than they are from the normal allelomorph, and even if this should be true for individual genes it is no more than is true for other ‘‘normal’’ genes, some of which mutate more readily than others. Emerson has shown for corn that one allelomorph of a series is more likely to mutate than others and we have shown for Drosophila that certain normal genes, as the one that mutates to produce vermilion eye color, are more likely to mutate than are others. When Jennings? tries to interpret this evidence of con- tinuous series of allelomorphie characters as breaking down all real distinction between mutation and continuous variation, he leaves out of account certain very funda- mental considerations. For example, De Vries himself has always urged that mutations may be very small so far as the character change is concerned; the Svalöf evi- dence shows this in a very striking way, and Johann- sen’s beans have been for several years a classic ease illustrating how minute the characters depending on 3 Loc. cit. 7 : 524 THE AMERICAN NATURALIST [ Von. LI genetic differences may be. It comes, therefore, some- what as a surprise when Jennings states: ‘‘ Certain serious, difficulties appear in this view of the matter; I shall mention merely two of them, for their practical results. One is the very existence of the minutely differing strains, which forms one of the.main founda- tions of the genotype theory. How have these arisen? Not by large steps, not by saltations, for the differences between the strains go down to the very limits of de- tectability. On the saltation theory, Jordan’s view that these things were created separate at the beginning seems the only solution.’’ It should be remembered too that it is possible to make up just as continuous a series of characters with genes belonging to different allelomorphie pairs (even when they lie in different linked groups) as the continuous series from multiple allelomorphs. If there were any connections between the gradations of character in allelomorphic series and the order in which the characters appear, such a relation might appear to furnish a support to the view that the assumed fluctuation of factors is a sequential process, and that selection actu- ally helps forward the direction in which mutation is likely to take place, a view that Castle has at times ap- parently espoused. As a matter of fact, there is no such relation known—the known facts are exactly to the con- trary; for the actual evidence from multiple allelomorphs | shows that genes may mutate in all directions and also that extreme mutations such as white eyes arise sine me from red and not by graded steps In the fifth place, the most recent work on Didbbiphita has shown not only that every gene may act (and often does act) as a differential for characters conditioned by other genes, but also that there are genes whose most visible effect is only on certain characters which may therefore be said to be modified by the former. It would be a great mistake to suppose that these modifying genes are unique in any essential respect—the kinds of effects No. 609] THE THEORY OF THE GENE 525 that they produce grade off into effects that the ordinary genes produce. The chief interest in demonstrating (in- stead of speculating about) such genes is that they go far towards helping us to a clearer interpretation of certain evidence that was heretofore obscure or misinterpreted. Wherever the history of the origin of these genes is known it has been found to be the same as for other genes and their behavior in Mendelian inheritance is precisely the same. Nevertheless, Jennings has, in the paper al- ready referred to, left certain implications in regard to them that, if not clearly understood, may throw the sub- ject into worse confusion than before. He seems to imply —perhaps he does not really intend to do more—that since through such modifying genes a perfectly contin- uous series of modifications of a character may exist, all real distinction becomes lost between continuous and discontinuous variation. Now as a matter of fact per- fectly continuous characters, if due to overlapping of the separate modifications, can be statistically handled, as Johannsen has done for beans and as Jennings himself has done for size differences in paramecium. Other ways are also known by which the localization in the chromo- somes of modifying factors can be studied by methods that no student of Mendelian heredity can afford to re- ject. All of this is familiar, of course, to Jennings. He . means, however, to suggest that if the work on Drosophila continues for another fifty years, so many modifiers may be found that the characters will form a continuous series. But suppose the mutants do become so numerous that it is impossible to distinguish between any two by inspec- tion., Are we then to reject all the body of evidence that is fast accumulating that the modifying genes are ordi-_ nary Mendelian factors? It would be the height of ab- surdity to throw overboard all this experimentally deter- mined evidence as to the actual method of origin and Ll inheritance of these genes because a time may come when members of a series have become so numerous that we 4 See ‘‘Mechanism of Mendelian Heredity,’’ pp. 192-4. 526 THE AMERICAN NATURALIST [ Vou. LI will be too much bored to make the tests that will dis- tinguish a given new member from some one or other of the old ones. But Jennings may reply, suppose the selec- tionist claims that his material is already in this finely triturated condition! If, so, the answer is that by suitable selection experiments an analysis may in many situations still be made, and, secondly, the evidence, even from Castle’s rats, is far from establishing that he is dealing with such a sublimated process. On the contrary, there is much in them to indicate that they may be capable of being handled by rather simple Mendelian methods, as MacDowell has shown. As a matter of fact, when indistinguishable characters are the product of one or another modifier, the identifica- tion of the two genes involved, as independent, is per- fectly easy and certain by means of linkage relations. If a particular material is not sufficiently worked out to make this test possible, is that a sufficient reason why we should refuse to accept evidence where it can be obtained? And if there are indistinguishable characters that are the product of one or of another allelomorph, of course it can not be determined which allelomorph produces the result; but as, ex hypothesi, each allelomorph produces the same indistinguishable result, a discussion of such a question would be as profitable as the ancient one of the number of angels that can stand on the point of a needle. In conclusion then it may be said that by stable or onstant genes we do not necessarily mean that the gene is absolute in the sense that a molecule is absolute, for we | \/ean not know this at present. We might mean by stable Aai | genes that even if there is a variability of the gene the -variation takes place about a mode; and if in a given individual the extreme of variation was caused by a corre- _ sponding extreme in the variation of the gene, still the experimental evidence shows that in the many cell-genera- tions through which that individual’s germ cells pass to produce the sperms or the eggs, the genic variation, if there i is any, is still about the original mode and that no No. 609] THE THEORY OF THE GENE 527 new mode has been established unless a mutation has occurred. This latter interpretation is indeed in contra- diction to the idea that the gene is a single molecule, for molecules are not supposed to vary about a mode. A present either interpretation is compatible with the evi- dence, which does not discriminate between them. (C) Non-Contamination of Genes At the time when Darwin wrote and for many years afterward it was supposed that any new or unusual trait of character would become obliterated by repeated cross- ing with the normal or average individual of the species. This was perhaps the most serious difficulty that Dar- win’s theory of natural selection met with. It will be re- called that in order to overcome it Darwin made a con- cession that in principle amounted to an abandonment of the origin of characters through natural selection of chance variations. He admitted that only when a new character appeared in a large number of individuals at the same time was there an opportunity for its per- petuation. n sharp contrast to this earlier view, all the evidence from Mendelian heredity goes to show that however often ne EY ea a new character, that rests on a genetic change in the erm plasm, may have been kept out of sight by crossing to dominant individuals, whenever the character emerges from the cross, it shows at once that its gene has not been contaminated by contact with other genes. This conclu- -sion is an enormous gain for the theory of natural selec- tion based on chance variation, and at the same time is an equally strong argument to show that genes remain stable, and are not infected or mixed in the presence of other contrasting genes.” Let me illustrate by a case of my own. 5 Those who in their haste try to show that Darwin must have meant by fluctuating variation small mutations, since he assumed such fluctuations to inherited, might well ponder the difference between the two kinds of variation cited above. If Darwin had realized the difference referred to, he would not have had to make the damaging concession forced upon him by his critie, a professor of engineering, Fleeming Jenkin, in the North British Review (June, 1867). eRe 528 THE AMERICAN NATURALIST [ Vou. LI There is a mutant called ‘‘notch’’ (Fig. 1) character- ized by a serration at the ends of the wings. The factor that causes this is sex-linked, dominant in regard to the n Fic. 1. Notch female. wing character but recessive in its lethal effect. A female with notch wings carries the gene in one of her X-chromo- somes and the normal allelomorph in the other X-chromo- some. Half of her sons get the former and die, the other half get the latter X-chromosome and live. As there are no lethal bearing males, the females must in every gen- eration be bred to normal males. For twenty generations such matings have been made. Each time there have come together in the same female one X-chromosome carrying the gene for notch and its mate carrying the normal allelomorph. Selection of those females that No. 609] THE THEORY OF THE GENE 529 showed the least amount of notch, changed, after a few generations, the outward character of the notch stock so that at least half of those females that carry the notch gene came to have normal wings. It might have seemed that the gene itself had changed, possibly through con- tamination with the normal gene, were such a thing pos- sible. On the contrary, if these females with normal wings are outcrossed to a male of any other stock, all the daughters that carry the notch gene have the notch in the original (atavistic) condition, showing that the gene still acts in all its original strength. Moreover, suitable ex- periments have shown that as a result of selection, a modifying gene, already present in the original stock, has been isolated. This gene modifies notch (although it produces no visible effect on the normal wing) in such a way that the notch is less likely to appear. The evidence furnishes the twofold demonstration that the gene for notch has not changed through contamination, and that there is present a new and definite gene that does account for the change. (D) Methods of Inheritance that are not Mendelian It has been claimed that Mendelian inheritance is only one kind of inheritance and applies to only a limited group of characters. It has even been implied that the kind of characters involved in the process of evolution can not be inherited in this way because, it is affirmed, evolu- ee ee ee ee tionary characters are not like Mendelian characters. It is known that certain plastids, such as chloroplasts, that lie in the protoplasm are transmitted as a rule only through the egg protoplasm. There can be no doubt that this sort of transmission takes place. In principle it is not different from the transmission of certain kinds of bacterial diseases like that of pebrine in the egg of the silkworm moth. pe tambak. INg. 24001 9 No. E — | 2 Allie of St. ou Saks Pogis 3d E Lambert No. 512 9 H | Kathleen of St. Lambert NM s | s No. 28353 Q No. 2238 d'No. 1259 g É i Stoke Pogis = Stoke Pogis 3d 2 No. 3239 j Q May Day Stoke Marjoram wl Pogis No. 5109 2 No. 1066 F Z 3 May Day of St. Lord Lisgar < Fip ARO? |. Ioia ọ 8| zZ Jerne H| œ No. 15175 g'|No. 13656 F |No. 4558 3 a OO Ida's Ridter of OO Bachelor of St. Lambert 5 St. Lambert No Lambert" INo. 24991 Q|No. 2 E o o Ate of St. OL aka Pogis 3d g Lambert No. 5 Q 4 ou Kathleen of St. Lambert + 2 No. 43671 9 |No. 8388 d'|No. 6036 — g Š à e George of St. Lambert = Canada’s John Bull = No. 12968 9 Allie of St. Lam- Nymph of St. Lambert bert 2d š ts No. 24991 Q|No. 2238 2 Fa es Alle of 8i on Stoke Pogis 3d : : Lambert o| 6 2 |Z ons Q On: Taidan of St. Lambert. 4 Referred to the propositus, Letty’s Fancy Lady (241551). 550 THE AMERICAN NATURALIST TABLE II [ Vou. LI PEDIGREE OF LETTY’S Fancy (160320), Dam or LETTY’S FANCY Lapy (241551) Ancestral Generations 1 2 3 4 5 >œ | © No. 13657 d'No. 4558 o'iNo. 3143 rol F Bachelor of St: Oriont Lambert -i -INo. 6638 9 į Exile of St. Charity of St. Lambert 2 Lambert No, 24991 Q No. 2338 B4 4 a Allie of Si tnia O Stoke Foead 2 bert No. 5122 9 a O Kathleen of St. Lambert ° $ |No. 73475 Q No. 10481 g'|No. 6036 J SI © Sir George of St. Lambert mn Diana’s Rioter 5 No. 6636 © Diana of St. Lambert fan Letty Rioter 2 No. 48128 2 No. 10481 g a è ẹ Diana’s Rioter Dj + | Letty Coles 2d ais No. 23351 Q m A Letty Coles |. |No. 17408 g |No. 8388 J No. 6036 F 2 3 } 8 ó Canade's6 aa O Sir George of St. Lambert ia No. 12968 ` 9 O Nymph of St. Lambert _ |St. Lambert Boy No. 14880 9 No. 5248 g 8 le Lorne $ Oakland’s Nora ~ 2 No. 5123 g w Pet of St. Lambert 2 : ~_ 3 |No. 124201 9 /No. 17408 d'No. 8388 J = ës Sh Lenten O © Canada’s John Bull A Boy No. 14880 o Lady Letis O © Oakland’s Nora tT] 8 B Lambert" iNo. 48128 Q No. 10481 F N Age te Siz oe: Lotty Coles O © Diana’s Rioter gis vis No. 23351 9 O © Letty Coles With these data in hand we may proceed to the evalua- tion first of the total inbreeding. We have in Table III the pedigree elimination table for this purpose, which lists the primary reappearances indicated by solid circles. ` No. 609] STUDIES ON INBREEDING 551 TABLE III PEDIGREE ELIMINATION TABLE FOR THE TOTAL INBREEDING OF TY’s FANCY LADY Ancestral Generation in which Primary Name of Animal Primarily Reappearing SS ae 3 4 5 King of St. Lambert. heh Ge Gee VOLG o= 1 2 4 Allie of St. Lambert. ooon] -— — -— 2 4 Canadivé Jobn Palkia ai — — === 1 2 St. Lambert ey veri. eeren oe — -= 1 2 ety Cols fa eee eee | — — — 1 2 Bachelor 3 A RISO es okies ual eo ak bred a == 1 Biako FUNG I E E a ss ne ]-|— 1 Sir George pbo st e a A ee eo S a — | — — — 1 Haas A O e ee a a ri | — si ei i 1 E ee oe | 6 1 7 | 18 Whence, by the usual method, using the tables of Pearl and Miner,’ we have the following values: TOTAL INBREEDING COEFFICIENTS FoR Letty’s Fancy LADY Z,.=0, Z:=12.50, Z= 43.75, Z, = 56.25. Let us next consider Table IV, which gives the pedigree elimination for the pedigree of the sire, as given in Table I, considered by itself, the primary reappearances listed being those indicated by solid squares. It must be par- ticularly noted that the primary reappearances listed*in this table are referred to the ancestral generations of the pedigree of Letty’s Fancy Lady, and not to the pedigree of Rioter’s St. Lambert King, her sire, with whose pedi- gree we are dealing. TABLE IV PEDIGREE ELIMINATION TABLE FOR RIOTER’S St. LAMBERT KING Ancestral Generation® in which Primary Name of Animal Primarily Reappearing ee Curt 1 2 3 4 5 King oF St. Lambert. 2 i id — — 1 2 4 Allie of St. Lambert. EE T Gar — — 1 2 Sipe Poo oa ee ee es — = -— a 1 ORME eS NCS Se ey OO oe OA 1 3 7 5 Pearl, R., and Miner, J. R., Maine Agr, Expt. Stat. Ann. Rept. for 1913, pp. 191-202. ê Referred to the pedigree of Letty’s Fancy Lady. 552 THE AMERICAN NATURALIST [Von LE - In Table V exactly corresponding data are given for the pedigree of Letty’s Fancy, the dam of Letty’s Fancy Lady. The primary reappearances here are those indi- cated by solid diamonds in Table IT. | TABLE V PEDIGREE ELIMINATION TABLE FOR LETTY’S FANCY estral G tion® in which Primary Name of Animal Primarily Reappearing ES 1 2 3 4 5 St. Lambert Boy.. Se MES Ghee Recep nee -— 1 2 Letty Coles 2d.. UO cata E Oe apap — | — — 1 2 Dikawa hota. o e se — | — — — 1 Wiha oN E he rer Otho 0 2 5 Combined Totals of Tables IV and V ....... 0 0 1 hr ep I2 Difference between combined bara and | | totals of Table III (total inbreeding). Oe Boe | | 0 | 0 2 6 3 ~ COEFFICIENTS Pass “sL Bos Eae 7 £ 19 Z ‘ GENERATIONS $ È Fic. 1. Diagram showing the inbreeding and relationship curves for Letty’s Fancy Lady. Total inbreeding coefficients—solid line and circles; relationship ot . The smooth eding curves for continued brother x sister, parent x offspring, bay eee and double cousin x cousin mating. These are inserted for compari From the last line of Table V we deduce relationship coefficients as follows: No. 609] STUDIES ON INBREEDING 553 Ke, 2%: Rs 10x5 — 25.00, eR ae re — 37.50 Expressed in words these coefficients mean that Rioter’s St. Lambert King and Letty’s Lady are related to the amount or degree of 25 per cent. in the third, and 37.5 per cent. in the fourth ancestral generation. In Fig. 1 are shown the total inbreeding (solid line and dots), the relationship (dashes and open circles), and the partial inbreeding (dots and crosses) eurves for Letty’s Fancy Lady. Finally we have, from (iv), the ac coefficients of partial inbreeding due to relationship. ezna w an ay KZ, = 9A) — 28.57, K Zam poe = 83.83, We thus see that of the total inbreeding observed in the third ancestral generation of Letty’s Fancy Lady, none is due to relationship between her sire and dam; of that observed in the fourth ancestral generation, 28.57 is due to such relationship; and finally, of that observed in the fifth ancestral generation, one third arises because of re- lationship between sire and dam. 4, Let us next consider an example of measuring rela- tionship independently, altogether apart from considera- tion of inbreeding. We may take a very simple case af- forded by the two milking shorthorn cows, Imp. Milk 554 THE AMERICAN NATURALIST [ Vou. LI Maid 211032, and Imp. White Queen 545726. The pedi- grees of these animals follow in Tables VI and VII. The problem before us is to measure and express numerically the degree of relationship or kinship between these two animals. TABLE VI PEDIGREE OF Imp. MILK Marb (211032) 1 2 3 Ot se No. 409267 o'|No. 409193 a x Inspector wa as Morning Sun > No. — Q ao ? Bessie 44th iA] + 2 INo. Q No. 40909 g sy A Dainty Bean aea ee ee Tulip 28th z >i ulip = ó No. Q od Zi Tulip 23d = | œ No. 433648 F |No. 80356 eH a Arkin Beau b Border Stamp a E No. ——— [e] 4 White Sunshine x re So 9 3 È No. — 2 |No. 425402 F a o A Balmoral Pearl È oy Lady Balmoral } S BENS Z Z Lady Benedict’s Farewell We see that in these two pedigrees there is, in the first ancestral generation, one ancestor (Ireby Signet) which occurs in both. Hence we have ida he — 50.00. In the second generation there are three ancestors (Morning Sun, Tulip 28th and Border Stamp), which occur in both pedigrees, whence it follows that 100 (3 K,= wee) 75.00. In the third generation there are six common ancestors No. 609] STUDIES ON INBREEDING 555 TABLE VII ` PEDIGREE OF IMP. WHITE QUEEN (545726) Ancestral Generation | 1 | 2 3 Or boas No. 409267 g No. 409193 g z | ‘ Inspector a | Morning Sun ponte Meee 2 | „20 Bessie 44th | ae bom Ne. Q No. 409093 g el ek ae i] Dainty Bean Bal) age se Tulip 28th Ed 6 No. ———— Q ge rA Tulip 23d S| oO No. 433648 g No. 80356 g i Arkin Beau = z Border Stamp = S No. ——— Q & White Sunshine = 3 11 @ a No Q No. 501767 F ait Loa Levens Guardsman : Sayre = Diamond Queen SAPPY | No. ——— ọ zZ | A | Landford Diamond (all involved from the second ancestral generation) and ence K. == == 75.00. 100 (6) 8 So that we may say that Imp. Milk Maid and Imp. White Queen are 50 per cent. related in the first ancestral generation, and 75 per cent. in the second and third. This case will illustrate the superiority of the present exact numerical expression of relationship over the ordi- nary verbal expression. These two cows are half sisters, both having the same sire (this degree of relationship is indicated numerically always by K, = 50). But they are more closely related than two individuals which are only half sisters, because they have also one grandsire (Border Stamp) in common. Their total degree of relationship is simply not expressible verbally, by any term of kinship known to me in the English language. Yet by the method 556 THE AMERICAN NATURALIST [ Von. LI here described it is exactly expressible in the form K,=50, KiK 15. 5. It will be perceived that the form of relationship coefficient here proposed leads to precisely the same numerical results in simple pedigrees, with not too in- volved inbreeding or kinship, as that given in my former paper’ except for the fact that I have here changed the subscript designation of the K’s to bring them into con- formity with the total inbreeding coefficients. The earlier form proposed for these coefficients would always give the same numerical values as the present one if certain rather complicated rules of application, which were not clearly or rigorously set forth in the earlier paper, were to be followed. But the present simplified form does away en- tirely with the need for these complicated rules of pro- cedure. 6. It is of interest to set forth in tabular form the values of the relationship coefficients for the commonly recognized degrees of kinship. This is done in Table VIII, in which the different degrees of kinship are ar- ranged in descending order of closeness, in general. In some cases, as, for example, parent and offspring and half brothers (or half sisters), groups of two or three different sorts of kinship showing the same numerical degree of relationship should be regarded as bracketed, since there is no more reason for placing one of these first than another. - From this table a number of interesting points emerge. We note that the absolute maximum of closeness of re- lationship is that of brother and sister. The parent and offspring relationship is one half as close. Uncle and nephew (or niece), or single first cousins, are twice as closely related as grandparent and offspring. Some of these comparisons made obvious by the table may seem at first thought to give unexpected results, but if one will take the trouble to write down pedigrees for the stated 7 Pearl, R., AMER.: NAT., Vol. XLVIII, pp. 513-523, 1914. No. 609] STUDIES ON INBREEDING 557 degree of kinship, he will see upon careful consideration the reasonableness of the numerical result. TABLE VIII VALUES OF THE RELATIONSHIP COEFFICIENTS FOR VARIOUS DEGREES F KINSHIP Degree of Kinship | Kı | Kə Ky | K4 | Í Í | | Meee, “=a prore (or sister). . .-| 100 | 100 | 100; | 100 Par PR URA PBT TAE OER e O a Be Half-brother anh half-brother (or half-sister)....... _ 50 | 50 | 50 50 PGubie first CONA. fc. co ie el ee | 0 100 | 100 | 100 Single first cousins. TE ae | 50 | 50 | 50 Uncle and nephew (or niece) apes Oi. 30) 804i 7 BO Grandparent and off eg | 95 | 25 | 25 Quadruple second eousing..; -im cl. fea ae. [PO TO | 100 | 100 Double second Cousins. ss Sie ois ee 0 | 0 | 50 | 50 Single second cousins CVE SET Ee SRR ae 0 Qe 28 6 in once removed........... ee Do ae Oe P ropositus and first cousin twice removed . a 5 Ol 0] 0 | 12.6 . 7. There are two points in the development of relation- ship coefficients in this paper which may seem open to criticism. The first is that according to the definitions and formule of this paper, the degree of relationship be- tween two individuals is not affected by the number of times the same common ancestor occurs in the pedigree of either of the two individuals. The fact that such an- cestor occurs at least once in both pedigrees makes it a common ancestor. If it occurred more times it would not be a more common ancestor, because after all it would still be, all the time, just the same identical individual, made up of the same germ plasm. Put in another way, it is community of ancestry of two individuals which makes kinship. But the multiple appearance of the same indi- vidual in two pedigrees does not make any more ancestors common to the two related individuals than if this an- cestor occurred only once in each pedigree. Consider an individual A which is rather intensely inbred with refer- ence to an ancestor X. Consider another individual B which is also inbred to some extent with reference to the - Same individual X. Because they have a common ancestor X, A and B are related. But, according to the concep- 558 THE AMERICAN NATURALIST [ Vou. LI tion on which the present method of measuring kinship is based, the fact that A and B happen both to be inbred in respect to X, does not make them any more closely related to each other than if they were not so inbred. It may be of interest in this connection to point out, not as adding to the scientific exactitude of the position here taken, but as indicating what the common sense of men who have given thought to the subject of consanguinity has been, that the position here adopted that in determining degree of kinship a common ancestor counts but once as such, appears to be exactly in agreement with the position of both the canon law and the civil law on the same point. The second point in regard to which criticism might seem to be possible is the method of referring the in- breeding or relationship to the ancestral generations. In all of these Studies the inbreeding or relationship is re- ferred to the generation of the more remote (from the propositus) of the two appearances in a pedigree of a repeated ancestor. The logic of this procedure, rather than the reverse, is found in the circumstance that the fact of inbreeding (or kinship) does not establish itself until the more remote reappearance is reached. Thus it is impossible to know that a mating is of uncle and niece until the grandparental generation is reached. Ancestral generation... Fi ais the uncle of x, the common ancestors being b and c, but this fact is not known until the second ancestral generation is reached. The only logical method of rep- resenting these facts exactly in a numerical way would _ seem to be to say, in effect, that up to and including the first ancestral generation of a and «x there is no evidence that these individuals are at all related, and therefore No. 609] STUDIES ON INBREEDING 559 K,=0. In the second ancestral generation, on the con- trary, it appears that two ancestors, b and c, in the pedi- gree of x are the same individuals as appeared in the first ancestral generation of a. Therefore it now appears that a and x are related to the extent of 50 per cent. by the existence of community of ancestry in the second an- cestral generation. It would seem only logical to attach the numerical measure of relationship to the generation in which it is first proved to exist. Again, this is precisely the point of view regarding the matter which has been taken by the canon law and Roman civil law. These two points, which seem so obvious to the writer as to be difficult to discuss, are taken up here because correspondence has shown that they have been a source | of difficulty with some who have undertaken the study of inbreeding in domestic animals by the methods set forth in these studies. It is hoped that the simpler and more precise definitions of both inbreeding and relationship constants given in this paper may help to clear up such difficulties, which must arise, it would seem, from a lack - of a thorough grasp of the characteristics of pedigrees. SUMMARY In this paper the basic concepts of inbreeding are re- defined in a simple and rigorous manner, and on the basis of these definitions a new and more accurate method of measuring and expressing numerically the degree of kin- ship between any two individuals whatsoever, whose pedi- grees are known, is set forth and illustrated by examples. A new constant, the partial inbreeding index, is de- scribed. Its purpose is*to indicate numerically the part of the total inbreeding exhibited in the pedigree of any individual which is due to relationship between the sire and the dam of that individual. MULTIPLICATION. BY FISSION IN HOLOTHURIANS! DR. W. J. CROZIER THERE is to be found in various text-books the state- ment that certain pedate holothurians are capable of spontaneous transverse division, each part so formed producing a new individual (Lang, 1894, p. 1095; Morgan, 1901, p. 144). This opinion seems to be based, so far as I can learn, upon the observations of Dalyell (1851, p. 74; Pl. XIV), although Morgan says that ‘‘more recent observers have confirmed this discovery.” Chadwick (1891), also, found small individuals of Cucumaria planci to undergo self-division, and in one instance the posterior portion so formed also divided. These are records of division in holothurians which were being kept in small aquaria, and there has been no evidence, so far as I am aware, going to show that self-division of adult pedate holothurians is a method of propagation among these animals in their normal surroundings. Hence the possi- | bility of non-sexual reproduction in this way is usually stated with reserve (ef. Lang, 1894). In other classes of echinoderms (aside from echinoids) the expedient of reproduction by fission is of course not unusual; but in ophiuroids and in such starfishes as Linckia (Clark, 1913) and Coscinasterias (Crozier, 1914) ,? we are dealing with the division or fragmentation of a 1 Contributions from the Bermuda Biologtcal Station for Research, No. 66. 2 I have been able to secure further evidence regarding the fragmentation of Coscinasterias (Asterias) tenuispina (Lam.), which proves conclusively that the rules previously deduced (Crozier, 1915a@) regarding this process are indeed valid. This evidence will be published in connection with a de- scription of experiments on the direction of progression in Coscinasterias. The presence of great variation in ray-length, as well as of a variable num- ber of madreporites, gives an opportunity to test out in this species the validity of certain ideas concerning ‘‘ physiological polarization” in as- teroids. 560 No. 609] MULTIPLICATION IN HOLOTHURIANS 561 many-armed creature relatively deficient in morpholog- ical centralization; whereas in Cucumaria and in Holo- thuria the body is compactly built, the animal much more of a unified individual. Consequently the self-division of these holothurians is not without interest, especially since in these cases the plane of separation is anatomically per- pendicular to that employed among the astroradiates, and it is the purpose of this paper to present evidence which proves that adult specimens of at least one species, Holo- thuria surinamensis Ludw., do as a matter of fact divide transversely into two parts under conditions which must be regarded as normal. Since these divisions are not in- frequent in large numbers of specimens, if not in a single life-history, we must conclude that fission represents a regular means of mutliplication in this species. A few years ago I found (Crozier, 1915b) that about 10 per cent. of the examples of H. surinamensis which were studied showed a condition of either the oral or of the cloacal end which—on the basis of observed regen- erations following experimental eutting—I interpreted as representing regeneration, possibly as a consequence of spontaneous self-division. Similar conditions have been noted by others for some other species of holo- thurians, e. g., by Benham (1912, p. 136) for Actinopyga (Miilleria) parvula (Selenka), but they have usually been referred to regeneration after injury by such bottom feed- ing fishes as small sharks. Dr. H. L. Clark informs me that he has found a corresponding state of affairs in some Australasian holothurians, at least in regard to the oc- currence in nature of specimens showing posterior re- generation. I subsequently obtained young individuals of H. cap- tiva Ludw., about 6mm. in length, which were observed to divide spontaneously in the laboratory (Crozier, 1914, p. 18), precisely according to the procedure figured by Dalyell (1851) and by Chadwick (1891). Only a single adult H. captiva has been discovered, however, in which there was evidence of normal regeneration; this indi- 562 THE AMERICAN NATURALIST [Vou. LI vidual, which was 40mm. long, was obtained among a group of 47 taken from under a large rock in April, 1916.% The anterior end for a distance of 7 mm. from the tip was light greenish yellow, with ten very feebly developed tentacles; there was a sharp line of demarcation between the light yellow surface and the dark olive green of the rest of the body. If H. captiva undergoes division nor- mally, it can only occur in very young stages. With H. surinamensis, however, the case is quite dif- ferent. In Table I there are summarized results of the examination of several series of these animals collected ai different times for this particular purpose. It will noted that in these collections from 2.5 to 16.9 per su (on the average about 11 per cent.) of the individuals show a condition of either the oral or of the cloacal end which is interpreted as representing regeneration. This seems to be about the proportion of such instances which is to be met with in general collecting, although numerical records have been kept only in the cases cited. The specimens represented in the tabulation were obtained TABLE I THE RELATIVE NUMBER OF CASES IN WHICH Holothuria surinamensis WAS FOUND TO BE REGENERATING IN NATURE bie Wo wi Regenerating zy weit Oral Cloacal Total % June, July, 1913. 200 z 13 20 10.0 July 30, 1916 ......... aa 39 0 1 1 2.5 n SO LIO... isser 84 6 5 ki 13.1 ve A! Bi GPR iio ee 70 6 2 9 12.8 Jan. 31, 1917.. „i 4 5 9 16.9 Totala E coleaki eee 446 23 27 50 11.2 Ratio of regenerating oral to cloacal wi: Iar: from one locality, Fairyland Creek, where they notably abound; but the species has been collected at many other stations, where also the regenerating individuals are to be found in approximately the same proportion. The 3 It is known that some holothurians tend to congregate together in con- siderable numbers at their time of breeding. Graber considered this to indicate the presence of a chemical sense (Delage et Hérouard, 1903). No. 609] MULTIPLICATION IN HOLOTHURIANS 563 season of the year seems not to influence the occurrence of regenerating specimens. The criterion of regeneration in these cases consists in the presence of an anterior or posterior terminal part of the body characteristically different in appearance from the normal buccal or cloacal end, the surface being clearly marked off from that of the rest of the body. In typical examples these regenerated ends of the animal are more sharply pointed than is usual; they bear feebly developed tentacles (at the anterior end), tube feet, and dorsal papille, which are less reactive than those on animals judged to be not regenerated; and these appendages are very lightly pigmented (Fig. 1). These characteristics a b Fic. 1. Regeneration found occurring naturally. at atk ae sketches, showing differences in pigmentation of a, cloacal end, b, oral end. Natural size. undoubtedly become less prominent with time, the colora- tion tending, however, to remain pale on the ventral sur- face (trivium). The first pigment to appear is the green fluorescent one (cf. Crozier, 1914, p. 9 and 1915b); the dark brown substance develops more slowly, just-as in the growth of the pest larval holothurians of this and re- lated species. The spicules of the podia and skin seem fewer than in corresponding non-regenerating parts, but are of the usual sizes and shapes. The tentacles on re- generating buccal ends are always fewer (9-15) than on the normal individual (20). 564 THE AMERICAN NATURALIST [Vor. LI È That the oral and cloacal terminations just described do in reality represent regeneration, has been verified by observation of the course of regeneration in the labora- tory after experimentally cutting the holothurians ün various ways (cf. Crozier, 1915b). Certain specimens also have been tabulated as ‘‘regenerating’’ when their appearance (Fig. 2, c), backed up by dissection, suggested that they had just completed division and had not yet begun to regenerate. These specimens lacked either a cloaca, or the stone ring and buceal structures, depending, obviously, on their former relation’ to the complete indi- vidual from which they were derived. at the division area is also indicat The evidence that the regeneration found occurring under natural conditions results from the self-division of adult holothurians, involves two considerations. The first concerns (a) the relative size of the regenerating animals, and (b) the relative frequency of anterior and posterior ends noted as regenerating. The second has to _do with direct observations of self-division. No. 609] MULTIPLICATION IN HOLOTHURIANS 565 One hundred H. surinamensis collected in Fairyland Creek ranged in length from 6 to 18 cm., with the mode at 14cm. The regenerating specimens ranged in length from 4 to 9 cm., with the mode at 7em. While no numer- ically exact argument can be based on these figures, since the length of any one holothurian is variable, the fact does stand out that the regenerating animals are about one half the length of the non-regenerating ones. There is also the significant fact that not a single instance has been found in which both a (supposedly) new oral and new cloacal end were present. If self-division has occurred, then we should expect to find new oral and new cloacal extremities in equal frequency; among the rather small number of cases available, we find their ratio to be as 1:1.17, an agreement sufficiently close to favor belief in self-division. The evidence concerning the second point is even more conclusive. I have seen, in all, nine cases in which a holo- thurian (H. surinamensis), in the laboratory, divided itself into two parts. The animals concerned seemed healthy, and bore no visible signs of having been in any way injured. In no case did the halves so formed redi- vide, although in two cases the resulting portions lived in the laboratory for a month (Aug. 3 to Sept. 5, 1916), during which time, even in the absence of food, missing structures were regenerated. In one case the process of division occupied five days; in another, twenty-four hours. Probably it is executed more rapidly in the field. The details of the division were not notably different from those described by Chadwick (1891) for Cucumaria, except perhaps in one particular. The intestine is not drawn out between the separated halves, as found in Cucumaria and as I have observed in the young of H. captiva. Division begins midway of the length of the body with a deep insinking of the ‘‘dor- sal” bivium. A powerful circular constriction, accom- panied by some slight local disintegration of the integu- ment, completes the separation (Fig. 2). During the progress of division the animal is quiescent, although it 566 THE AMERICAN NATURALIST [ Vou. LI may be adhering firmly by its tube feet to the vertical wall of the aquarium. When the constriction and separation of the skin- and muscle-layers is completed, a short length of the intestine usually remains for a time connecting the two pieces; it may rupture close to one of them, or may disintegrate completely. The point to be noted is, that the resultants of the division do not move apart, but re- main quiescent. Lastly, after considerable searching, I found in the field one case in which division had evidently just been completed. The halves were still joined by an exposed portion of the gut On the basis of all this evidence there is certainly reason to believe that Holothuria surinamensis, in the adult state, normally multiplies. its numbers by a process of binary fission. The resulting organisms readily complete their missing parts, but probably do not undergo a second divi- sion until after the lapse of a considerable interval, if they do at all. It would be of some interest to determine the nature of the sexual products in the animals which thus result from ivision. AGAR’S ISLAND, BERMUDA REFERENCES Benham, [W. B.] 1912, Report on Sundry Invertebrates from oe Kermadec Islands. Trans. N. Zeal. Inst., Wellington, Vol. 44, pp. 135-138. Chadwick, H. C. 1891. No ae on Cucumaria planci. Trans. Liverp. Biol. 0C., pp. 81-82, Clark, H. L. 1913. pean in Lacon. Zool. Anz., Bd. 42, pp. 156-159. Crozier, W. J. 1914. The Orientation of a Holothurian by Light. Amer. Jour. Physiol., Vol. 36, p 0. 1915a. On the Number of Rays in cool tenuispina Lamk. at Ber- muda. AMER. NAT., Vol. 49, 36. 19156. The Sensory Reactions of Holothuria bE Ludwig. ete., Vol, 1, London, 286 pp., 70 i Delage, Y., et Hérouard, E. 1903. Les échinodermes. Traité de Zool. concrète, T. 3, x + 495 pp., 565 figs., 53 pls. Lang, A.. 1894. Lehrbuch der TENNA Anatomie der wirbellosen Thiere. xiv + 1197 pp., 854 Fig. Jen Morgan, T. H. 1901. Regeneration. Sets: Univ. Biol. Ser., VII, xii + 316 pp., 66 figs. SHORTER ARTICLES AND DISCUSSION AN ATTEMPT TO MODIFY THE GERM PLASM OF (ENOTHERA THROUGH THE GERMINATING SEED WHEN a new character appears in a homozygous race or species it may be either a mutation or an acquired character. If a mutation it has been produced because the germ plasm has in some way been affected and the succeeding generations may be expected to show the same variation. If an acquired char- acter it will be present for a single generation and is then lost unless the cause that produced the somatic change also modified the germ plasm in such a manner that it may develop the same character, in succeeding generations. By ‘‘acquired characters’’ is meant any and all changes that are wrought in the soma of the organism by the environment considered in its broadest sense. It is the creed of modern biology that acquired characters are not inherited unless the environmental influences also play on the germ cells even while focused on the body tissues, pro- ducing at the same time potential alterations in the former and visible changes in the latter. The causes of mutation are in dispute. It is fairly obvious, however, that in the reported instances of mutation the varia- tions arose as the result of changes in the germ plasm. A muta- tion, therefore, has its origin from within and this origin has no very evident connection in any way with external condi- tions. If a case of an acquired character is shown to be in- herited it is clear that the germ plasm of the organism must — have been affected. The stimulus to change, therefore, in con- trast with the cause of mutation, would have come from without. While the end result—the alteration of the germinal constitu- tion—is the same in both cases, the method or cause by which it is brought about is different. From this point of view, the relative value in evolution of mutation and of hereditary acquired characters is open to various interpretations. I think that in the case of mutations it may soon be possible to demon- strate that some of the so-called examples of ‘‘mutations’’ are due to or are associated with irregularities of karyokinesis. It is not at all inconceivable that outside conditions producing ac- 568 THE AMERICAN NATURALIST [ Vou. LI quired characters may at the same time effect chromosome struc- ture and behavior in the germ plasm. The cause for unusual activity on the part of the chromosomes in either case may, how- ever, well be totally different. : Interesting as are the effects of natural influences on the germ plasm, of greatest importance is the problem whether or not modifications may be produced and controlled artificially. The fact that experimental research is tending to show the specificity of certain chemicals for various organisms or parts of organisms makes hopeful the outlook for finding a specific or many specifics that will act upon the mechanism of heredity or different parts of that mechanism. At present our investigations can only be empirical, trying this or that method more or less blindly. Con- siderable work in this field has already been done. Without entering into a full discussion of the numerous investigations, it may be said that alcohol, temperature, humidity, ether, zine sulphate and radium have been used in attempts to alter the germinal constitution. Changes indeed have been produced in some cases but the effects have been generally physiological in nature either interfering with development or influencing color, length of hair, ete. In all cases where the offspring re- sembles the altered parents the results have been readily inter- preted on the assumption that eggs or embryos are influenced at the same time as were the forms producing them. MacDougal reports having produced changes in one of. the cnotheras through ovarial injections. Although it may be possible that he produced modifications, the plant he selected for experimenta- tion was unfortunate since the natural variability of the ceno- theras is in most cases great, and the gametic purity of his material was not clearly demonstrated. In many ways plants offer the most favorable material for the study of experimental variation. Many parts of the plant are adapted to experimentation. The pollen may be subjected to treatment, the ovaries may be injected with chemicals or otherwise handled, the seeds offer themselves to various manipu- lations, many stages of growing plants are available for experi- mentation, a variety of experiments may be conducted and, if there are numerous branches, controls may be maintained on the same plant.. Coupled with these advantages are possibilities of cultivating pure lines for many generations. It would be difficult indeed to find an animal about which all of the above statements could be made. No. 609] SHORTER ARTICLES AND DISCUSSION 569 Two methods of inducing germinal variation seem practical. The first is the treatment of the germ cells. MacDougal’s ovarial injections lie in this class. The other method is to attack the buds or growing points of the stems. I do not at present believe that exact results can be expected from ovarial injec- tions since the ovules are generally so tightly packed that fluids can not readily circulate and it is impossible to know the quan- tity and the strength of the fluid that may reach the germ plasm. The treatment of pollen may offer greater possibilities unless unforeseen technical difficulties are encountered. By subject- ing growing points to treatment the germ plasm may more readily be affected and the complete chromatin equipment may be placed under the influence of the materials used. If germinat- ing seeds or seedlings are immersed in a given chemical solution we may have reason to expect that the cells of the growing points are probably in contact with the solution or some derivative of it. If the chromosomes of a growing point can be influenced it is pos- sible that the organs that are developed from this point may be altered. Two years ago, in connection with a program of study deal- ing in general with problems of development, I had the oppor- tunity of making some experiments on seeds and seedlings of nothera biennis L. (the Dutch biennis) from a pedigreed line which had been inbred for at least eight generations. Mno- thera biennis L. is one of the most stable species so far studied in this genus and its rare ‘‘mutations’’ are known from the research of Professor DeVries. Consequently it seemed justi- fiable to anticipate that any results obtained through experi- mental treatment may readily be recognized. To Dr. Bradley M. Davis I am indebted for suggesting (nothera biennis as a favorable plant for study, for many ‘pedigreed seeds and for the complete freedom of his garden. Without his aid in the study of the plants and without his advice I should have made little progress. In preliminary studies of this sort all the in- investigator cah hope to accomplish is to determine certain solu- tions which produce suggestive effects. The results of my studies are here brought together in hope that they may be of some help to other workers. _ As has been pointed out, studies of this character must at present be largely empirical. The problem is to find chemicals that will modify the structure of the germ plasm or bring about irregularities in the distribution of the chromosomes. Some of 570 THE AMERICAN NATURALIST [ Vou. LI the chemicals used are those frequently employed in fixing fluids; others were selected for various reasons. As shown in the List of Experiments the seeds and seedlings were soaked for varying lengths of time in the solutions. The material was either thoroughly washed before being placed on moist filter paper in petri dishes to complete germination or it was placed on paper which had been moistened with the same solution as that in which the material had been soaked. Since the seeds of Gnothera biennis are about 96 per cent. viable, seed sterility was not an important factor in the results of the experiments. Dr. Davis’s large cultures under normal conditions were used as controls. About one hundred seeds were used in each experiment. LIST oF EXPERIMENTS Fluids Percentage beaa Prante pesbies sooo BOOUG ME Ce Ae eee 0.125 | Seeds 4 days 25 0.125 Seedlings| 3 S 0.625 | Seeds 29 days 40 Butyee Held: oo eke oe eae ee Seeds 34 days 40 ; 0.5 Seeds 34 days 0 1.0 Seedlings| 1 day 1.0 Seeds 5 days 0 Cherm hydrate... A FS 0.75 Seedlings) 5 days 0.75 Seeds 34 days 50 0.375 s 24 days 0 0.187 | Seeds 24 days 50 Chrotic tod. seauu A 0.03 Seedlings | 11 hours ° 0.015 | Seeds 4 days 65 0.015 s 3 days 60 0.015 | Seedlings 21 days 0.015 s ays 60 Ethyl sobol.. oia a a 5.0 Seedlings| 8 days 5.0 Seeds 27 days 0 1.0 Seeds 18 days 40 0.5 Seeds s 60 0.5 Seeds 24 days 80 0.25 Seeds 24 days |A few germi- nations Mothyi-aloghol i o. Aa aa 1.0 Seeds 18 days Normal ger- mination Amylie MOURO. 0 OSS a 1.0 Seeds 18 days 0 Butylic alobal. ieo eis. Gas 1.0 Seeds 18 days 0 Propylic MOOD. ana ea es 1.0 Seeds 18 days 0 mine sulphate, i 0 hie es 10.0 Seeds 18 days eas 5.0 Seeds 18 days (A few germi- nations Stryhn. a ssa es Seeds 10 days Pot. bromide and iodide......... Seeds 10 days 0 POMC MN Ce or a as 4.0 Seeds 10 days 0 0.5 Seeds 14 days 19 No. 609] SHORTER ARTICLES AND DISCUSSION 571 RESULTS In the acetic-acid solutions mold grew vigorously and possibly interfered with the growing plants. The percentage of germina- tion was low. In the young plants the cotyledons were rather more pointed than normal, although this modification was not marked. The leaves of young rosettes also appeared more nar- row and pointed, but these peculiarities disappeared as the plants matured. All the seeds and seedlings treated with butyric acid died. Chloral hydrate produced no effect other than retarding the period of germination, reducing its percentage, and weakening the plants. Chromic acid produced by far the most interesting results. In the various solutions used germination was prompt (about the usual three to four day period) but the percentage was materially lowered. The seedlings produced were vigorous in appearance, although the root system was in most cases stunted. There was a slight though not, I believe, significant modification of the cotyledons which were somewhat less pointed than in the type. Some of the seedlings were bright red and practically all had a reddish or pinkish tinge. Growth after planting was slow but all the plants finally developed normally. Ethyl alcohol produced no modification of structure, although I believe that it will be worth while to continue this line of ex- perimentation. In all cases where seeds were allowed to soak in ethyl alcohol the solution became thick and gelatinous from a substance extracted from the seeds. The percentage of germina- tion was much reduced. Methyl alcohol retarded germination but the resulting plants were fairly normal. Amylic, butylic and propylie alcohols all inhibited germina- tion in the strengths employed. _ Germination was also inhibited by the solutions of zine sulphate, strychnine, potassium bromide and iodide and by four per cent. ferric alum. In general it may be said that the treatment of seeds and seed- lings in the experiments has resulted, as in the experiments of others, in reducing the percentage of germination or in a general weakening of the plants rather than in specifically modifying the germinal constitution. The results from the experiments with chromic acid and possibly with chloral hydrate and Sia alcohol suggest the desirability of further studies. In futu Ea +. 572 THE AMERICAN NATURALIST [ Von. LI work the concentration of the agent and the length of treatment should be studied in greater detail. Rosert T. Hance BOTANICAL AND ZOOLOGICAL LABORATORIES, UNIVERSITY OF PENNSYLVANIA, June, 1917 CONCERNING A MORPHOLOGICAL PREDICTION FROM DISTRIBUTIONAL DATA AND ITS SUBSEQUENT VERIFICATION? : ALL species of the genus Salpa are notable for their two alter- nating generations. One, known as the ‘‘solitary’’ generation, produces offspring by budding and the buds, when fully mature, constitute what is known as the ‘‘aggregate’’ generation, or, as I shall call them, zooids. Each zooid produces, sexually, one of the solitary generation, the embryo being nourished and carried within the body of the zooid until after it has begun to form zooids of the succeeding generation. The relation between the two generations, however, is not simple; and as there seems to be a widespread misconception re- garding this relation, it is well to correct it. One sees the statement in much of the literature that the sol- itary generation is asexual, that the aggregate generation is sexual (hermaphroditic), and that the developing embryo of the solitary form is carried within the body of its mother. Such, however, is not the ease; Brooks having clearly demonstrated in his classical volume ‘‘The Genus Salpa,’’ that the solitary form is, in the most literal sense, a female while the zooid it produces by budding is a male. In brief, the curious sexual relationships are as follows: The solitary form, or female, produces immature males by budding, within the body of each of which the mother tucks away one fully developed and ripened egg together with its fol- licle. Before fertilization the egg is suspended by means of a fertilizing duct, which opens into the cloaca, into one of the blood channels of the newly formed zooid. The spermatozoa which are _ drawn into the pharynx of the zooid with the sea-water, are swept past this opening by the contraction of the muscles in swimming, and some of them enter, one of which penetrates to the egg and _ fertilizes it. The embryo, at an early stage, pushes into the 1A paper read before the Western Society of Naturalists, Friday, April 6, 1917, at Stanford University. No. 609] SHORTER ARTICLES AND DISCUSSION 573 cloaca, carrying its wall before it, thus becoming inclosed in an epithelial capsule. This capsule is soon cast off and, in its place, a placenta is formed which, communicating with the blood- channels of the zooid, nourishes the growing embryo until birth. While the embryo projects into the cloaca it is not, at first, ex- posed to the water, but is enclosed by an epithelial capsule and, after its disappearance, by an embryo sae resembling somewhat the amnion of vertebrates. Later, but still while very young, this sac is distended and finally broken by the growing embryo and from then until birth, the embryo is directly exposed to the water in the cloaca, being fastened to the zooid only by the pla- centa and a narrow band of ectoderm which connects the neck of the placenta to the walls of the cloaca. After the embryo has reached maturity and made its escape into the water, the testis matures and the zooid becomes a fully developed male. Vir- tually, then, the mother salpa gives rise, by budding, to its sons; and each son serves first as a depository for one of its mother’s unfertilized eggs, then as a living incubator within which its mother’s daughter is housed and reared, and lastly as a father. Such is, in brief, the life-cycle of the Salpe. In Salpa democratica, to which the distributional data to fol- low relate, and which is the tiniest of all the Salpæ, the mech- anism of budding is as follows: Within the body of the solitary salpa, at its caudo-ventral extremity, lies a compact intestinal tract called the nucleus. Closely associated with it is an organ, ‘‘the proliferous stolon,’’ which by segmenting distally gives rise to the zooids. The zooids are budded off rapidly and no evi- dence is at hand that the stolon ever exhausts its capacity to pro- duce them. The zooids remain attached to each other and to the stolon in the form of a chain and, as segmentation advances, they become crowded and pushed along a spiral path encircling the nucleus. The stolon undergoes more or less regular periods of segmentation and rest so that the zooids are developed in sets or blocks, all individuals in a single block being of approximately the same size and in the same stage of development. Each block contains from forty to sixty-five zooids, and as there are from three to four blocks present when the distal end of the chain is _ready to emerge from the test-cavity to the exterior, the stolon carries in the neighborhood of two hundred zooids at one time. In other species the chain remains intact and attached to the stolon after, as well as before, the zooids are protruded into the * 574 THE AMERICAN NATURALIST [ Vou. LI water, but no evidence of this relative to S. democratica is re- corded in the literature. During June and July, 1908 and 1909, thirty surface-net hauls were made in the vicinity of San Diego when the temperature of the water was between 15°.9 and 18°.3 C., and forty-six when it was between 18°.4 and 20°.8 C. Solitary forms were captured in greater numbers per unit volume of water filtered from the warmer water, and zooids from the colder water. But one or more of both solitary forms and zooids were captured from a larger per cent. of these unit volumes in the colder than in the warmer water. This is better shown by the following table: TABLE SHOWING SURFACE DISTRIBUTION OF S. democratica RELATIVE TO M AND COLD WATER Pancake te | No. of | _ Solitary Forms Zooids See eer eee Ee eg Abund.! | Freq? EOE en Be Fe wey o o wo oo E 4-20.8..........| 46° | 160 | 73 t27 l 80 These data show that, while zooids occurred in greatest num- bers in the places they frequented the most as would be expected, solitary forms occurred in greatest numbers in the places they frequented the least, or, to state it differently, they were found most often where they occurred in smallest numbers. What does this apparent paradox mean ? Believing that these relations must have been consequent upon a freak result of chance or random sampling, and discovering that only 9 of the 30 cold-water hauls were made during the night 6 p.M.—6 A.M.) while 33 of the 46 warm-water hauls were made at night, the day hauls and night hauls were separated, and the data concerning each were retabulated with respect to the same two groups of temperatures. The results are given in the table on p: 575. Be it noted that, although the magnitudes of the differences vary, the directions of the differences are exactly the same as revealed by the data in the above table, i. e., solitary forms are most abundant and least frequent in the warmer water, while 7 zooids are most abundant and most frequent in the colder water. = Still being in the position of an unbeliever, the data were re- 1 Abundance or number of individuals per unit volume of water filtered. 2 Frequency or per cent. of unit volumes containing at least one individual. No. 609] SHORTER ARTICLES AND DISCUSSION 575 tabulated in every practicable manner: first, by eliminating all hauls except those made between 6 and 10 a.M.; second, by elim- inating all except those made between 10 a.m. and 2 P.m.; third, by eliminating all except those made between 6 and 10 P.M.; fourth, by eliminating all except those made during June, 1909, and fifth, by eliminating all except those made between June 12 and 30, 1909, and within a distance of approximately half a mile of each other. But, in every case without exception solitary forms were most abundant and least frequent in the warmer water, while zooids were most abundant and most frequent in the colder water. Furthermore, the frequency of solitary forms was found to be usually identical with that of zooids. Finally, the probability of such an identical series of relations being due to the effects of random sampling was computed, and the odds against it were found to be more than 1,600 to 1. Obviously, then, the apparent paradox has a significance. What is it? DayLicHt HAULS (6 A.M.—6 P.M.) | Solitary Forms Zooids Hauls | j : | Abund. | Freq. Abund. Freq. Cold water E e T 710 96 Warm water $cc). Bb E 158 | 86 NigHtT HAULS (6 P.M.—6 A.M.) . Cold water bee BS ikio Bits bead) 3.81 Warm water | 33 im. AB Lop Let a solitary form be symbolized by a cork, a zooid by an iron weight, warm water by the surface of the ocean, and cold water by the bottom of the ocean. Flotation is then analogous to ac- cumulation in warm water, and sinking to accumulation in cold water. The problem may then be restated as follows: Since corks float and iron weights sink, what is the relation between them that necessitates taking some corks from the bottom of the ocean whenever a number of iron weights are taken there- from, and that necessitates taking some iron weights from the surface whenever a number of corks are taken therefrom? Stated in this symbolical language, it is obvious that the only feasible answer is that at the time the corks and iron weights -were removed from the ocean, they were tied together. Now, if, 576 THE AMERICAN NATURALIST [ Von. LI by experiment, we find that one cork will barely float say six iron weights, corks with more than this number attached would sink, while those with less attached would float. Moreover, if those corks with more than six weights attached usually outnum- bered those with less, while occasionally those with less far out- numbered those with more, then both corks and iron weights would be so distributed with respect to surface and bottom that, while corks in the long run would be obtained in greater num- bers from the surface than from the bottom, at least one would be obtained in a greater percentage of bottom than of surface hauls. In other words corks would be most abundant and least frequent on the surface, while iron weights would be most abun- dant and most frequent on the bottom—the exact parallel of the distribution of the two generations of S. democratica. he conclusion, therefore, seems unescapable that the zooids, after being pushed to the exterior on the proliferating stolon of the solitary salpa, remain attached to each other and to the stolon in the form of a protruding chain. More positive proof being mandatory, each of the seventy-six hauls were carefully examined but, although certain statistical facts pointed toward ithe existence of protruding chains, no two zooids were found attached together. But, being convinced against my own prejudice, that such chains were encountered and broken up during the processes of towing and washing the net, hauls of brief duration were made, and in some of them several fragments of chains were discovered. Comparison of the size of these zooids with that of the largest ones found within the test-cavity of the solitary salpa proves them to have remained attached to their progenitor for a considerable period, probably until at least one entire block of zooids had reached the exterior. The distributional data carry many other significant implica- tions which can not be discussed in this limited time. Suffice it to say that a complete report is now nearly ready for the press and will be published in the Zoological series of the University of California under the title ‘‘Differentials in Behavior of the Two Generations of Salpa democratica semeei to the Tem- perature of the Sea.’’ Ertis L. MICHAEL Scripps INSTITUTION FOR BIOLOGICAL RESEARCH, La JOLLA, CALIF, THE AMERICAN NATURALIST Vou. LI. October, 1917 No. 610 THE MUTATION THEORY AND THE SPECIES- CONCEPT! R. R. GATES Iw the early days of natural history, when the concep- tions of special creation held sway, it was supposed that any one could determine species who was capable of ob- serving the differences between existing forms. Linnæus crystallized this sentiment into the dictum that there are as Many species as were created in the beginning, imply- ing that any one with sufficient powers of discrimination could determine exactly how many species there were in each group. But with the introduction of the theory of evolution, species came to be viewed more and more as dynamic entities, and questions of origin have entered progressively into the species-concept. The latter has grown continually more complex, and yet Darwin’s an- ticipation that systematists would cease to discuss how many Rubi there were in Britain or how many Crataegi in North America, has not been realized. On the contrary, with this increase in the complexity of the conception of species, the extreme views as to what constitutes a species have become more and more di- vergent, until the ‘‘lumpers’’ and ‘‘splitters’’ among sys- tematists usually differ radically in their interpretation of the species in a given genus. This diversity of opinion among systematists has been partly a direct result of our increasing knowledge of the complexity of species, de- 1 Presented at the Pacific Coast meeting of the American Association for the Advancement of Science at San Diego, August, 1916. 577 578 THE AMERICAN NATURALIST [ Vou. LI rived from studies of variation and geographical dis- tribution and from the experimental study of evolution. FH aranana aaa aBa na TAAL 2 a TE HHH aunt s X R kë CECH s s ATRA T EATE aan poo For dam uae io Geography, History. Civica, Kecnomics, tc. Prepared by J. Peat Gonde. Published by The Calverstty of Chicago Prom, Clone Mitata Copyright 1910 by The University of Chicago Fic. 1 Key to Map 1. Otus asio asio (Linn.) Ridgway (= floridanus). 2. Otus asio naevius (Gmel. ay. 3. mecallit (Cassin) Ridgway. 4 hasbroucki Ridgway. 5 aikeni (Brewster) Ridgway. 6 manwelliae (Rid 7 macfarlanet (Brewster) Ridgway. 8 kennicottii (Elliot) Ridgway. 9, brewsteri Ridgway. 10. bendirei (Brewster) Ridgway. 10a quercinus Grinnell. 12 13. (Brewster) Ridg. 14. Otus trichopsis (Wagler) Ridg. 15. Otus vinaceus (Brewster) Ridg. If we look for a moment first at the complexities which have been added to the original simple concept of species, we find all grades and kinds of difference within the species itself, such as subspecies, varieties, forms and © races, ending finally in the differences between single in- ividuals. Some of these conceptions as ordinarily used are also related to geographical distribution. To. them No. 610] MUTATION THEORY AND SPECIES-CONCEPT 579 must be added from experimental work the conceptions of mutants, Mendelian units, phenotypes and genotypes, and pure lines differing only in the position of their modal condition and requiring statistical analysis for their demonstration. With such an array of apparently (though not really) conflicting concepts before him, it is small wonder that the systematist is inclined to cast them all aside and de- scribe his species according to his own ideas of what they are and how they have originated. Nevertheless, for him as for the experimentalist the question What is a species? is more or less colored, if not determined by the question What is its origin? How didit appear? It might be said that the systematist should pay no attention to origins at all, merely describing what he sees: Some systematists doubtless adopt this plan. But obviously, for an under- standing of the characters and relationships of species all possible facts and conceptions bearing on their origin should be considered, and in this way systematics may ultimately hope to become something more than a purely descriptive science. If we examine the ideas which form the background of thought of the systematist in his work of constructing species, we find almost invariably that they are based en- tirely upon the Darwinian conception of natural selection by the gradual accumulation of:slight individual differ- ences. We are then concerned to ask, Is the systematist justified in assuming that all specific and varietal differ- ences have originated in one way? I can find no reason in logic or philosophy why this should be the case; and for ourselves, we believe there is no single method of species formation, but we think the conclusion will ulti- mately be reached that the methods of species formation are multiform, though certain of them are doubtless more widespread and important than others. Among the more important factors in speciation which we wish to consider here should therefore be mentioned (1) local adaptation of races through natural selection or by direct response 580 THE AMERICAN NATURALIST [ Vou. LI to environment, (2) mutations occurring more or less in- dependently of the environment and not necessarily of adaptive value, (3) orthogenesis, whatever that may im- ply. These are by no means mutually exclusive, and we can see no reason why all, and others as well which we have not time to consider, should not have been at work producing the result we call organic evolution. It is not difficult to find, particularly among birds and mammals, instances of specifie variation giving rise to local geographic races which are apparently the result of response on the part of the organism to local conditions. Again, it is easy, particularly among plants, to find vari- eties, species and even genera which have arisen appar- ently through sudden mutations and without anything in the nature of adaptational response. Finally, paleon- tology teems with apparent cases of orthogenetic phylogeny which are not at present clearly explainable in -terms of natural selection or mutations. Before ex- amining concrete cases which come under each of these categories let us diverge for a moment to consider the effect which natural selection as a theory has had upon biological conceptions. We see at once that the philosophical conception of con- tinuity took an extraordinary hold on the minds of biol- ogists. Largely as a result of the great influence of Dar- win, towards the end of the nineteenth century continuity in the origin of species became almost a fetish, and all efforts were directed to showing how every character whatsoever might have originated through the selection of a series of gradually intergrading infinitesimal steps. Yet it is more than doubtful if Darwin himself would ever have been led into such an extreme position. Biolog- ical philosophy has thus been ridden with the conception that if a character could be shown to have arisen in a gradual, piecemeal fashion its origin was thereby ex- plained or accounted for, even though natural selection could not be shown to have operated in its development. On the other hand, the appearance of a character sud- No. 610] MUTATION THEORY AND SPECIES-CONCEPT 581 denly, at one step, was considered practically equivalent to its creation by a miracle, and the type of argument in- volving this view is still not infrequently leveled against the mutation theory. But where lies the necessity for assuming that either continuity or discontinuity is universal? Surely the matter is one to be determined by direct observation, and not by a priori argument. The continuity concept of origins appears not to have influenced other sciences to the extent it has biology. True, Lyell first introduced it by developing the doctrine of uniformitarianism in geol- ogy. Nevertheless the geologist has continued to deal with large and relatively catastrophic effects occurring at irregular intervals, such as landslides, floods, earth- quakes and volcanic eruptions. The phenomena of ge- ological history are then continuous only in a limited sense. Similarly, no chemist supposes it necessary to think that, for example, carbon and silicon were gradually dif- ferentiated from some previous substance which pos- sessed certain qualities of both. On the contrary he sees his atoms built of definite units, the electrons, combined in various ways and numbers to give a variety of prod- ucts, the elements, which are for the most part stable from the first. Hence, while perhaps little can be gained for the biologist by reasoning from analogy with other sciences, yet we at least realize that concepts of discon- tinuity are quite as widespread in science at large as are those of continuity, and that the origin of a character is not explained merely by breaking it up into infinitesimal steps through which it may not have passed at all. Let us consider now some concrete instances. And here we shall select chiefly cases of discontinuity, since we are considering especially the bearing of the mutation theory on the conception of species. In examining species and genera of plants and animals, we find very often, par- ticularly in plants, characters which almost certainly had a discontinuous origin. Perhaps the majority of generic 582 THE AMERICAN NATURALIST (Vou. LI characters in higher plants have originated in this way. Such morphological generic characters are found in num- bers wherever one turns. They indicate a great variety of marked changes, in addition to those involving altera- tion in number of parts; they most often concern the flower structure, on which generic differences usually de- pend; and in many cases at least they can not reasonably be supposed to be of any special value to the plant. If we turn to the lily family and compare the well- known North American genera, Smilacina and Maianthe- mum, we find the following differences: Smilacina Maianthemum Perianth segments 6 Perianth segments 4 Stamens 6 Stamens 4 Ovary 3-celled Ovary 2-celled Style short and thick Style about as long as the ovary Stigma 3-grooved or 3-lobed Stigma 2-lobed or 2-cleft Leaves oblong or lanceolate Leaves usually cordate at base These generic differences are almost entirely in the number of parts in the flower. Otherwise, in foliage and habit Maianthemum might be considered a reduced boreal subgenus of Smilacina. Can it be doubted that Maian- themum originated from the ancestors of certain species of Smilacina through a mutation, in which the flowers changed suddenly from the hexamerous to the tetramerous condition? All these changes in flower-parts would then have occurred at one stroke. It can not well be imagined that they passed through a series of gradual transition stages which have since been lost. When one remembers the almost universal occurrence of 3-parted flowers in Monocotyledons, this change becomes all the more strik- ing. The whole order Crucifere, among Dicotyledons, must have originated in the same way, through a sudden change from pentamery to tetramery. If we examine the species of Maianthemum and their varieties we find evidence that similar processes of dis- continuous variation are going on at the present time. The genus contains three species, M. canadense in Amer- No. 610] MUTATION THEORY AND SPECIES-CONCEPT 583 ica, M. bifolium in Europe and M. dilatatum in western America and northwestern Asia. M. canadense differs from M. bifoliwm chiefly in leaf-shape and in being typi- cally glabrous. The pubescence probably was lost at one stroke, just as numerous glabrous varieties arise. An- other step is sufficient to account for the alteration in leaf-form, so that two steps are ample for the transition from one species to the other. M. dilatatum resembles M. bifolium except for its relatively gigantic size and the fact that it is glabrous like M. canadense. Again two definite steps are sufficient to account for its origin. Turning now to the geographic variations of M. cana- dense particularly as regards pubescence, a detailed study shows that over the greater part of its extensive range it is absolutely glabrous, but that pubescence has appeared especially in two localized parts of its range. A heavily hirsute variety interius Fernald, occurs in the Black Hills of South Dakota, an exceedingly arid region on or near the western extremity of the range of the species. This- variety is apparently restricted in its distribution to the arid portion of western South Dakota, and the most rea- sonable interpretation appears to be that it has originated here through a marked variation, and has thus enabled the species to extend its range into this arid region. Further east, chiefly in Minnesota and Wisconsin, a semi- pubescent form occurs, and this may be the form from: which the much more marked hirsute variety arose. The evidence, when closely examined, favors then a discon- tinuous rather than a continuous manner of origin of this heavily pubescent condition. ‘The condition itself is nevertheless an adaptation, enabling the plant to survive in extreme conditions of aridi The monotypic genus K nikada is related to Streptopus in much the same way that Maianthemum is related to Smilacina. It was originally described in the ‘‘Flora ‘Rossica’’ by Ledebour as Smilacina streptopoides from eastern Siberia. His name indicates his idea of its rela- tionships. Baker afterwards, from plants without flowers 584 _ THE AMERICAN NATURALIST [Vou. LI collected in Oregon, described what has proved to be the same form under the provisional name Streptopus brevi- pes. Kruhsea streptopoides then, although it agrees © with Streptopus in foliage and in fruit and seed char- acters, differs remarkably in its flowers. They are very small, the perianth nearly rotate, dark purple; the sta- mens altered; and in the absence of a style the discoid stigma rests directly on the ovary. It is possible that connecting forms between Kruhsea and Streptopus may yet be found in Siberia, but at any rate the differences between these two genera can not be reasonably sup- posed to have arisen through natural selection. Kruhsea appears to have originated through a few definite ger- minal changes and to have since been momen by heredity. ` Another pair of genera which is of inwth interest in this connection is Platystemon and Platystigma, two Cali- fornian genera of the Papaveracee. Both occur abun- dantly as spring flowers, occupying similar habitats. Their main differences are as follows: ome: Platystigma Stamens numero Stamens 6-12 Filaments eg ji flattened ilaments narrower, flattened or Carpels 6-20, forming a compound filiform ovary, which in fruit breaks up by Carpels 3, forming a 1-celled, 3- constrictions into 1-seeded joints valved or terete ovary which in fruit forms a 3-valved dehiscent capsule These genera are almost exactly alike in habit, foliage, pubescence, color of flowers and general form of the sta- mens. They differ chiefly in the pistils, and these differ- ences only become conspicuous as the seed capsules mature. Platystemon has acquired numerous carpels which are connivent or coherent in a circle. In develop- ing, the carpels separate and their free margins cohere with each other. Each carpel then becomes torulose by constrictions between the seeds. How shall we account for the origin of such a condition except through a marked No. 610] MUTATION THEORY AND SPECIES-CONCEPT 585 variation, which is perpetuated by heredity and not be- cause the plant has any advantage or disadvantage in life compared with Platystigma. Only one species of Plat- ystemon (P. californicus Benth.) and two of Platystigma have been generally recognized, although Greene? has described some 50 species based on minor differences. Another significant difference between Platystemon and Platystigma is in the variations of the petals. In Platystemon the number varies from 6 to 10 or more, and all the petals of a flower or a plant vary together in color from dark yellow through light yellow to white. In Platystigma, on the other hand, the number of petals ap- pears to be uniformly six, and the outer three vary in color independently of the (alternate) inner three. Thus in Platystigma lineare Benth. (which Greene calls Hes- peromecon pulchellum) the outer petals may be dark yel- low, or with a more or less extensive wedge-shaped dark yellow mark at the tip, while the inner petals are light - yellow or white.® The peculiarity in the carpels of Platystemon acquires added interest from the fact that, as Lindley pointed out,* it is by no means unique, but contraction of the sides of the carpels, forming a torulose structure, has occurred equally and must have originated independently in Hype- coum of the Papaveracex, in such genera as the radishes among Crucifere, in Ornithopus among the Leguminose, and in other families. We may look upon this condition as apparently the result of parallel mutation in different families, independent of utility; and countless other cases of a similar kind occur among higher plants. From the few instances I have cited, which could be added to indefinitely, and from the abundant evidence of marked variations which we have from experiment, the 2 Greene, E. L., 1903, ‘‘ Platystemon and its Allies,’’ Pittonia, 5: 139-194, 3 A figure of P. lineare in Bot. Reg., T. 1954 (1837), from the Russian River, Cal., shows the petals alternately yellow and white. Another interenting point, to which Mrs, K. Brandegee has directed my at- tention, is the abundant occurrence of tiny plants bearing a single minute flower, intermingled with the larger plants. 4 Bot. Reg., T. 1679 (1834). 586 THE AMERICAN NATURALIST (Vou. LI conclusion seems clear that many marked morphological characters in plants have arisen independently of func- tion and without the aid of natural selection. This con- clusion is all the more probable because form is so much more loosely tied to function in plants than in animals. In many plants it makes little or no difference what is the shape of the leaf so far as its chlorophyllian function is concerned, nor what is the shape of the anthers so long as they produce pollen. Another matter, which I have touched upon elsewhere,® is the geographic relationships of the most closely related species of plants. It appears that Jordan’s well-known law that the most nearly related species occupy adjacent areas, although widely applicable especially to the sub- species of mammals and birds, is by no means so generally true in regard to plants. But we shall come to this point again. Referring now to animals, the North American screech owls afford an interesting case in which two kinds of variability can be clearly contrasted as regards their geo- graphic relationships. These two types of variations are (1) those in which apparently continuous or nearly continuous variations occur progressively over certain geographical areas, with no two forms occupying the same area, and (2) those in which two or more sharply marked forms occupy the same area. The accompanying map, compiled largely from Ridg- way’s data, shows the distribution of the various sub- 5 Gates, R. R., 1916, ‘‘On Pairs of Species,’’ Bot. Gazette, 61: 177-212. . 1. 6 Ridgway, Robert, 1914, ‘‘The Birds of North and Middle America,’’ Bull. U. S. Nat. Mus., No. 50, Part VI, pp. 882, pls. 36. Ridgway says (p. 683): ‘‘In the main, geographic variations [in Otus] are more or less marked and constant; but occasionally specimens oceur in a given area which are with difficulty, if at all, distinguishable from the form inhabiting another—sometimes distant—geographic area.’’ He further com- ments on the fact that, while 0. choliba in South America is remarkably uniform over a vast area, O. asio shows great change of coloration within relatively short distances, indicating an organization sensitive to slight changes in the physical environment. No. 610] MUTATION THEORY AND SPECIES-CONCEPT 587 species of Otus Asio Stephens (formerly known as Meg- ascops asio Kaup) over the North American continent. While such a map is only approximately accurate, it shows that in general only one subspecies occupies a given geographic area.’ There are, however, certain excep- tions. Thus in Central Colorado Otus asio aikeni and Otus asio maxwellie both occur, the former finding here its northern limit from Texas and Mexico, the latter the southern limit of its range from Montana. It is stated, however,’ that they occur in Colorado chiefly at different altitudes, maxwellie up to 6,000 ft. and aikent from 5,000 to 9,000 ft. This is the reverse of what might be expected, since aikeni is the more southern form. But Mr. Aiken states? that at Colorado Springs maazwellie occurs only in winter and aikeni only in summer, indicating a slight migration. Again, gilmani and cineraceus—the latter somewhat darker with coarser pencilings and averaging slightly larger in size—both occur in southwestern Ari- zona, but, according to Swarth,’® although both birds may occasionally be taken in the same locality, this is only in winter when cineraceus comes down from the higher alti- tudes to the different life zone of the hot Lower Sonoran valleys occupied by gilmani. The differences between these various subspecies are chiefly in density of coloration and in size! Thus Otus asio nevius is larger than Otus asio asio and is also lighter in coloration, with more white on the under parts. The subspecies mccallai in Texas and northern Mexico is in- termediate between these in size, but is paler than either 7It may be pointed out that there is sometimes discernible a tendency for systematists to call a form a subspecies or a species according to whether or not it is the only form in a given area, thus making the geographical rela- tions of the form their criterion, rather than the degree of its distinctness. 8 Cooke, W. W., 1897, ‘‘The Birds of Colorado,’’ Bull. No. 37, Agric. Expt. Sta., Fort Collins, Colo., p. 78. 9 Cooke, W. W., 1898, ‘“‘Further Notes on the Birds of Colorado,” Bull. No. 44, Agric. Expt. Station, Fort Collins, Colo. 10 evar, H. S., 1916, ‘‘The Sahuaro Aak Owl as a Recognizable Race,’’ Condor, 18: 163-165. 11 I am indebted to Dr. Grinnell for permission to examine series of speci- mens in the Museum of Vertebrate Zoology of the University of California. 588 THE AMERICAN NATURALIST [ Von. LI and more coarsely mottled. Hasbroucki, very limited in known range (see map), is decidedly larger and darker than mecallii, with much less buffy gray above and broader transverse bars. Maawellie, another northern form, is decidedly larger but paler than aikeni. It is the palest of all in color, with more extensive pure white than even nevius. West of maawellie, in Washington and Oregon, is macfarlanei, which is larger and very much darker, almost agreeing in coloration with bendirei of California. The Pacific coast forms comprise an interesting series running down the coast, beginning with kennicotti, which occurs from Sitka through British Columbia to the south- ern border of Washington State. It is very large like macfarlanei, but much darker, and browner rather than gray. The remaining subspecies extending down the coast region and into the desert become progressively paler and smaller. Thus brewsteri in Oregon is smaller and less brownish than kennicottii. In California occurs bendirei which is lighter again and smaller. Grinnell*? has segregated from bendirei in the more arid region of southern California another form under the subspecific name quercinus, considered to be paler dorsally and with less or no ferruginous markings around the head. But I. confess that this difference, if it exists as a constant dis- tinction, is too fine for me to appreciate. On the contrary, specimens of bendirei from Palo Alto appeared to me somewhat lighter on the breast than a series from Pasa- dena. Whether or not this very close form is distinguish- able from bendirei, the next in the series are cineraceus, gilmani and xantusi, becoming progressively lighter with finer vermiculations, the two former in southern Arizona and xantusi confined to the tip of the peninsula of Lower California, smaller and with the toes less feathered. - Thus the subspecies appear to be arranged progres- sively in passing from one geographic area to another, and there is little overlapping. But this conception of 12 Grinnell, 8 1915, ‘tA New Subspecies of Screech Owl from Califor- nia,’’ Auk, 32: No.610] MUTATION THEORY AND SPECIES-CONCEPT 589 gradual and progressive change can be overdone when it is attempted to correlate the alterations observed with climatic or other environmental features. Thus the pro- gressive lightening in color from kennicottiu to xantusi, first by lightening and restriction of the brown until it practically all disappears and then by paling and diminu- tion of the gray, is believed to be associated with the de- creasing moisture in the northern part of the range and the increasing aridity in the south. There are of course many well-known cases of paler races of birds and mam- mals occupying desert areas. Yet it is not clear that the coastal region of Oregon, where the less dense brown brewsteri occurs, is any less humid than the correspond- ing part of Washington where kennicotti is found. Simi- larly wantusi on the peninsula of Lower California can not be supposed to exist in a drier habitat than gilmani or cineraceus. Of course in none of these cases is it known just what feature in the environment acts as the critical factor nor how the race responds to it. The experimental studies of Tower!® and others show that a race may re- spond in the same way (i. e., by showing the same vari- ations) to different environmental stimuli or in different ways to the same stimulus. But studies of this character are still too few to furnish a basis for interpreting these reactions on the part of species of the higher animals. The experiments being carried on by Sumner'+ with the white-footed mouse, Peromyscus maniculatus may be ex- pected to throw further light on this important question of the origin of local subspecies. Again, it is not certain that such races as kennicottiu, brewsteri and bendirei form an absolutely graded series with all intermediates. On the contrary there appears to . be some evidence that although their boundaries are con- tiguous there are definite though small steps from one to 13 Tower, W. L., 1906, ‘An Investigation of Evolution in Chrysomelid Beetles of the Genes Leptinotarsa,’’ Carnegie Inst. Publ. 48, pp. 320, figs. 31, pls. 30. 14 Sumner, F. B., 1915, ‘‘Genetie Studies of Several Geographic Races of California Deer-mice,’’ Am. NAT., 49: 688-701, with map. 590 THE AMERICAN NATURALIST (Vou. LI the other. This may conceivably be explained through the principle of invasion and reinvasion. Grinnell, * who is doing so much towards a detailed knowledge of the Pacific coast fauna, has considered this principle and also the part played by barriers in the development of geo- graphic subspecies or races, in connection with the discus- sion of many specific cases of distribution in birds and mammals. Walter P. Tayler,!® in a recent study of the western beavers, concludes in agreement with others, that migration, geographic isolation with adaptation to local ecological niches, and final reinvasion of earlier-ocecupied localities, will account for the origin and present distribu- tion of geographic subspecies such as we have been con- sidering. This explanation seems as likely as any other at the present time, but it is beyond the purpose of the present paper to discuss these aspects of speciation in birds and mammals. The intention is rather to show that the problems involved are entirely different from those concerned with another type of variability to be men- tioned ina moment. It may be pointed out, however, that although the theory of reinvasion as developed involves the conception of races isolated in certain geographic areas becoming gradually modified through environ- mental stress and fixed before the reinvasion takes place; that there is at the present time no definite evidence that fixation actually takes place gradually, in this way or in any other way. If intermediates between the various geographic subspecies do not occur, this may be because definite though small steps in variation are taken from one race to the other, which would do away with the necessity for assuming a long period of isolation during which the gradual development and fixation of the race 15 Grinnell, Joseph, 1914, ‘‘An Account of the Mammals and Birds of the Lower Colorado Valley, with Especial Reference to the Distributional Prob- lems Presented.’? Univ. Calif. Publ. Zool., 12: 51-294, 9 figs., pls, 3-13, and other papers. 16 hide Walter P., 1916, ‘‘The Status of the Beavers of Western North ca, with a Chusldpeaition of the Pasioni in Their Speciation,’’ Ma. Calif. Publ. Zool., 12: 413-495, No. 610] MUTATION THEORY AND SPECIES-CONCEPT 591 occurred. However, the process does appear to be gradual at least in comparison with the other type of variability, which is fundamentally different in its geo- graphic relations. The second type of variability in Otus asio to which I have reference, consists in the occurrence of gray and reddish or rufous phases of coloration in the same area of distribution. Thus all the eastern subspecies, asio, nevius, mecallii and hasbroucki, produce both gray and red birds. These phases are sharply marked, and inter- mediates rarely occur. Hasbrouck!" attempted an ex- planation of this dichromatic condition, but some of his conclusions were justly criticized by Allen.1s The gray phase occurs more commonly in Florida and in the north- ern part of the range of nevius,!® while the red phase oc- curs commonly in the Central Atlantic states, perhaps to the exclusion of the gray in some localities. The red phase is unknown in the western forms of Otus asio. Nevertheless grayish and rufescent phases of the small O. flammeolus, which is found in the mountains of western America from British Columbia to Mexico, occur in this region. The red phase is found also in O. trichopsis (see map). Similarly, brown and rufous phases are found in O. choliba which extends over a large part of South America, and also in the Central and South Amer- ican species O. cassini, O. guatamale, O. barbarus and O. vermiculatus. Owls belonging to other genera also exhibit two phases. For example Bubo virginianus (Asio magellanicus),” the single species of Bubo occurring in all North and South America, with many geographic varieties, shows dichro- matism in various parts of its range. The same is ap- parently true of various Old World owls. 17 Hasbrouck, E. M., 1893, ‘‘ Evolution and Dichromatism in the Genus Megascops,’’ AMER. NAT., 27: 521-533, 638-649, 4 maps. 18 A (llen), J. A., 1893, Auk, 10: 347-351. 19 Oberholser, H. C., 1904, ‘‘A Revision of the American Great Horned Owls,’? Proc. U. S. Nat. Mus., 27: 177-192. 20 The red phase is stated by Allen to be rare in Maine. 592 THE AMERICAN NATURALIST [Vou. LI Dichromatism is then, both geographically and sys- tematically, a widespread phenomenon in owls. The red phase appears to be quite independent of geographic locality in its origin. Hasbrouck attempted to show with regard to Otus asio that the red phase had arisen gradu- ally from the gray, which it was slowly supplanting in certain areas. He believed that the grays inhabited regions of greater humidity (Florida, northern range of nevius) and the reds the drier interior, yet grays occur in Florida and reds are found, though uncommonly, in Maine. He also reported reds as occurring exclusively in the relatively humid Mississippi valley. But, as Allen pointed out, any such correlation with climatic or environ- mental factors hopelessly breaks down because both types are found indiscriminately over at least the greater part of the eastern range. All writers agree that the two types of plumage are independent of age, sex or season, and that in many localities at least both occur together and freely interbreed. Hasbrouck states, however, that gray males far outnumber red males, while red females out- number gray females 4:1. Confirmation of this point is to be desired, as it suggests sex-linkage of the red con- dition. It is further stated that on the continent of Europe the red owls are said always to be females and the grays males. All young birds of Otus asio are gray in the down, the red first appearing in the feathers. Observations go to show that red birds mated with red may give (1) all red offspring, (2) all gray, or (3) both red and gray. When one parent is red and the other gray, the same three re- sults may follow. Further, Hasbrouck claims that gray gray gives always only gray young. This is probably true, but since the result is based on observation of only six matings of this kind in regions where reds occur, it is much to be desired that further observations on this point should be recorded. The obvious hypothesis to explain ee facts is that the red phase appeared as a mutation from the gray, and No. 610] MUTATION THEORY AND SPECIES-CONCEPT 593 that it is inherited as a simple Mendelian dominant char- acter. The results of the various matings between red and gray would then be as stated above, according to whether the red parent were homozygous or heterozygous, but the offspring from red X red should seldom be all gray, since this would be only a chance result when both parents were heterozygous. It is not impossible, how- ever, since the screech owls usually have only three or four young in a nest, or sometimes only two. Since the red phase occurs in various species as well as subspecies it is not improbable that it has originated through independent variations in different species. In any case the geographic ranges of the red phases show that, having appeared as variations, they are inherited without any conspicuous advantage or disadvantage in competition with the gray. The present frequency of the reds in certain areas and their infrequency in others may be merely an indication of the localities where the original mutations took place, and from which as centers they have gradually spread. Although the western forms have no red phase, yet Otus asio kennicottii exhibits in addition to its usual tawny-brown phase a relatively rare gray phase. This fact is indeed an argument favoring the assumption that the brown phase of Kennicottii also arose at one step and has since nearly supplanted the original gray form. If now we compare the two types of variability that I have described in Otus asio, we find them sharply con- trasted in several respects: (1) the former is clearly re- lated to geographical distribution, a single race occurring in each locality: the latter has no such relation, but two forms may occur interchangeably in the same place; (2) the former is essentially continuous as a form of vari- ation, the latter markedly discontinuous; (3) the former appears to be related to environmental (climatic) con- ditions, the latter apparently bears no such relation. As regards their evolutionary significance, there can be little doubt that the former or apparently continuous type of 594 THE AMERICAN NATURALIST [ Vou. LI variations is more important in this case, for they appear to have given rise to the geographic subspecies now recog- nized, and, moreover, the specific differences in the genus are merely an exaggeration or intensification of the kinds of difference shown by these subspecies. It seems evi- dent, then, that the differentiation which has gone on in the evolution of the genus Otus is for the most part of the kind exemplified by the small differences now existing be- tween geographic races or subspecies occupying different areas. The same thing is true of many other birds and mam- mals, but this condition is by no means universal even in these groups of animals. On the contrary, it is not diffi- cult to find instances in which the discontinuous type of variation, independent of environment or function, has been the main factor in speciation. I will merely mention the case of the North American flickers, Colaptes auratus and C. cafer, set forth by Bateson,” since the latter is a Californian bird. These species differ remarkably in their color markings, the most conspicuous differences being (1) yellow or red quills, (2) a black or a red malar strip in the males, (3) the presence or absence of a scarlet nuchal crescent in males and females. C. auratus pos- sesses the first of each pair of characters and C. cafer the second. C. auratus extends from Alaska diagonally across Canada and the United States to Texas and eastward to the Atlantic, while C. cafer occurs in its pure form from Oregon through Utah, California and Arizona into Mexico. Each possesses 3 or 4 geographical subspecies. Where the ranges of the species overlap over a large area a mixed population of forms occurs which is usually in- terpreted as a series of complex hybrids, but this will bear further study. It is clear, however, as Bateson points out, that the differences in range of the species can not be associated with any constant environmental difference in the habitats, and that the species can not have differentiated from this mixed population of inter- 21 Bateson, W., 1913, ‘‘ Problems of Genetics,” Yale Univ. Press, p. 146. No. 610] MUTATION THEORY AND SPECIES-CONCEPT 6595 mediate forms. However these species originated, they can not be reasonably supposed to have developed through gradual adaptation, but the color differences probably play no more part in the economy of the species than is the case with the red and gray phases of the screech owls. Something in the germinal organization of Colaptes doubtless determines the definiteness of its color patterns, and it is probable that each element of the pattern was changed by a marked step rather than through a series of gradual stages. This view is strengthened by the fact that a third species, C. chrysoides in Lower California, is essentially a cafer with yellow instead of red quills. Thus even in birds our second type of variation, non- adaptational and not related to local conditions, is appar- ently an important factor in speciation, although in Colaptes too geographical races occur as well. In dis- tribution also these species do not follow the rule for geographic subspecies, for they overlap over large areas. The fact that each species has its own geographic sub- species shows that the origin of these species antedates the development of their geographic varieties. I have endeavored to show that in plant and animal species there are two distinct types of variability, having different geographical relations. The one is discon- tinuous, independent of environmental or functional in- fluence, and has given rise to many specific and generic characters, notably in plants but also in higher animals. The other is continuous and apparently represents the results of the stress of the environment on the species in its dispersal, leading to the gradual differentiation of local races or subspecies whose peculiarities are ulti- mately intensified and fixed. The latter type of specia- tion is notably exemplified in birds and mammals, organ- isms in which, unlike plants, the individuals can migrate from place to place and so substitute for a stress result- ing from overpopulation an environmental stress caused by a new set of climatic or physiographic conditions. FURTHER OBSERVATIONS ON THE EFFECTS OF ALCOHOL ON WHITE MICE? L. B. NICE In a former paper (711) it was found that white mice were not markedly affected when given alcohol in their food. Since this paper appeared Stockard (’12, 713, 716) has brought forth some striking and conclusive results demonstrating that guinea pigs are very sensitive to alcohol and decidedly injured by it. He administered the alcohol to his animals through the lungs by placing them in a tank containing alcohol at the bottom so that they had to inhale the fumes. His work raised the question as to whether similar results might be obtained with white mice by using the same method. Therefore it was decided to repeat my experiments, using the inhalation method. For these experiments white mice eleven weeks old were obtained. They were all from one strain inbred to the fourth generation, two entirely distinct strains having been united to form this strain.. They were divided into four lines, viz., a control line, a double alcohol line, that is, both males and females were subjected to alcohol, a female alcohol line and a male alcohol line. There were three cages in each line, each cage contained two females and one male, thus making six females and three males in each line. Two cages were made up of second generation al- coholized mice; both males were from the male alcohol line, two females from the same line and one female from the female alcohol line. The same cages were used as in my former experiments. They were made of 8-mesh wire and were 6 inches wide, 6 inches deep and 12 inches long. The mice were kept in a laboratory room heated by steam. It was attempted to keep the room at a uniform temperature, but fluctuations occurred. 1 From the Laboratory of Physiology in the University of Oklahoma. or a review of the literature see Nice (’11 and 12). Also Stockard (712, °13 and ’16), 596 No. 610] ALCOHOL AND WHITE MICE 597 All the animals were fed the same food, consisting of wheat and kaffir corn with bread and milk once a day. Every day except Sunday the double alcohol line, par- ents and young, the males of the male alcohol line, the females and young of the female alcohol line and the sec- ond generation of alcohol mice with their young were placed in a galvanized tank 26 inches long, 20 inches wide and 14 inches deep. Alcohol had been poured on to cotton which was placed under a wire mesh situated about 2 inches from the bottom of the tank, so the mice had to breathe the fumes. The mice were kept in the tank each day until they became intoxicated, as shown by their stag- gering gait or inability to stand up. At the beginning of the experiment the time necessary to intoxicate them was about one hour. Later they would often be kept in the tank for two hours. This shows that the mice acquired a tolerance for alcohol. This tolerance was shown after they had been treated about a month. To make sure that the mice were being heavily alecoholized, a few times they were left in the tank so long that they would not recover from the effects for three or more hours, and in some cases they did not recover from the intoxication, but died. At the beginning of the experiment when the mice were placed in the aleohol tank they would sneeze, their eyes water and they would rush about in their cages, showing great uneasiness. Later they ceased to be so much dis- turbed, yet during the course of the experiment there was no indication that the mice liked the aleohol fumes. THE WEIGHT oF THE ADULT MICE The adult mice were weighed at the beginning of the experiment and once each month thereafter, to get an in- dication as to their health. Since they were nearly the same age and closely related, their average weights would be expected to be about the same unless the alcohol treat- ment had an injurious effect on them. By referring to Table X, it will be seen that there is only a slight dif- ference between the various lines. The average weight 598 THE AMERICAN NATURALIST [ Vou. LI of the mice not treated with alcohol was 22.4 grams and of those treated with alcohol 21.8 grams. TABLE I AVERAGE GAINS oF ADULT MICE Those Not Given Alcohol Average g Average Line Sex No.of Mice| Gain in |No.ofMice| Gain in Weighed | Gramsat4| Weighed | Grams at 7 Months Months Control Male 3 4.6 3 4.9 Co ki. Female 6 4.3 6 9.3 Male alcohol Female 6 | 4.3 6 6.4 F le al Male 3 5.8 Average gain of all mice not given alcoho! poas piue 15 7.2 TABLE II AVERAGE GAINS OF ADULT MICE Those Treated with Alcohol Average Average Line No. of Mice Gain in No. of Mice} Gain in Sex Weighed Forks paa at Weighed Grams at 4 Months 7 Months Double alcohol.................| Male 3 Hed 1 4.0 Double alcohol................. Female 8 4.8 2 4.2 Male alcohol wiiiis. a E Male 3 3.0 3 3.5 Female alcohol................! Female 4 3.5 Average pain of all mice given alcohol) 18 3.9 6 3.8 Tables I and II give the average gain of the different lines for four months and for seven months. The mice that did not receive alcohol gained more than those that were treated; the former gaining 4.6 grams on an average for four months and the latter 3.9 grams; in seven months the untreated mice ening 7.2 grams and the alcoholized 3,8 grams. | -Itis possible that dia Fasihi of the mice and the extra exercise they took in the excitement of being alcoholized might account in part for their growing less than the un- treated mice. In the 1911 experiments the control mice carried 7 months gained only 2 grams on an average while the alcohol mice gained 6 grams; in the second generation carried four months the controls gained 1 gram each and No. 610] ALCOHOL AND WHITE MICE 599 15 ; pdi l4 13 12 ul 19 9 8 7 6 5 4 3 Control line 31 mice 2 -m ao eee è , LSE alechol line 25 mice -e 6 a comes § Femsle alcohol line 21 mice w =n ass aw ane Double alcokol line 13 mice | ; -m = æ eee = æ Seconåd generation alcohol line ll stice o 0 | 2 3 4 5 4 Fic. 1. Curve showing the growth of the mice. The abscissas — the age of the mice in weeks and the ordinates their weights in gram: 600 THE AMERICAN NATURALIST [Vou. LI the alcohol mice 2 grams. In that case the alcohol was given in the food and water and apparently had a fat- tening effect. None of those animals were handled except for weighing them. - Recorps oF THE Young oF HacH FEMALE A record was kept of the young of each female. Many of the young in each of the lines were eaten by their par- ents, Tables III to VIII show the number of months each female was carried; the number of litters each had; the total number of young born, and the number that died ap- parently from lack of vitality. TABLE III RECORD OF THE YOUNG OF EACH FEMALE CoNTROL LINE Female No. of Mo. 3 No, of Young | No. of Young ove ti * Born That Died = sheath ER eO | ped ph poa eee Y jet .. Total F; TRANW AUR OS o;joooooooeo wej ORAS ee oO * Females G and H are second generation controls. TABLE IV MALE ALCOHOL LINE Female ane | O ee | 7e on ate Taai Aiai Canai 7 2 12 1 B 7 1 12 0 Chae 7 2 10 0 D ‘ 7 1 3 3 E 7 1 5 0 Mas | bi 3 24 0 eine | 7 o | a 4 No. 610] ALCOHOL AND WHITE MICE 601 TABLE V FEMALE ALCOHOL LINE No. of Mo. No. of Y No. ot Y sven p Saio | OM Mie T That Dien A 4 1 7 0 B 4 1 7 0 G 4 2 12 3 YS soa 4 2 15 1 E 2 0 0 0 F. 2 0 0 0 aa A A A O R 6 41 | 4 2:2 | Note.—Females A, B, C, and D were killed by being left in the alcohol tank too long. Females E and F were killed by accident. TABLE VI DOUBLE ALCOHOL LINE Female ` pores i No. of Litters oig Haag na oer 4 4 | 1 5 0 B 4 | 1 7 0 a 9 | 0 0 0 D 9 | 3 20 0 p 4 | 2 15 z F PE N 15 0 G 4 | i 5 1 T 4 | 1 6 2 Total ee a i. 73 5 2:9 | j * These mice from A to F were left in the alcohol tank so long one day that only C and D survived the experience. After this accident one more cage, G and H, were made up from the original stock. TABLE VII SECOND GENERATION ALCOHOL LINE Hoot. No. of specs Mo. Observed | NO of Litters i aa Bom |. That pied” A 23 1 6 1 B 24 1 10 0 g. 23 1 9 0 TOMS... SOLE 24 3 25 1 VIABILITY OF THE Youna Table VIII shows the number of litters and the number of young born in each line; also the number that died from lack of vitality. 602 THE AMERICAN NATURALIST (Von. LI TABLE VIII RECORD OF THE YOUNG OF EACH LINE Summary of Tables III to VII Tip No. ot | No. ot | No. ot | No.of |No.That| Per Cent. Mice | Months | Litters | Young Died ‘ThatDied ice cn | Control | | 1 4 9} AT 0 0 aS 3 | Male alcohol E. 7 10 | 66 4 6 Female alcohol.................... | 4 + 6 | 41 4 9.8 Double alcohol....,.............0+. | { 6 > ake! 13 we be Second generation alcohol...... | 3 24 Se ce 1 4 As in my former experiments none of the control young died of lack of vitality. The alcohol lines show a small percentage of deaths—4 mice or six per cent. in the male alcohol line, 5 mice or 6.8 per cent. in the double alcohol line and 4 mice or 9.8 per cent. in the female alcohol line. The second generation alcohol lines had 1 death or 4 per cent. of all of their young. In the former experiments (711) the fatalities were somewhat greater—9 young or 11.1 per cent. in the first generation of aleoholized mice and 7 young or 12.5 per cent. in the second generation. Stockard (716), with his guinea pigs, had a fatality of 43 per cent. in the male alcohol line, 52 per cent. in the female alcohol line, 46 per cent. in the double alcohol line and 16 per cent. in the control line. FECUNDITY Table IX gives the average number of litters, average number of young, and average number in a litter for one female of the control line and one female of the male alcohol lines for seven months; for one female of the female alcohol line and double aleohol line for 4 months; and for the second generation alcohol line for 24 months. On account of the difference in the length of time the different lines were carried, it is impossible to make a direct comparison. However the greater fecundity of all the alcohol lines over the control line is striking. Though No. 610] ALCOHOL AND WHITE MICE 603 TABLE IX FECUNDITY OF THE DIFFERENT LINES Average of One Female of Each Line tne | Months | Average in |Average No.|Average No. Observed a Litter of Litters of Young Control | 7 | 5.1] 1.3 6.5 ale alcoho Ei 6.6 1.66 11.0 Female alcohol pona 6.83 1.5 10.25 Double alcohol | 4 6.5 1.4 9.3 Second generation alcohol | 24 8.3 1.0 8.3 the control mice were carried longer than any line except the male alcohol line, they have next to the lowest num- ber of litters—the lowest being the second generation alcohol line carried only one third as long as they. They have the fewest young of all the lines and the smallest litters. The male alcohol line can be compared directly with the control line since they were both carried seven months. They have somewhat larger litters, somewhat greater average number of litters and nearly twice as many young as the controls. It is not possible to compare them directly with the lines that were only carried four months, but since the averages of these lines are almost as high as those of the male alcohol mice, it follows that the male alcohol mice were not as fecund as the female alcohol and double alcohol lines. The female alcohol mice show the greatest fecundity of all the lines, while the double alcohol and second generation of aleohol mice come next. The three lines in which the females were alcoholized were somewhat more fecund than the line in which the males alone were alcoholized and decidedly more so than that in which neither parent was aleoholized. These re- sults confirm those obtained in my former work (711) where the control mice carried 7 months had 2.2 litters or 13.3 young on an average and the alcohol mice had 2.8 litters and 16.1 young; the second generation of control mice carried 4 months had 1.5 litters and 7.1 young, while the corresponding alcohol line had 1.8 litters or 12.4 young. | Why the mice had fewer young in these experiments 604 THE AMERICAN NATURALIST [ Vou. LI than in the former is not clear. It may have been due in part to the greater fluctuations of temperature in the laboratory building used here. Whatever the reason, the control mice in these experiments after the first few months occupied themselves in growing fat instead of having young. Stockard’s results on guinea pigs are directly contrary to these; his alcoholized animals had decidedly fewer young than the control guinea pigs. CoMPARISON OF THE GROWTH OF THE YOUNG IN THE VARIOUS LINES In comparing the weights in Table X and Curve I it should be remembered that all of these young were al- coholized except the male alcohol line and de course the controls. TABLE X WEEKLY GROWTH OF THE YOUNG 3,/% r legis ie jas Fls 131% Line BR) BA |Sa)og ney arian BE| BES tie o 20/82 Bigsigeigeig 2o % v os P eg (erifs\e6/g6lee RERE fa 88 | En <"|< Ja |*8/< |< |< < |< |< EEE | 'ontrols 22.6| 21.5| 31 | 1.4 2.3 3.8 49 6.6 8.7 10.2) 11.2) 11.6 Male alcohol 23.2) 24.8 25 | 1.2) 2.1 3.7, 4.8 7.4 9.3 11.1 13.1) 13.6 male 20.2) 21.4| 21 | 1.1| 2.5) 3.5| 3.8] 5.5) 7.7) 9.5) 10.3 14.3 Double EMRA EEPE 21.1| 23.7 13. 1.2) 2.6 3.5 5.9) 8.4 9.0) 10.9 11.3 18 Second generation alcohol... 13.5117 | 11/ 1.1) 2.4 3.8! 4.7' 6.4) 8.8' 11.5 14.5: 15.1 The weights of all the lines at birth and for the first two weeks are quite similar. After that variations began. The young of the double alcohol line surpassed all for four and a half weeks, while the young of the female alcohol line fell behind all the others at the beginning of the third week and remained below up to the seventh week and at the eighth week they were next to the highest. The young of the control line, the male alcohol line and the second generation of the alcohol line grew at about the same rate up to the fifth week. At this time the weight of the male alcohol line slightly surpassed all the others; then the No. 610] ALCOHOL AND WHITE MICE 605 & second generation of alcohol mice outgrew all and con- tinued ahead until the end of the experiment. After the sixth week there were rather wide variations and this con- tinued as long as they were weighed. At the eighth week the weights of the different lines stood in the following order: second generation of alcohol mice 15.1 grams; female alcohol mice 14.3 grams; male alcohol mice 13.6 grams; the double alcohol mice 13 grams; and the controls 11.6 grams. In my former experiments (711) the alcohol young surpassed the controls in the rate of growth. COMPARISON OF THE DIFFERENT LINES.. The control line had the fewest young of any of the lines; they had no deaths from lack of vitality ; the growth of their young was slower than that of any of the other lines except the female alcohol line. The male alcohol line was more fecund than the con- trols, but less so than the other alcohol lines; their death rate from lack of vitality was four mice or 6 per cent.; the growth of their young was better than that of the controls and female alcohol lines. The female alcohol mice were the most fecund of all the lines; their death rate was four mice or 9.8 per cent.; their growth was even slower than the control mice until the last week, when they made a large gain and outgrew all but the second generation of alcohol mice. The double alcohol mice were slightly less fecund than the female alcohol line; five mice or 6.8 per cent. of their young died; they grew a little faster than the controls. For the second generation of alcohol mice, two males, offspring of the male alcohol line, were mated with two females, young of the same line and one female from the female alcohol line. Thus one grandmother and all but one of the grandfathers were alcoholized, the second gen- eration were all alcoholized after they became adult and. one from birth and their young also were alcoholized. The fecundity of the second generation of aleohol mice was high; they had one death from lack of vitality, or 4 « 606 THE AMERICAN NATURALIST [Vou. LI per cent. of all their young, and their young grew the fastest of all the lines. It is a matter of regret that owing to an accident— over-aleoholization one day—the second and third gen- erations were not carried farther. However, as far as they went, no injurious effect from alcohol is apparent in fertility, nor vigor of growth, and but a small one in viability. From indications in our results it would seem to be dangerous to draw far-reaching conclusions from data obtained on a single species. Although this work was not carried as long as it was planned, yet as it corrob- orates my former experiments in practically every detail it goes to prove that mice are to a degree resistant to aleohol whether it is fed or inhaled by them. From re- sults obtained in bacteriological laboratories it is well known that mice are very resistant little animals, in com- parison to sensitive animals like guinea pigs. Mice are immune to the toxin of the tetanus bacillus. It seems reasonable to expect that an animal which is immune to such a virulent toxin might have a considerable degree of resistance to the effects of alcohol. SUMMARY 1. The white mice given alcohol by the inhalation method gave much the same results as those that received it in their food in my former experiments. 2. The fecundity of the alcohol mice was greater than that of the control mice, as in my former study. 3. Six per cent. of the young of the male alcohol line, 6.8 per cent. of the double alcohol line, 9.8 per cent. of the female alcohol line and 4 per cent. of the second gen- eration alcohol line died from lowered vitality, while none of the control young died. Similar results were obtained in my former experiments, except that the alcohol line had a higher death rate—11.1 per cent. in the first gen- eration and 12.5 per cent. in the second generation. 4. The growth of the young of all the alcohol lines ex- No. 610] ALCOHOL AND WHITE MICE 607 ceeded that of the controls, as in my former experiments. The young of the second generation alcohol line outgrew all the others. 5. There were no abortions, no still births and no mon- sters obtained in these experiments, nor in the former. BIBLIOGRAPHY Nice, L. B I9. e EAA Studies on the Effects of Alcohol, Nicotine, To- bacco Smoke and Caffeine on White Mice. I. Effects on Re- r 3, p. 133 pro 1912. Studies on the Effects of Alcohol, Nicotine and Caffeine on ite Mice. II. Effects on Activity. Jour. Exp. Zool., Vol. Stockard; C. R. 1912. An Experimental Study of Racial Degeneration in apeg treated with Alcohol. Archiv. Internal Med., Vol. 10, p. 1913. The Effect on the Offspring of Intoxicating the Male he and the pire Z the Defects to Subsequent Genera- dá s. AT., Vol. 47, p. 641. 1914. A mA PA of Further onnon of Mammals treated with i oc. Soc. Exp. Biol. and p 1916. A Further Analysis of the Hereditary Transmission of Degen- i and Deformities by the Descendant of Alcoholized Mam- Am. NAT., Vol. 50, p. 65. LINKAGE IN LYCOPERSICUM DONALD F. JONES CONNECTICUT AGRICULTURAL EXPERIMENT STATION, New Haven, Conn. Tue known cases of linkage of hereditary factors in plants are not as yet so numerous but that it seems de- sirable to place on record all instances of this condition. With that end in view I wish to call attention to some scattered data obtained several years ago, before much was known about linkage, and presented in publications which are probably not widely circulated. Part of the data to be considered resulted foauhs an in- vestigation, started by Hedrick and Booth, shortly after the beginning of the awakened interest in Mendelism, which was designed to test the inheritance of Mendelian characters in the garden tomato (Lycopersicum esculen- tum Mill.). The results were published in the Proceed- ings of the Society for Horticultural Science in 1907. Two different crosses were studied. One cross was made between two varieties which differed in one character only, viz., standard and dwarf habit of vine. The other cross was between two varieties which differed in three characters, habit of vine, shape. of fruit and color of fruit. It is this second cross which gives evidence that there is a genetic linkage between the factors for habit of vine and shape of fruit. In this latter cross the varieties used are known under the varietal names of Quarter Century and Yellow Pear. The Quarter Century variety is described as having a dwarf type of vine, red-colored fruit which is shaped like that of the common garden varieties, i. e., more or less spherical. The Yellow Pear variety has a standard or spreading vine, fruit yellow in color and pear-shaped. The first generation plants grown from this cross were standard in habit of vine, with red-colored fruit which differed in shape from either parent, being oval rather 608 No. 610] LINKAGE IN LYCOPERSICUM 609 than spherical, but not constricted like the pear-shaped fruit. The second generation gave the two parental types of fruit shape together with the heterozygous fruit shape in approximately the ratio of 1:2:1. The actual numbers obtained are as follows: 114 plants with fruit shaped like the Quarter Century parent, 208 plants with fruit shaped like that of the F, plants and 130 plants with pear-shaped fruit. The F, shape of fruit was considered to be more like that of the Quarter Century parent. Both types dif- fered from the Yellow Pear shape by not having the con- stricted neck. The non-constricted type is thus incom- pletely dominant over the constricted. The cross therefore received one dominant factor from one parent (standard vine) and two from the other (red and non-constricted fruit). Any linkage between the fac- tors for vine habit and shape of fruit would in this cross be a case of spurious allelomorphism or repulsion accord- ing to the English designation of this condition. The second generation segregated into 12 distinct cate- gories according to the Mendelian expectancy for a tri- hybrid where two factors show complete dominance and one factor incomplete dominance. The actual results ob- tained compared with the theoretical expectancies are tabulated by Hedrick and Booth, and given here in Table I. TABLE I HEDRICK AND BootH’s Data SHOWING THE DISTRIBUTION OF THE F, PLANTS FROM A CROSS IN WHICH THE PARENTS DIFFERED IN THREE CHARACTERS Plants Fo E Standard vine, Fruit Quarter Century shape, red....... 48 63 9/16 yellow.....°16 213/16 Fruit Hybrid shape, re loch ins T 148 127 1/8 WOU. enc. 4s 42 3/8 Fruit Yellow Pear lat i Vinay ies 98 63 9/16 ou 29 213/16 Dwarf vine, Fruit Quarter Century shape, sie ja Gee 37 213/16 yellow ...... 13 71/16 Fruit Hybrid shape, vais a ea a ee ae 17 42 3/8 MCC cia eaves 3 141/8 Fruit Yellow Pear kerai ar E ta Sit 2 213/16 fs 4 Parental combinations. Total..... 452 459 610 THE AMERICAN NATURALIST [ Vou. LI From the results as given it can be seen that the plants with the combinations of habit of vine and shape of fruit obtained in F, which duplicated the parental combina- tions, are more numerous than expected, whereas the two new combinations, with respect to these two factors, are less than expected. The writers observed these facts and commented upon them as follows: The percentage of plants which fall into each class are, however, quite different from those of Mendel. This is of importance in that it indicates the number of plants which it is necessary to grow in order to get a plant with a certain combination of characters. Theoretically, 64 plants should have included all the combinations we secured; actu- ally, with 452 plants there is one combination with only one representa- tive. In place of one representative, there should have been seven. Our results would indicate that it is necessary to raise seven times as many tomatoes as are theoretically necessary in order to secure a desired combination. There is apparently a method to this variation. The tendency seems to be for the second generation hybrids to go back to the same combinations of characters as the parents, rather than to form new ones. Thus it will be seen that the tomato with Quarter Century fruit on Yellow Pear vines is less than theoretical considerations alone would indicate, while the number of tomatoes with Quarter Century fruit on Quarter Century vines is more than theory would require; the same being true for the yellow pear. Inertia seems to be a factor and the preserva- tion of the status quo an object among tomatoes as among men (p. 23). In the light of more recent investigations of factorial linkage it is recognized at once that the above statement fulfils the conditions of linkage between at least two of the allelomorphic pairs concerned. Let us then examine the data more closely to see if a clear case of linkage can be made out. . Since the deviations above and below the expectancies are about the same in both the red-fruited and yellow- fruited plants, it indicates that color of fruit is an inde- pendent factor and that habit of vine and shape of fruit are partially linked with frequent breaks in the linkage. Combining the figures for yellow and red fruit and putting the 452 plants into 6 categories instead of 12, the con- densed results given in Table II are obtained. 1 The italics are mine. No. 610] LINKAGE IN LYCOPERSICUM 611 TABLE II DISTRIBUTION OF THE F, PLANTS WITH RESPECT TO THEIR HABIT OF VINE AND SHAPE OF FRUIT Categories Found Expected |F diE ted gomhinanone ot Standard vine, sphericalfruit 64 84} New combination F, fruit 188 1693 379| 339 Pear-shaped fruit 127 844 Parental combina- tion Dwarf vine, spherical fruit...... 50 | 28} Parental combina- tion F, fruit 20 56} 73| 113 Pear-shaped fruit........ 3 28} New combination Total 452 | 452 452| 462 These figures bring out more clearly the fact that the parental combinations are in excess while the new com- binations are deficient when compared with the theoretical expectations with independent assortment. Combining the numbers of the three types of standard plants and the three of dwarfs brings out another fact, viz.: that the standards exceed the dwarfs far more than is to be ex- pected. This result can not be accounted for on the basis of linkage, because it makes no difference whether habit of vine is or is not linked with any other factor; the ratio of the total number of standards to the total number of dwarfs should approach a ratio of 3:1, if the two char- acters form a simple allelomorphie pair free from any other complicating factors. The same deficiency of dwarfs was noted by these investigators in the other cross re- ported in both the F, generation and the F, generation - from heterozygous F, plants. The numbers they ob- tained were as follows: Found _ Expected Stone X Dwarf Aristocrat, F,: Standards ....2,289 2,176 ! nEs Dwarfs s.ro. 61 25 Stone X Dwarf Aristocrat, F,: Standards ....1,086 1,026 Dwarfs. sssi. 382. With regard to this Painia of dwarfs Hedrick and Booth suggest that . the smaller number may be due to a lesser vigor on the part of the dwarf as compared with the standard plants, and an unconscious selec- 612 THE AMERICAN NATURALIST [Vor. LI tion by the man pricking out the young plants from the seed boxes, of the larger, that is, the standard plants. This point had been antici- pated and the workmen cautioned to take the plants just as they came, but it is against all of a gardener’s training to throw aside a good vigorous plant and take one half the size. ; However, in the F, plants given above, from F, segre- gating plants, all the seeds which were planted and lived were grown to maturity, so that the latter source of error, of unequal sampling, was avoided. Still there was the same deficiency of dwarfs. Craig (1907), in the same publication, reports large numbers of the same cross which also showed a deviation in the second generation, of too many standard plants. He does not state whether or not an attempt was made to grow all the plants obtained from the seed pE His figures are as follows: Found Expected Stone X Dwarf Aristocrat, F,: Standards ....2,499 2,367 Dwaris F068: 657 789 Stone X Dwarf Aristocrat, F,: Standards .... 154 155 Dwarfs ....... 52 51 Both Halsted (1905) and Price and Drinkard (1908) give figures on the proportions of standard and dwarf plants obtained in F, populations. I have tabulated their data as follows: Halsted’s Data (pp. 450-462) Standards Dwarts Dwarf Champion X Magnus, Py... is.. iss. 5 20 Dwarf Stone X Golden Queen, F, ................ 25 5 Dwarf Stone x Extra Early Tree, F, ..........:. 14 6 x Dwarf Champion, BP, ...........; 18 3 Total COUR. cc... seo pee dk 122 34 Prpa 6 fais 5 ook see Seeds Se ee ee yee nee 117 39 Price and Drinkard’s Data (Table XI, p. 40) Standard Dwart | Dwarf Champion X Red Currant ................. 21 3 Potato Leaf x Dwarf Champion ................+ 15 9 Total found ... ee) 12 Esrpocted ...... 55% 36 12 No. 610] LINKAGE IN LYCOPERSICUM 613 In these last two tabulations many of the crosses show a deficiency of dwarfs, although the results as a whole agree closely with expectations. However, the numbers are too small to place much weight upon. In connection with another investigation I have ob- tained considerable data on the inheritance of this char- acter, by simply growing the seedlings in flats from 6 to 8 weeks, and then counting the dwarfs and standards with- out setting the plants in the field, as in all the previous cases cited. It is not always possible to distinguish all of the two types of plants, with certainty, at this stage espe- cially, if the plants are crowded and there are many small stunted plants. However, counting the plants at this time removes the possibility of unequal sampling when only a part of the seedlings are set in the field, and also the pos- sibility of differential viability in the field. The distribu- tions in 5 F, and 16 F, populations from heterozygous F, plants gave the following results: Standards |= Dwarfs Dwarf Champion X Stone, 5 F, populations .... 1,103 437 Dwarf Champion Earliana, 1 F, population... 186 67 Dwarf Champion X Stone, 12 F, populations ... 1,707 730 Dwarf Champion X Earliana, 4 F, populations .. 571 149 COLNE COUN eker ho uy 6c eaan ay vee 3,567 1,383 TOROS C6. LEGO Uso e te sich ieee ee 3,713 1,237 Here the deviation from expectation is in the opposite direction. There is an excess of dwarfs. It would seem that too many of the small plants were classified as dwarfs when they were really standards.. Two of the above F, populations were grown longer than the others in flats which were not so crowded, so that the errors in classification, I believe, were more nearly overcome. The following results were obtained: Dwart Champion X Stone, F, .:........0052e00es 268 88 Dwarf Champion X Earliana, F; ...........++-++ 186 67 FORE SOURS Ae UE AE 454 155 614 THE AMERICAN NATURALIST [Vou. LI From these data it seems justifiable to conclude that dwarfness and standardness form a simple allelomorphic pair, free from any genetically complicating factors. I have gone to this length to demonstrate the normal behavior of this character in order to be able to correct Hedrick and Booth’s data according to the proportion of dwarf and standard plants, which presumably. they should have obtained if all the plants had been grown to maturity, and if there had been equal viability. More- over, whether or not the deficiency of dwarfs which they obtained is due to unequal sampling, differential viability or some unknown cause, there is no reason to suppose that the cause, whatever it is, has anything to do with the linkage between the factors for habit of vine and shape of fruit. I have, therefore, in Table III increased the num- TABLE III CORRECTED DISTRIBUTION OF THE F, PLANTS WITH RESPECT TO THEIR HABIT OF VINE AND SHAPE OF FRUIT—CHARACTERS WHICH SHOW LINKAGE ad Combinations of y Characters S] kei For- Characters of F2 Plants PER = bad A; Oo 284 New combinatio AB.... Standard vine, oon ead Ese -m ce 95 Parental combination Ab..... Standard vine, constricted fruit...... 7 aB.....|Dwa vine Danii. pre 70| 85121 95 Parental combination sb... Dwarf vine, constricted fruit. ......... 3 28 ew combination al ai ...'452 EEP SE, ber of dwarfs to the number theoretically expected, keep- ing the proportion of the two different kinds of dwarfs the same with respect to shape of fruit. From Table II it can be noted that 379 standards were obtained. The- oretically the dwarfs should have been one third of this number, or 126.3. There were actually only 73. This number would have to be increased 1.73 times in order to bring the number of dwarfs up to the expected number. Combining both the standard and dwarf plants in two classes each, those with and those without constricted fruit, and multiplying these two classes of dwarf plants No. 610] LINKAGE IN LYCOPERSICUM 615 by 1.73, the figures given in column 3 of Table III are ob- tained. The figures in this column represent the number of plants which presumably should have been obtained in the four different categories, if the expected number of dwarf plants had been obtained. . These corrected numbers can then be compared with the closest theoretical ratio where the gametes, instead of being produced in the equal proportion of 1 AB:1 Ab: 1 aB:1 ab, were produced in unequal proportions (where A and B represent the two dominant factors—standard vine and non-constricted fruit). In this case if the gametes were formed in the proportion of 1 AB:4 Ab: 4 aB:1 ab, the agreement between the corrected result and the theoretical expectation is surprisingly close. Corrected numbers ....... ron Be eos 262 9) 27 FSF :5 WOPPOCIed PALO... ore ics gute bole ae Poulos. 50.4: 25.4:° 24.2: 1 Theoretical TAGON ior VAS ocins cs awe SE 24 24 :1 (1: 4: 4: 1 gametie series) Tt is seen that the data obtained by Hedrick and Booth give a clear indication of linkage of the factor for stand- ard vine with that for constricted fruit and dwarf vine with non-constricted fruit. Frequent breaks in the linkage occur to form the two new combinations. On the chromo- some hypothesis the data show, in this case, that crossing over occurs in 20 per cent. of the gametes formed. TABLE IV CORRECTED DISTRIBUTION OF THE F, PLANTS WITH RESPECT TO THEIR IT OF VINE AND COLOR OF FruritT—CHARACTERS WHICH Do not SHow LINKAGE ig $ kzi 4 Bers Characters of Fe Plants ‘3 $ $3 4 TOE mulæ a R S| # i S oj A ‘AW..,.\Staridard vine, red hait 254 204 284 Sau otidan Ab..... Standard vine, yellow fruit........... 85) 85 85) 95/Parental combination aB.....| Dwarf vine, red fruit 6 85 97 95 Parental combination ab: Dwarf vine, yellow fruit. 17, 28| 29| 31|New combination Total 1505/505) i | | 616 THE AMERICAN NATURALIST [Vou. LI The data also show that there is no linkage between the other two combinations of factors reported, viz., vine habit and fruit color, and fruit color and fruit shape. Cor- recting the number of dwarfs in the same way as in Table TII the results for these two combinations of factors are given in Tables IV and V. TABLE V CORRECTED? DISTRIBUTION OF THE F, PLANTS WITH RESPECT TO THEIR SHAPE RUIT AND COLOR OF FRUIT—CHARACTERS WHICH Do not SHow LINKAGE | Genetic p x] 3| E g bs : 5 | Combinations of oe Characters of Fz Plants ag 2j E ~+ A Chardenels EA 8) al AB....|Non-constricted, red fruit 250 254 (289 284| Parental combination Ab...../Non-constricted, yellow fruit......... $ | 84| 95| New combination aB..2..| Constricted, red fruit 00| 85101) 95 ‘New combination SD. -srs Constricted, yellow fruit 30 28| 31| 31 Parental combination | Total 459 452 R05 505! | | EET From these two tabulations it will be seen that the agreement between expectation and observation, when the number of dwarfs is increased to the number expected, is reasonably close, and the deviation from the expected is not such as to suggest linkage between any of these factors. Both Halsted, and Price and Drinkard, in the publica- tions previously mentioned, give a large number of crosses of tomatoes where the inheritance of many dif- ferent characters are studied. Unfortunately, in most cases the data are presented in such a way as to show the inheritance of only one character pair at a time. Halsted gives a dihybrid cross between two varieties differing in habit of vine—standard (A) and dwarf (a), and margin of leaf—serrate (B) and entire (b). The 2 Since habit of vine is not concerned in this cross it is, of course, un- necessary to correct for the low number of dwarfs as in the two previous tables. I have done so simply to show that it does not affect the goodness of fit to any great extent. No. 610] LINKAGE IN LYCOPERSICUM 617 cross was made in such a way that one dominant factor entered from each parent. The numbers obtained AB: Ab: aB:ab OURO ree eee seks heb AN see ee cs Fhe 49:16: 13:7 Mipaeted.. ies 4,08 Fes ES. oh POT oA it 48:16: 16:5 do not indicate any linkage between these factors. Price and Drinkard’s data indicate that there is no linkage between shape of fruit and color of fruit in two different crosses (agreeing with the data given in Table V), none between foliage color and fruit color, and none between foliage color and fruit shape. In these crosses the numbers are too small to be sure of the conclusions with regard to linkage. They give the results of a cross, however, which shows complete linkage between green foliage color and two-celled fruit, as opposed to yellow foliage color and many-celled fruit. Only 24 F, plants were grown, which were of two types only, duplicating the parents. These characters, foliage color and loculation of ovary, can not be the expression of the same factor because many varieties are known with these characters combined in the other ways. In fact the majority of the common garden varieties have green foliage and many-celled fruit. Neither does it seem probable that these dissimilar char- acters form a series of multiple allelomorphs as some cases of complete linkage, for instance, cob and pericarp colors in maize, are considered to be. Although the num- ber of plants is small, as the writers state, it would seem that among 24 plants at least one new combination would appear if the factors were independent of each other. Larger numbers of a similar cross, studied by back crosses in the more favorable way, will probably show these factors to be partially linked. Crane (1915) reports a cross between varieties of toma- toes differing in rather complex characters of inflores- cence and fruit shape. He obtained figures which indicate partial linkage in these characters, but states that ‘‘the 618 THE AMERICAN NATURALIST [ Vou. LI ‘numbers are not sufficiently large to form any conclusion as to the intensity of the coupling, nor to establish the existence of the same with certainty.’’ A number of clearly segregating characters are known in the tomato. Halsted lists 7 alternative unit character pairs, while Price and Drinkard give 13. However, from their own statements in regard to the behavior of these characters, and from my own rather limited experience with tomatoes, the number of different character pairs which they list should be reduced. For instance, only two allelomorphic pairs are known for color of fruit, viz., red and yellow flesh or endocarp, and yellow and colorless fruit skin or epicarp, while Price and Drinkard give four, and Halsted three, character pairs of fruit color. Different combinations of skin colors and flesh colors give the dif- ferent colored fruits. For example, colorless epicarp over red endocarp gives pink-colored fruit. TABLE VI MENDELIAN CHARACTERS IN THE GARDEN TOMAT (Revised from a lists given by Halsted and by Price nek Drinkard. ) Dominant | Recessive Fruit hapt tA 1 Spherical (non-con- Pyriform (con- tricted) stricted ) ....:- molars tritubercular. * The upper levels of the Almagre of New Mexico are perhaps equivalent to the Lysite. Granger, Bull. Amer. Mus. Nat. Hist., Vol. XXXIII, 1914, p. 207. ; In certain cases the paleontological evidence is in- decisive, as between the hypothesis. of successive muta- tion in loco and the hypothesis of continuous evolution in an unknown center of evolution followed by. discon- tinuous immigration of later stages into the region under observation. Such a case is described by Matthew® as follows: 6 Bull, Amer. Mus. Nat. Hist., Vol. XXXIV, 1915, p. 316. 632 THE AMERICAN NATURALIST [Vou. LI Osborn in 1902 pointed out the evolutionary progress observable in the species of Hyopsodus from successive stages of the Lower and Middle Eocene; this is in general confirmed and extended by the far larger collections [comprising more than a thousand specimens] now available and the somewhat wider geologic range of the genus; but it is evident that not one but three or four phyla are present in each horizon; the relations of the Lower Eocene species to those of the Middle Eocene are not wholly clear, and the geological overlap of stages of each struc- tural phylum suggests rather progressive displacement of older by newer stages coming in from some other region, than gradual evolution in loco. It might equally well be interpreted as the displacement of older by newer “ mutants,” in the DeVriesian sense of the term. However this may be, the Lower Eocene species are distinguished from those of the Middle Eocene by the less molariform premolars, and this is most noticeable in H. eis from fhe: lowest horizon, while the Lost Cabin species [from the up of the Lower Eocene] approach nearest to those of the Digar gi nii: Eocene]. Examples of this kind might be multiplied, tending to show that the evolution of Tertiary mammals has often been more or less continuoùs, or by small successive changes, at least during the relatively brief geological periods that are represented by a large series of speci- mens from closely sequent levels of an uninterrupted stratigraphic series. And although mutations may well be a paleontological reality, there is little danger that ver- tebrate paleontologists are likely to draw false inferences regarding the history of structures and of races through mistaking independent contemporaneous mutants for suc- cessive stages, for the simple reason that their observa- tions are based on long series of specimens which are ar- ranged in their true chronological sequence, from ascend- ing geological horizons covering the whole Tertiary eriod. In this connection I submit an accurate diagram (Fig. 1) by Mr. Granger, which is fairly representative of the kind of evolution demonstrated among many, but not all, known races of mammals during the Tertiary and Quater- nary Periods, a period of time conservatively estimated at 4,000,000 years. The character of the evidence tending to show that the paleontologist is dealing with truly successive stages and No. 610] GENETICS VERSUS PALEONTOLOGY 633 y napia caballus Recent. anes EGE oe pe Pea complicatus Pleistocene. -4 FA Hypohippus sp. Loup Fork. 4 T _ Mesohippus sp. White River. 4 A Epihippus uintensis Uinta. Orohippus uintanus Bridger. 4 Eohippus venticolus Wind River. 2 > Eohippus cristonensis Wasatch. Frc. 1. ecurate outlines of lower upea aja of Equidæ. peda pank point of mtg Pro apaidans e diameter i series, ranging ca eng hird molar in the Wasatch species to the third oe in the modern hen After W, pira een! gnati Mus. Nat. Hist., Vol. XXIV, Art. XV, si a ies. 1908.) 634 “ THE AMERICAN NATURALIST [ Vou. LI not with an -arbitrarily selected series of mutants is further illustrated in the following note by Dr. Matthew. Of the hundreds of specimens examined no horse from the Lower Eocene has ever been found which had any fully molariform premolars. No horse out of the Middle Eocene has either more or less than one molariform premolar in the lower jaw, on each side. Out of the Upper Eocene all horses have two molariform premolars. In the Oligocene all have three. All Oligocene and older horses have brachyodont molars without cement. All Miocene horses are progressively hypsodont with a progressive increase in the amount of cement. The milk teeth of Miocene horses have almost no cement. Those of all Pliocene and later horses are heavily cemented. At each successive stage of evolution the cement appears at an earlier stage in the ontogeny of the tooth. These are simply a few out of many progressive changes in the teeth, and they are accompanied by equally clear progressive changes in the skull and skeleton. Every one of these progressive stages is as exactly lim- ited in time as the ones cited. Geneticists who are examining the nature of the paleon- tological evidence regarding modes of evolution would do well to realize that only a small part of the available ma- terial bearing on the subject is either exhibited or pub- lished. The scientific staff of the American Museum of Natural History would be very glad to exhibit to their col- leagues the great wealth of accurate data, concerning the chronological sequence of specimens, which has been gath- ered during twenty years of close exploration; they would also be pleased to place before them any of the extensive series of specimens, sometimes amounting to several thousands of individuals, which appear to throw light upon the problem of continuity vs. discontinuity. In conclusion, paleontologists can show that evolution- ary changes have involved progressive and measurable emphases or suppressions of earlier structures or of ear- lier proportions (allometric evolution, Osborn); and when the progressive emphases are manifested as focal outgrowths they seem like ‘‘new’’ structures (rectigrada- tions, Osborn). Paleontologists, however, are not in a position to say which characters would be transmitted according to the Mendelian ratio, nor can they prove what were the cytological causes of the evolutionary changes No. 610] GENETICS VERSUS PALEONTOLOGY 635 which they record or infer. In that direction lies oppor- tunity for consultation with the men who study enzymes, chromosomes, heredity and variation. The Batesonian hypothesis that both the progressive differential emphases or suppressions of organs, and the focal outgrowth of new structures, have been due to a secular, differential stopping down of inhibitory factors inherent in the germ-cells seems to the present writer quite consistent with the observed facts of evolutionary change; but apparently no observations that the paleon- tologist can make could furnish any critical tests of this hypothesis; it therefore has for him a stimulative philo- sophical value, but hardly constitutes a working hypothe- sis for the discovery of new facts and principles in his limited field. The nature of later events being determined in part by the nature of their precedent events, no matter how many causal series may be interwoven in the final outcome, it follows that paleontologists, like other historians, con- tribute to a partial understanding of existing conditions merely by arranging past events in their true chronolog- ical sequence. The characteristics of existing Cetacea are determined in part by the germinal and somatic charac- teristics of their remote quadrupedal ancestors, as well as by the conditions of the pelagic life into which they somehow drifted; so too the characteristics of man, as a bipedal, hiwanous. anthropoid Primate are determined in part, as I believe, by the fact that the remote ancestors of the man-greatape stock were arboreal, quadrumanous, lemuroid Primates of the Lower Eocene. For such reasons, I must continue to hold that ‘‘pro- gressive adaptation’’ when cleared of all implications as to the mode of evolution, stands for a historical and verifiable process; that the time for developing phylo- genetic conclusions and for revising comparative anat- omy and classification is always now, as fast as the evi- dence can be gathered and analyzed. SHORTER ARTICLES AND DISCUSSION STUDIES ON INBREEDING. VIII. A SINGLE NUMER- ICAL MEASURE OF THE TOTAL AMOUNT OF INBREEDING* 1. In the earlier numbers of these Studies, and particularly in VII, methods have been given for measuring the amount or de- gree of inbreeding exhibited in a particular pedigree by a series of inbreeding coefficients, Zi, Za Za . . ., Zn, one for each àn- cestral generation. The inbreeding for the whole pedigree is indicated by an inbreeding curve, formed by plotting and con- necting by a line the several coefficients. 2. From the earliest stages of this investigation the writer has - been aware of the desirability of a single numerical measure, to supplement or replace the inbreeding curve as a designation of the total inbreeding exhibited. Such a designation has now been found, which, it is believed, uniquely and rigorously meets the requirements. It is the purpose of the present paper to describe this new constant. Í Consider Fig. 1. This gives, in the heavy line and solid circles, the inbreeding curve for 9 ancestral generations of the Brown Swiss bull, Saxton (2668). The values of the inbreeding co- efficients are: Z,.= 0, Z,=12.50, Z, = 26.95, J. == 0; Z,=17.19, Ž, = 28.91, * AE - B= 21.09, Z,= 29.30. The smooth curve of Fig. 1 is the inbreeding curve for con- tinued brother X sister mating. This represents the closest or maximum degree of inbreeding possible in sexually reproducing organisms. It is clear from inspection of thin diagram, that Saxton is much less intensely inbred in fact than he would be if in all his ancestry the matings had been of brother sister out of brother 1 Papers from the Biological Laboratory of the Maine Agricultural Ex- periment Station. No. 118. 2 Pearl, R., AMER. NAT., 1917, in press. 3 Cf. Pearl, R., these Studies, I. Amer. NAT., Vol. XLVII., p. 603, 1913. No.610] SHORTER ARTICLES AND DISCUSSION 637 X sister, etc. This is evident, in the first place, because the or- dinates of the Saxton curve, a c are nowhere as high as those of the brother X sister curve, a b. But it would also be equally clear that Saxton-was less inbred than the maximum possible amount if the last ordinate at c, for example, had a value of 99.6, as does the corresponding ordinate of the maximum curve. * Me g COEFFICIENTS aa ea é é : 70 12 “4 GENERATIONS Diagram showing the inbreeding curve of Saxton, a Brown Swiss bull, in the heavy lines and solid circles. The smooth, light line curve is the curve of maximum inbreeding (continued brother x sister breeding). For further ciple nti: see text Upon consideration it appears that the real measure of com- parative amount or degree of inbreeding, considering the gree as a whole, is given by the area included by the hates inbreeding curve under discussion, as compared with the corre- sponding area of the maximum (brother X sister) curve. Thus in Fig. 1, Saxton is less inbred than the maximum possible amount to an extent proportionate to the amount by which the area T (a c d) is smaller than the area M (a b d). This con- sideration gives us the desired method of uniquely expressing the total amount or degree of inbreeding. It only remains to consider practical methods of calculation. , : 3. Theoretically one should integrate the maximum (brother X sister) curve, and the observed curve, and compare the areas derived from such integrations. Practically this is not possible, - because many observed curves of inbreeding can not be fitted by 638 THE AMERICAN NATURALIST [ Von. LI any simple, readily integrable theoretical curves. We shall hence be compelled in this case to make use of the expedient so fre- quently employed in applied mathematical problems of all sorts, namely, to take finite summation as a sufficiently close approx- imation to integration. Doing so, we may take as the simple ex- pression of total inbreeding, to and including the nth generation, the following: 100° Fy.’ : where 3 denotes summation of all values between the inclusive limits indicated, and F z, is a constant having the value set forth in Table I. Fz, is of course the total area of the maximum brother X sister curve up to and including the n -+ 1-th genera- tion. Since these successive values are constant they may be tabled once for all. TABLE I VALUES OF F 7, THE wg soe AREAS OF THE MAXIMUM INBREEDING CURVE Ancestral generation F. Ta E Se oe a 4 50 e E Me gE 2 125 Be DA Vr 3 212.5 OAT ey TE P 4 306.25 Gare 5 403.125 Ree. see a 6 501.5625 AEF ay Saati, So a 7 600.78125 1 PSS iat LE E 8 700.390625 s LE EE ae N ie 9 800.19531254 EE. E Saray 10 900.09765625 ee eee eee. 11 1000.048828125 Oe oa ee ee ee 12 1100.0244140625 Pe Oy UO. s 13 1200.01220703125 1S BoE SARS 14 1300.006103515625 Bis 6 BAR oe = 15 -1400.0030517578125 In using the form of total inbreeding coefficients shown in (i) there is one caution which must be carefully observed. This is that only so many generations should be used as to include the one in which the observed Z taken first reaches its highest value for the pedigree under discussion, and not any beyond that one. This will usually be for the earliest ancestral generation of the pedigree, but not always. , #One would not, of course, use in practical calculation such excessive numbers of decimals as are tabled from this point on No. 610] SHORTER ARTICLES AND DISCUSSION 639 4. We may now consider some numerical illustrations. Let us take first the bull Saxton, for which the several observed in- breeding coefficients have already been given. We have 22 = 142.19 and hence, from Table I, 14219 ; Zn = 800.1953 = 17.8. Or we may say that Saxton is inbred in ten ancestral genera- tions, taken together 17.8 per cent. of the maximum amount pos- sible in those generations. For comparison some other figures may be examined. Pearl and Patterson’ have given mean values of the inbreeding co- efficients for four groups of Jersey cattle: (a) Random sample bulls, (b) register of merit bulls, (c) random sample cows, (d) register of merit cows. It will be of interest to reexamine these figures by the method here described. The results are given in Table II. TABLE II i TOTAL INBREEDING COEFFICIENTS FOR JERSEY CATTLE. (PEARL AND PATTER- son DATA) _ Total Inbreeding Coefficients Zy, units | Lower Limiting | Upper Limiti er venues r "alan sa Vaca 1. General population (pasaon sample) bulls| 25.39 30.48 27.94 2. Register of merit bul 24.52 29.17 31.85 3. General aserra ( wii sample) cows) 27.46 “oy 74 29.60 4. Register of merit c 20.72 27.08 23.90 From this table we see that American Jersey cattle, as judged by random samples of the general population, are about 28 to 30 per cent. as closely inbred as the maximum possible amount, taking account of the first eight ancestral generations as a whole. It is not desirable to go further into the discussion of these Jersey data, since the purpose of this note is simply an exposition of method. This new method makes possible exact and unique numerical comparison between pedigrees in respect of the degree of inbreeding which they exhibit in the same i of ancestral generations. RAYMOND PEARL 5 Pearl, R., and Patterson, S. W., Proc. Nat. Acad. Sci., Vol. 2, p. 60, 1916. eee pari Ag Š THE AMERICAN NATURALIST Vou. LI. November, 1917 bi. No, G11 THE GENESIS OF THE ORGANIZATION OF THE. INSECT EGG! PROFESSOR ROBERT W. HEGNER ZOOLOGICAL LABORATORY, UNIVERSITY OF MICHIGAN I. THE COMPLEXITY oF ORGANIZATION OF THE Insect Eca -~ 1. Introduction THE morphological and experimental investigations of the germ-cell cycle in insects which the writer has carried on during the past ten years have resulted in the accumu- lation of many data which indicate the complexity of or- ganization of the eggs of these animals, and suggest hypotheses regarding the nature and genesis of this or- ganization. That the animal egg at the time develop- ment begins does not consist of a homogeneous mass of protoplasm, as the old theory of epigenesis required, but is a highly organized cell containing various kinds of protoplasm localized in definite regions has been proved conclusively by numerous investigators working with the eggs of many different species. The degree of organiza- ‘tion at the time of fertilization varies according to the species of animal, but all embryologists admit that the insect egg is one of the most highly organized of all. The nature and genesis of the different kinds of proto- . plasm in the insect egg is the principal problem discussed- in the following pages. This problem is the logical suc- cessor of those dealing with cell lineage and the organiza- tion of the egg at the time development begins. The re- 1 Presented to the Johns Hopkins Scientific Association on October 9, 1917. 641 642 THE AMERICAN NATURALIST * [VoL UI lations between this problem and the larger problem of heredity and development are very close indeed. The study of heredity is concerned not only with the adult animal, but also with every stage in the development of the adult since the fertilized egg, or embryo, or larva that arises from it, is an individual just as is the fully devel- oped animal. The organization of the fertilized egg is the result of the processes of differentiation that take place at each stage in the history of the egg and sperm from the time the primordial germ cells are segregated until the highly specialized gametes have become fully formed. The period extending from-the formation of the primordial germ cells to and through the growth period of the gametes is one of the least known in the entire his- tory of the individual. It is nevertheless a most im- portant period, for during this time, at least in the insect egg, the principal axes of the individual are established and different kinds of cytoplasm are elaborated and lo- ealized that are predetermined to form definite parts of the embryo. The stages in the germ-cell cycle in insects belonging to different orders, families, genera or species are often quite different, as is to be expected, hence the data on which this paper is based were derived from the study of a number of species. On this account it seems best to give an abridged description of the germ-cell cycle in one group and include, wherever necessary, data derived from the study of other groups. The principal work has been carried on with representatives of the orders Diptera, Hymenoptera and Coleoptera, and of these, the order Coleoptera has furnished the best material for experi- mental purposes. We will select, therefore, the eggs of certain chrysomelid beetles for deseriptive purposes. 2. The Structure of the Insect Egg at the Time of Depo- sition’ At the time of deposition (Fig. 1) the beetle’s egg con- sists largely of deutoplasm—a substance which is used ; 1 Hegner, 1909, Journ. Exp. Zool., Vol. 6. No. 611] GENESIS OF ORGANIZATION OF INSECT EGG 648 up during the growth of the embryo. This deutoplasm is composed of vitelline spheres, which contain refringent granules, the vitelline bodies, and of oil glob- ules. The deutoplasmic bodies are embedded in a viscid cytoplasmic matrix which, however, is very slight in amount as compared with the deutoplasmie material. At the periphery of the egg is a thin cortical layer of cytoplasm which is continuous with the cytoplasm in which the vitelline spheres lie. A short distance back of the anterior pole of the egg is a thickening of this cortical layer in which the maturation divisions of the oocyte nucleus occur. The cy- toplasm appears to be homogeneous. except at the posterior end, at which place, in many insects, inclusions have been discovered which appear to play a rôle in the formation: of the primordial germ cells.? terior end. = yolk. (From Hegner, 1909.) proren, Lae Fig, 1. Longitudinal section B ugh an egg of Calligrapha bigsbyana four hours ` deposition. gcd = germ-cell deter- ants. gn = germ-nuclei copulating. khb? pepr layer of cytoplasm. p = pos- om = vitelline bi, Yy These inclusions in the w somelid beetles take the form of a polar disc of granules which I have called germ-cell, keimbahn, or germ-line determinants.? 2 Hegner, 1909, Jou Similar inclusions (Fig. 2) have been Morph., Vol. 20. 3 Hegner, 1908, Biot, 2 Bull., Vol. 16. 4 Weisman, 18638, Zeit. f. wiss. Zool., Bd. 13; Metschnikoff, 1866, ibid., 644 THE AMERICAN NATURALIST [Vou. LI noted in Diptera,‘ in parasitic Hymenoptera,’ and in Hymenopterous gall flies.6 There should also be men- tioned in this connection minute bodies that have been Cc Í Fig. 2. Germ-line determinants in .the eggs of various animals. «a. Pole- plasm (p Pl) at the posterior end of the egg of Miastor (Hegner, 1914). 0. £ .) f. Ectosomen at end of the first cleavage spindle in the egg of Cyclops. (Amma, 1911.) discovered in the cytoplasm of the eggs of the carpenter ant, Camponotus, and in those of various other insects. Bd. 16; Ritter, 1890, ibid., Bd. 50; Noack, 1901, ibid., Bd. 70; Kahle, 1908, Zoologica, Bd. 21; Hasper, 1911, Zool. Jahrb., Bd. 31; Hegner, 1912, Science, Vol. 36; Wotan 1914, Journ. Morph., Vol. 5 Silvestri, 1906, Boll. Labor. Zool. E. Se. Agr. Portici, Vol. 1; Silvestri, 1908, ibid., Vol. 3; Silvestri, 1914, Anat. Anz., Ba. 47; Silvestri, 1915, Boll. Labor. Zool. R. Se. Agr. Portici, Vol. 10; Silvestri, 1916, Rend. D. R. Accad. D. Finit ci 25; Martin, 1914, Zeit. f. wiss. Zool, Bd. 110; Hegner, 1914, Anat. Anz., Ba. 46; Hagas, 1915, Journ. Morph., Vol. 26. 6 Hegner, 1915, Journ. Morph., Vol. eae, 1886, Festsch. nat med. Verein zu Heidelberg; Sule, 1906, No. 611] GENESIS OF ORGANIZATION OF INSECT EGG 645 These have been considered symbiotic bacteria, but their true nature remains yet to be definitely established. 3. Cleavage The first cleavage nucleus of the beetle’s egg lies some- what anterior of the center in a small island of cytoplasm that is continuous with the cytoplasm that surrounds the deutoplasmic bodies (Fig. 1, gn). During early cleavage no cell walls are formed, but after each division the daughter nuclei move a short distance apart and then divide again. Successive divisions and migrations of the cleavage nuclei (Fig. 3) finally result in the production of = pole disc, lastoderm stare. At the pésterior end are the primordial germ cells. (Heg- ner, 1909, 1914 ‘hundreds of nuclei, which come to lie just beneath the cortical layer of cytoplasm, and are each surrounded by an irregular mass of cytoplasm. The fusion of these cytoplasmic masses with the cortical layer then takes place, followed by the intervention of cell walls, thus form- ing a blastoderm of a single layer of cells, each of which contains a cleavage nucleus, part of the cytoplasm which it brought to the periphery with it, and a portion of the Stzber. bohm. Geselisch. Wiss. Prag.; Sule, 1910, ibid. ; Merceir, 1907, Arch. Protistenk., Bd. 9; Pierantoni, 1910, Zool. Anz Ba; 36; Buchner, 1912, Arch. Protistenk., Bd. 26; aigis, 1913, Bull. IU. St. Lab. Nat. Hist., Vol. 9; Hegner, 1915, Journ. Morph., Vol. 26 646 THE AMERICAN NATURALIST [Von. LI cortical layer. Not all of the cleavage nuclei take part in blastoderm formation, many of them remaining behind in the yolk to aid in breaking down this substance. 4. The Origin of the Primordial Germ Cells Blastoderm formation is interrupted at the posterior end of certain chrysomelid beetles’ eggs by the segrega- tion of the primordial germ cells (Fig. 4). Those cleav- mordial germ cells entirely separated from the egg. (Hegner, 1909.) age nuclei that encounter the granules of the pole-disc do not produce blastoderm cells, but continue their migra- tion and are finally cut off from the rest of the egg as distinct cells. These are the primordial germ cells, of which there are sixteen. Each of these cells contains a portion of the cortical layer that includes pole-dise gran- ` No. 611] GENESIS OF ORGANIZATION OF INSECT EGG 647 ules, thus differing in content from the blastoderm cells. - They differ also from blastoderm cells in size, being con- siderably larger. This is probably due at first to the in- clusion of the pole-dise granules, but later a greater dif- ference in size is brought about by the failure of the germ cells to divide as rapidly as do the blastoderm cells. An early origin of the primordial germ cells in a simi- lar manner has been described in a number of insects be- sides Coleoptera, especially in the midge, Chironomus,’ where these cells are derived from one of the cleavage nuclei at the four-cell stage; in the pedogenetic larva of the fly, Miastor,? where one of the first eight cleavage nuclei becomes the nucleus of the primordial germ cell, in the fly, Calliphora,° and in parasitic Hymenoptera." 5. The Formation of the Ovaries In chrysomelid beetles, Chironomus, Miastor and cer- tain other species of insects, the primordial germ cells undergo a multiplication period shortly after they are formed. This is followed by a period during which they become lodged within the embryo—either by the shifting of the surrounding tissues or by migration or by both these processes. At this time also they become separated into two groups; in chrysomelid beetles each group ap- pears to contain thirty-two germ cells; in Miastor each consists of four germ cells. One group becomes located on either side of the embryo and later gives rise to one half of the germ glands. The sex of the individual can be determined by the morphology of the germ glands be- fore the young hatches. The further history of the germ cells in female insects is in general as follows. From each primitive ovary a number of ovarian tubules arise each containing many germ cells (oogonia) which have undergone a multiplica- tion period. The oogonia pee cease dividing and the 8 Hasper, 1911, sap Jahrb., Bd. ® Kahle, 1908, p sae logica, Bd. 21; Hegner 1914, Journ. Morph., Vol. 25. 10 Noack, 1901, Zeit. f. wiss. Zool., Bd. 70. 11 Silvestri, l. c. 648 THE AMERICAN NATURALIST [Vou. LI ultimate oogonia are ready to enter upon the growth period. A period of differentiation may or may not in- tervene, according to the species, during which nurse cells are formed. When the oocytes have reached their full size they separate from the ovarian tubule, pass down ‘the oviduct into the vagina and are deposited. Each egg is surrounded by two membranes; a thin inner vitelline membrane and a thicker, outer membrane, the chorion. 6. The Complexity of Organization of the Insect Egg. (a) Comparison between Eggs of Insects and Those of Other Animals.—Insect eggs differ greatly from those usually employed for the study of egg organization, since they are, as a rule, laid in the air and not in the water, and because cleavage is of the superficial type, cell walls being absent until a comparatively late cleavage stage. - The eggs of chrysomelid beetles are particularly favor- able for study, since they may be subjected to the most violent experimental conditions without preventing their development.'? In insect eggs the character of the blastoderm cells depends, as in holoblastic eggs, upon the kinds of proto- plasm they contain, but all those phenomena connected with the position of the cleavage spindle, which have been so carefully studied in the eggs of mollusks, worms, ascidians and other animals, can have no influence upon the localization of different substances in various parts of an insect egg, because in the latter the volume of the egg is thousands of times greater than that of the cleavage spindle. Furthermore in holoblastic eggs differentiated substances are segregated in different cells during early cleavage and are there isolated by cell walls, and to this isolation is attributed in large part the progress of dif- ferentiation; but in the insect egg the different kinds of cytoplasm are in direct continuity until hundreds of cleav- age nuclei are present, and are not separated by cell walls until the blastoderm is fully formed. 12 Hegner, 1908, Biol. Bull., Vol. 16; 1909, Journ. Exp. Zool., Vol. 6; 1911, Biol. Bull., Vol. 2 No. 611] GENESIS OF ORGANIZATION OF INSECT EGG 649 (b) Results of Experiments with Gravity and Centrif- ugal Force.—A chrysomelid beetle, such as Calligrapha or Leptinotarsa, during the process of egg-laying clings to the under side of a leaf and the end of the egg that emerges first is glued to the leaf by a viscid secretion. Then the egg is pushed back away from the abdomen and another is laid'® (Fig. 5). In this way from four to Fie. 5. Diagram showing a chrysomelid beetle, Calligrapha bdigsbyana, clinging to the underside of a willow leaf and layin er eggs. The relation between the orientation of the egg before and after deposition is indicated by the letters. a= anterior. d= dorsal. = left. p = posterior. r= right anterior ventral surface where a spot of India ink was placed as a guide for orienting the eggs during experiments. (Hegner, 1909.) eighty eggs are laid in one group within a period of about an hour. These eggs hatch in approximately five days. A few hours before they hatch the young can be seen dis- tinctly through the semi-transparent egg shell. An exam- ination of hundreds of eggs at this stage in their develop- ment has established the fact that the posterior end of the egg is attached to the leaf and the anterior end is free. In every other respect the orientation of the young in the egg corresponds to that of the egg as it lay within the body of the mother before deposition; that is, the ends and various surfaces of the egg are definitely determined before deposition and correspond to the orientation of the mother as indicated in the diagram (Fig. 5). This rigid correspondence between the orientation of the egg and that of the adult is known as the ‘‘law of orientation’? which was first discovered by Hallez in 1886. Since the eggs of these beetles are usually attached to 18 Hegner, 1909, Journ. Exp. Zool., Vol. 6. : 650 THE AMERICAN NATURALIST [Vou. LI the underside of leaves, it was suggested that the orienta- tion of the young might depend upon the force of gravity, but eggs that were first marked with India ink and then placed in every conceivable position with respect to grav- ity proceeded to develop as though undisturbed.’* It seemed from this, therefore, that the position of the young must be predetermined in the undeveloped egg. Several kinds of experiments were performed in order to discover the complexity and fixity of this apparent organization. First, the eggs were subjected to a force greater than gravity by means of a centrifugal machine. Hun- dreds of eggs were revolved at different rates of speed for vari- ous lengths of time and in many different positions. A description of one experiment will serve to illustrate the results obtained.’® In this experiment freshly laid eggs were placed in cavities in a block of paraffin with the pos- terior end toward the center of rotation, and were revolved in a hydraulic centrifuge for sixteen hours. The heavier substances were thrown to the outer end and p. Side view of a freshly laid egg of Calligrapha multipunctata, which was cen- trifuged for sixteen hours and aken out and allowed to the lighter protoplasm accumu- lated at the inner end, where an embryo developed (Fig. 6). It is perfectly evident that the pro- toplasm from the various parts of the egg has, in its new posi- tion, developed into the tissue that it would have given rise to if it had been left undisturbed. Normally the yolk would _ be surrounded by the embryonic tissue and would be en- 14 Wheeler, 1889, Journ. Morph., Vol. 3; Hegner, 1909, Journ. Exp. Zool., Vol. 6. * ee 15 Hegner, 1909, Journ. Exp. Zool., Vol. 6. No. 611] GENESIS OF ORGANIZATION OF INSECT EGG 651 closed by the mid-intestine, but in this case a dwarf em- bryo has developed without growing around the nutritive material. The effects of centrifugal force upon insect eggs are. different from those produced upon the other types of eggs that have been employed for such experiments. In the eggs of worms,'® mollusks,” etc., apparently the ma- terials that undergo stratification under the influence of - centrifugal force have no influence upon the ‘‘ground sub- stance” which is ‘‘the seat of polarity and pattern of organization of the cell.’’ In the insect egg, the organ- ized protoplasm is almost entirely limited to the cortical layer and this layer may be shifted away from the periphery by a sufficient force and may become massed at the inner light end when an undeveloped egg is cen- trifuged. Since the cytoplasm develops in its new situation and proceeds to build up an embryo as nearly normal as is possible under the conditions imposed upon it, it is evi- dent that the potencies of the cytoplasmic areas are pre- determined at the time the egg is laid. It was hoped by means of these experiments with cen- trifugal force to throw the pole-dise granules and the cytoplasm containing them into some other part of the egg. If the germ cells arose from this material in its new position the conclusion would have been convincing that these substances were necessary for the formation of the reproductive cells. Unfortunately, although the cortical layer at the posterior end was shifted by the centrifugal force, it was impossible to locate accurately the germ cells in the embryos that developed from the eggs that were thus operated upon. (c) Relation between Cleavage Nuclei and Egg Organi- zation.—During the course of my early studies of chry- somelid eggs it occurred to me that the nuclei that result from early cleavage might be definite in number and in 16 Lillie, F. R., 1906, Journ. Exp. Zool., Vol. 3. 17 Conklin, 1917, Journ. Exp. Zool., Vol. 22. 652 THE AMERICAN NATURALIST [Von. LI distribution and that they might be qualitatively differ- ent. If this were true, then nuclei of one sort might always migrate into one part of the egg and might deter- mine the nature of the tissue that developed there, and nuclei of other sorts might likewise become located in other predetermined parts of the egg. Careful studies of the origin and migration of the cleavage nuclei'® have led to the conclusion that the distribution of these nuclei is adventitious and that they are all potentially alike— ` that is are totipotent—a view that is now held by most embryologists regarding the relative importance of nucleus and cytoplasm during cleavage. That the nuclei may play a part in the differentiation of the cortical layer of cytoplasm during the cleavage period is highly im- probable, since definite cytoplasmic organization already exists before cleavage begins. The factors brought into the egg by the spermatozoon, however, have an oppor- tunity at this time to modify the initial organization and thus the early embryo may exhibit paternal characteris- tics. Whether or not such an influence is exerted at this time is not known. The kind of tissue that develops from any part of the egg, therefore, depends upon the kind of cytoplasm en- countered by the nuclei. (d) Complexity of Organization as Indicated by the Development of Parts of Eggs.—More convincing evi- dence of the presence of a complex and fixed organiza- tion in the cytoplasm was derived from operations per- formed upon eggs with a hot needle. Parts of the freshly deposited eggs were killed by being touched with a hot needle und these parts were thus prevented from taking part in development. The living portions of the eggs continued to develop and in every case produced those parts of the embryo that they would have formed if the egg had not been injured.!® This seems to prove that every part of the egg cytoplasm is set aside for the pro- ` 18 Hegner, 1914, Journ. Morph., Vol. 25. O Fegin, 1911, Biol. Bull., Vol. 20. No. 611] GENESIS OF ORGANIZATION OF INSECT EGG 653 duction of a definite part of the embryo, and hence of the larva and adult, and that the cytoplasm is therefore highly organized at the time the egg is fertilized. After such experiments there is no regeneration of substances. As stated above, the cortical layer of cytoplasm is vis- ibly alike throughout except at the posterior end, where it has embedded in it the pole-disc granules. One of the most interesting results of the operations performed with the hot needle was obtained by killing the posterior por- tion of the egg containing the pole-dise (Fig. 7). Eggs B Fic. 7. Diagrams showing the results of killing parts of the eggs of Lepti- Teea oman ta with a hot needle. (Hegner, 1911. ongitudinal section through an egg twenty-four hours old. The pos- sia end _(k) was killed just after the egg was deposited (conditions as in Fig. . o germ cells were Oe uaa tral view of an embryo three igi gs The posterior end (k) was uta ae oe the egg was S denaii toi. N p ya: remained aliye gave rise to the head va pean half of the thora ew of a Sag five days ah: The aniditoe end (k) was killed in the hinstoderm stage ir . 8, d). The part that remained alive produced the te nUe f alte nd Cae tak of the thorax (t). o Ven ral view of an egg three days old. Thé posterior end (k) was ket when the embr Faih was two days old, The anterior Se continued to de- pendence of the tissues is indicated by the minute end of the Meee i pener riit normally after being a from the rest of the embryo. The two parts of the egg underwent a revolution of ninety degrees steep the twenty-four hours succeeding a ment. = abdomen. bl= blastoderm. gcd = pole-disc. h = head. k= portion of egg ini with hot needle. t= thorax. tf — tall fold. y= yolk. : thus modified produced embryos without germ selli, prov- ing that this cytoplasmic region is necessary for their formation. The castration of the individual may also be performed in a similar fashion after the germ cells have been extruded from the egg (Fig. 3, d), and it is interest- ing to note that sexless chicks have recently been pro- 654 THE AMERICAN NATURALIST [Vou. LI duced in a similar fashion, by removing the region of the embryo from which the germ cells arise.?° The blastoderm is of course definitely organized, since its cells contain organized cytoplasm, and by killing parts of the eggs in the blastoderm and later stages results were obtained similar to those produced when fresh eggs were operated upon (Fig. 7, b, c, d). 7. Summary of Part I Summarizing the data briefly given above, we may say: 1. Morphological and experimental studies have proved that the eggs of animals are more or less highly organ- ized at the time of fertilization. 2. We know almost nothing about the nature and gen- esis of this organization. 3. Descriptions are given of the condition of the eggs of certain chrysomelid beetles at the time of deposition, of.the stages of cleavage and blastoderm formation, of the origin of the germ cells, and of the principal stages in the germ-cell cycle. 4. The eggs of certain chrysomelid beetles and of other insects are definitely organized when deposited as indi- cated by observations on normally developing eggs and . by experiments with gravity. 5. This organization exists in the cytoplasm as indi- cated by a morphological study of cleavage, by experi- ments with gravity and centrifugal force, and by killing with a hot needle parts of eggs in various stages of de- velopment. 6. These observations and experiments prove also that the nuclei up to the time of blastoderm formation are totipotent. II. THE Genesis or THE ORGANIZATION oF THE INsEcT Hee 1. Introduction. I have decided to consider the organization of the egg only in this discussion, since it contains everything neces- : 20 Reagan, 1916, Abstracts 14 annual meeting Amer. Soe. Zool. No. 611] GENESIS OF ORGANIZATION OF INSECT EGG 655 sary for the production of a complete organism. In many species of insects and other animals, parthenogenesis is a normal phenomenon and in many species whose eggs must ordinarily be fertilized development may be ini- tiated by artificial means.?™! Among these are the eggs of the silkworm moth. It also seems certain that the cytoplasmic regions of the insect egg have reached a high state of morphological and physiological differentiation before fertilization, judging from the results outlined in Part I. of this paper. The cleavage nuclei may possibly exert an influence upon the cortical layer of cytoplasm before the blastoderm is formed, thereby enabling the paternal contribution to the zygote to act, but besides be- ing very improbable, such a phenomenon would, of course, follow rather than precede the establishment of the ad- vanced state of organization that exists in the undevel- oped egg. | 2. Constitution of the Primordial Germ Cells In certain beetles, flies and parasitic Hymenoptera, the primordial germ cells are visibly different from the rest of the embryonic cells that arise at about the same time. This difference is primarily due to the inclusion within them of visible substances that are located in the egg ma- terial from which they originate. The eggs of these dif- ferent insects are similar in certain respects and dif- ferent in others. In every instance, however, this visible substance, which forms the germ-line determinants, is situated near the posterior end of the egg, and it is at this point that the primordial germ cells are formed. The origin of the germ-line determinants is not definitely known in any insect, but their position in the egg and their granular appearance are constant. It has often been pointed out that the primordial germ cells remain in a comparatively undifferentiated state until the individual in which they lie has almost reached maturity, and that they then undergo changes during 21 Loeb, 1913, ‘‘ Artificial Parthenogenesis and Fertilization,’’ Chicago. 6560- THE AMERICAN NATURALIST [Vou. LI which they reach a high state of specialization. The dis- covery of axial gradients of metabolism in the eggs of certain animals in an anterior-posterior direction?? sug- gests that this may also be true of insect eggs. If such gradients exist in insect eggs and if the metabolic activity decreases gradually from the anterior to the posterior end, then the primordial germ cells, which arise at the extreme posterior end, are actually the least active meta- bolically of all the cells of the embryo. Their early sep- aration from the egg would also tend to keep them in an undifferentiated condition since they are on this account less likely to be influenced by the rest of the embryo. The primordial germ cells in these insect eggs are thus visibly different because of the presence of germ-line de- terminants and are probably physiologically different, at least in part, because of their position at the posterior end of the egg. The contents of these cells are as follows (Fig. 4): (1) part of the cortical layer of cytoplasm, (2) part of the cytoplasm which surrounds the cleavage nuclei and which is collected from among the yolk globules, (3) part of the germ-line determinants, and (4) a nucleus with the full amount of chromatin. The fourth item is mentioned because in Miastor all of the nuclei that form somatic cells undergo a diminution process, being similar in this respect to Ascaris. This chromatin is in Miastor entirely maternal since the eggs of this fly that have been studied, develop parthenogenetically. Nothing very definite has been discovered regarding the arrangement of these substances in the germ. cells. The nucleus lies near the center in all of them; the two kinds of cytoplasm soon become indistinguishable; and the germ-line determinants may, at first, be more or less evenly distributed throughout the cytoplasm, as in chry- somelid beetles and Miastor, or may be clumped in vari- ous parts of the cell, as in Chironomus. In every case, however, the germ-line determinants evidently become 22 Child, 1916, Biol. Bull, Vol. 30. No. 611] GENESIS OF ORGANIZATION OF INSECT EGG 657 more or less evenly scattered since they cannot be distin- guished in later stages in the germ-cell cycle. 3. Differential Divisions during the Formation of Nurse Cells There is no evidence of any definite localization of sub- stances or physiological processes in the primordial germ cells when formed, nor do these cells exhibit recognizable polarity or symmetry of any kind. As described in preceding pages, they multiply; migrate into or are enveloped by the tis- sues of the embryo; sepa- rate into two groups from which the ovaries on either side of the body arise; and then pass through an- other period of multipli- cation. This brings them to the stage just preceding the growth period. At this time phenomena occur m. in the ovaries of certain 4 youme De ae A a a species of insects that ¢ompanying nurse cells (nc). gv= germ- have a direct bearing upon inal vesicle. (Hegner, 1914. our problem; these are concerned principally with the dif- ferentiation of oocytes and nurse cells. In Miastor the nurse cells are mesodermal in origin, and a group of nurse cells and one oocyte become enclosed within a sheath of epithelial cells (Fig. 8). As the oocyte increases in size it elongates, and then for the first time in its history ex- hibits recognizable polarity; the anterior end adjoining the group of nurse cells. Polarity may, however, have been present from the time the primordial germ cell was first formed, corresponding to that of the parental egg. The germinal vesicle soon becomes eccentric, but whether or not this indicates that bilateral symmetry has also been 658 THE AMERICAN NATURALIST [Vou. LI determined, as it does in certain other insects, is unknown. It is thus certain that polarity exists soon after the be- ginning of the growth period and that bilaterality is prob- ably also established at an early stage. The differentiation of oocytes and nurse cells in dy- tiscid and gyrinid beetles is of peculiar interest, although the early and later history of the germ cells in these in- sects is not known. In the diving beetle, Dytiscus mar- ginalis,?> a single oogonium gives rise to fifteen nurse cells and one oocyte. The oocyte and its mother cell, grandmother cell, and great-grandmother cell can be dis- ere ox Diagrams Boeman differential divisions ane the formation of nurse sees in the errand _— igi? nigrior. Hegner and Russell, 1916.) of an ovarian tubule showing two oocytes (0) each accompanied d by seven nurse sol a j: . Division of two-cell stage. An ultimate oogonium has divided forming a iah cell (ne) and an oocyte grandmother cell (ogmc) containing the oocyte determinant. 7 intercellul d. grammatic ae of oocyte differentiation. The plain circles Indicate nurse cells. = ultimate oogonium re the tie ds determinant within its nucleus. Pome oocyte grandmother c o= 23 Giardina, 1901, Internat. Monatssch. f. Anat. u. Hai Bd. 18; Debai- sieux, 1909, La Cellule, T. 25; Gunthert, 1910, Zool. Jahrb., Bd. 30. No. 611] GENESIS OF ORGANIZATION OF INSECT EGG 659 tinguished from the nurse cells by the presence of a pe- culiar ring of nuclear material within the cytoplasm and by their larger size. The gyrinid beetle, Dineutes nigrior (Fig. 9), resembles Dytiscus in general, but the ultimate oogonium passes through one less division, thus giving rise to one oocyte and only seven nurse cells.24 The im- portant fact is that during these differential divisions, in both cases, the nurse cells, which may be considered so- matic since they are unable to reproduce, are deprived of part of their nuclear material. Apparently they differ from their sister cell, the oocyte, in this one respect, and it is therefore the presence of this nuclear material that makes it possible for the oocyte to develop into a new in- dividual. This is one of the most striking cases of the passage of large masses of nuclear material into the cyto- plasm. No such differential divisions have been discov- ered in chrysomelid beetles nor in the other insects where the nurse cells arise from oogonia, but they may occur in some way that has not been revealed by our methods of research. The writer has discussed this subject rather fully with relation to the origin of nurse cells and oocytes in the honeybee.” In this insect a single oogonium gives rise to a rosette-like group of cells that are connected with one another by strands—probably of a mitochondrial na- ture—the remains of preceding mitotic divisions. There is no visible difference among the cells in a rosette which are hence apparently potentially alike. Nevertheless one or several from each rosette enlarge to form oocytes which are nourished by the rest acting as nurse cells. “What determines the differentiation of certain cells into oocytes is not known but the following hypotheses have been expressed. Three explanations have occurred to me: (1) There may be dif- ferential changes during the mitotie divisions in rosette formation as in Dytiscus resulting in one or more cells (oocytes) which differ in con- 24 Hegner and Russell, 1916, Proc. Nat. Acad. Sc., Vol. 2. 25 Hegner, 1915, Journ. Morph., Vol. 26. 660 THE AMERICAN NATURALIST [Vou. LI tent from the others (nurse cells). No visible changes of this sort were observed. (2) The polarity of the rosettes may influence the cells in such a way that those near the center of the ovariole and closest to the zone of differentiation tend to develop into oocytes. (3) Those cells of the rosettes which reach the zone of differentiation first are stimulated to become oocytes and by their growth and differentiation prevent the other cells of the rosettes from, similar changes. 4. Constitution of the Oocyte at the Beginning of the Growth Period Very soon after the nurse cells are formed and the oocytes begin to enlarge the main axis of the oocyte in all AESA ‘ Bi Fic. 10. The formation of oocytes in the honeybee. (Hegner, 1915.) a. Part of an ovariole showing the rosettes (r) each resulting from the division of a single oogonium. b. Part of an oyariole in the zone of differentiation showing five oocytes (0), many nurse cells: (n) and epithelial cells (e), insects seems to be established. The germinal vesicle at the same time changes its position from the center of the cell to a point near the nurse-cell chamber at the an- terior end as described above in Miastor (Fig. 8). At what stage bilateral symmetry becomes fixed has not been determined. The oocytes of insects at the beginning of the growth No. 611] GENESIS OF ORGANIZATION OF INSECT EGG 661 period differ from the other cells in the body in the fol- lowing ways. (1) In all cases where germ-line determi- nants occur the oocytes alone are provided with them and with the cytoplasm in which they are embedded. (2) In insects like Mzastor a full amount of chromatin is present only in the oocytes. (3) In Dytiscus, Dineutes and prob- ably other insects the oocytes contain nuclear material of which the nurse cells are deprived, but this may be in- terpreted simply as a means of inhibiting the reproduc- tion of the latter and of changing them into nurse cells. (4) The oocytes seem to have no influence upon the de- velopment of the individual in which they lie, as indicated by castration and transplantation experiments,” and are in a comparatively undifferentiated condition when the growth period begins. 26 Meisenheimer, 1912, Fest. 60 Geburtstage von Dr. J. W. Spengel III.; Kopec, 1911, Arch. Daie. -mech., Bd. 33. (To be continued) INHERITANCE OF FERTILITY IN SOUTHDOWN . . SHEEP? EDWARD N. WENTWORTH PROFESSOR OF ANIMAL BREEDING AND J. B. SWEET ASSISTANT IN EXPERIMENTAL BREEDING, UNIVERSITY OF WISCONSIN INTRODUCTION Youatt (14)? says: The disposition to twinning is undoubtedly hereditary: “Ewes yearly by lambing rich masters do make: e lambs of such twinners for breeders go take.” Flockmasters of the last century have made selections on this assumption, while the increased number at a birth in the progeny of ewes born multiparously as com- pared to the progeny of those born singly has been dem- onstrated by several investigators. In general there has been shown to be an increase in number produced at a birth as the average birth values of the animals lambing increase. Thus, Rietz and Roberts (13) present the fol- lowing in Shropshires, the number at a birth being repre- sented by the figures 1, 2, or 3: Sire | Dam Offspring No. Cases Ta e 1 1.3452 0.0059 3,059 PT 1 1.3946 0.0073 2,088 | ere ba 1.4171 + 0.0067 2,436 r 2 1.4548+ 0.0088 1,550 One parent a triplet 1.6076 + 0.030 158 Experimental investigations of the inheritance of twin- ning in sheep have been attempted in few cases. Ains- -~ 1Paper No. 5 from the Laboratory of Animal Technology, Kansas Ex- periment Station. 2 Reference is made by number to literature cited at close of paper. 662 No. 611] FERTILITY IN SOUTHDOWN SHEEP 663 worth-Davis and Turner (1) reported a preliminary in- vestigation on this subject, but their numbers are too small and results too contradictory, as published, even to be indicative of the method of inheritance. Arkell and Jones (7) at the New Hampshire station also instituted investigations along this line, but have published no re- sults. Due to the environmental and physiological factors in- volved in multiple births, as well as to the economic im- practicability of maintaining large flocks under rigid ex- perimental conditions, there are at hand no considerable masses of experimental data which yield evidence on this point, nor are the probabilities great that such exper- iments will ever be conducted on an adequate scale; hence the bulk of evidence on the inheritance of fertility must come from breeders’ flocks or from breed registry records. Tue FERTILITY PROBLEM High fertility obviously depends on three factors—the number at a birth, the frequency of reproduction, and the total number of successful gestations an animal may undergo. Unfortunately flock book records give available data on the first point only, although for specific cases some evidence on the second point (barring abortions and unregistered progeny) exists. For breeding purposes the number of successful gesta- tions is not a practical selective index, since the breeder ean not afford to withhold progeny from breeding until their dams or sires shall have completed their breeding eycles. Frequency of reproduction or regularity of breed- ing as termed by the breeder is a more practicable trait for purposes of selection, but since barren reproductive periods are so much more frequently due to pathological or physiological causes than to genetic, most sheepmen lay principal emphasis on the number of offspring at the given birth. There are two ways in which selection on this basis may be applied. The ewe may be selected on the basis of 664 THE AMERICAN NATURALIST [ Von. LI a particular lambing, or the basis of the best lambing she shows. From a genetic standpoint the second criterion would seem the better, but practical breeders would be very likely to use the first. Unfortunately, records in Southdowns on which a comparison can be based are few, forty-three animals only being available. Table I pre- sents the correlation of each individual record with the average lambing record for each ewe, while Table II pre- sents the correlation of the best record for each ewe with her average. Nine of the ewes had four lambings to their eredit, nine had three, while twenty-five had only two. The inadequacy of these data is recognized, since there is a false agreement between a single number and its average with another as compared to its agreement with its average with several numbers. Since, however, the material is suggestive from a comparative standpoint, it is presented, as the same actual error exists in each table: ; TABLE I CORRELATION OF INDIVIDUAL LAMBING WITH AVERAGE LAMBING PER EWE Average Lambing per Ewe Individual 1 1.5 1.67 1.75 2 2.5 F 1 38 6 4 4 1 Oo. | 8 2 0 6 Cede CES 1 58 3 0 0 0 0 1 1 2 Total $% | 12 kelep feck as 2 fens: The coefficient of correlation for this table is 0.81806 + 0.02099. TABLE II CORRELATION OF BEST LAMBING RECORD WITH AVERAGE OF EACH EWE Average Lambing of Ewe High Record 1.00 1.50. | 167 1.75 | 2.00 2.50 | F 1 16 0 0 0 0 ö ow 2 0 4 4 4 13 0 | 25 3 0 Q 0 0 1 1 | 2 ead 16 4 4 t] H Ii u The coefficient of correlation for Table IL was found to be 0.92354 + 0.01513. No. 611] FERTILITY IN SOUTHDOWN SHEEP 665 ‘While both records show a higher agreement with the average than probably exists in actual selections, the fact that the best record is more closely in agreement with the average than a random record makes high production a significant selection standard. The correlation between random records and the best records is presented in Table IIT TABLE IMI Best RECORD OF EWE Individual Record 1 | 2 3 F 1 38 14 1 53 2 0 55 3 58 3.. 0 0 2 2 Total 38 | 69 6 113 The coefficient of correlation here is 0.6518 + 0.03665. The relationship is not as great as between either of the records and the average record, as shown in Tables I and II. Since the correlation is not as high, and since an error (false agreement with the average) is introduced into each of the first two tables, it is well to determine whether the difference between the first two correlation coefficients is significant. Using the formula for the standard deviation of the difference between the first two constants presented by Pearl (10) : Error of (x — y)=V Ea? + E,? — 2 rey oe oy, where E refers to the error, x to the larger constant, y to the smaller, and rey to the correlation between æ and y. The error of the difference between the cor- relation coefficients of Tables I and II is 0.01598. Since the difference is .10548, it is greater than three times the probable error, hence it is justifiable to conclude that the highest number at a birth is a better indication of the average fertility of an animal than a random birth, although on the basis of the figures presented the latter relationship is high. Youatt (14) reports in 1837 that one ewe out of five in the average English flock produced twins, which would give 120 per cent. of lambs as the Seas of English flocks at that time. 666 THE AMERICAN NATURALIST [ Vou. LI Mansell (12) reports 168 per cent. of lambs in 11,668 English Shropshires in 1896, while Humphrey and Klein- heinz (6) from figures on the University of Wisconsin flock made the following breed comparisons: TABLE IV Singles | Twins | Triplets Breed im i AR il E E A Lor eet cng a No. Per Cent. | No. Per Cent. | No. Per Cent. Shropshire 42 Ve ESS eae E A 67.8 | 15 9.5 Dorset 10 33.3 | 20 66.7 | 0 0.0 * Southdown 27 PETOA IT 62.0 | 0 0.0 xfor 3 136 16 Tae i 3 13.6 tampshire 9 SLO | 20 69.0 | 0 0.0 hevi 9 31.0 20 69.0 0 0.0 s Percentage of faha as given by Mansell is, of course, only a rough indication of twin-bearers, since ewes having triplet and quadruplet births may be aneluded: Rietz and Roberts (13) show that 43 out of every 100 births in American Shropshires are multiple births, while they have determined from Heape’s (5) statistics of 1895-96 that 64 out of every 100 births in English Shrop- shires are multiple. Plumb (12) found in 20,037 Shropshire births 59.2 per cent. were singles, 39.2 per cent. twins, and 1.3 per cent. triplets, all recorded in the American Shropshire Flock Book, 1890 to 1899. Heape (5), from a study of the birth records of 89,000 ewes in English flocks, presents the following data to show the relative fertility of different breeds of sheep: TABLE V Per Cent. Lambs Per Cent. Twin Bearing per Ewe BUNGE oasis sss ceo te 141.77 52.22 BG PEE E Pn pas ag ol 124.05 31.38 Bouthdownh s.ro eevee 109.89 18.67 Hampshire s. css. tukso 114.69 24.09 Oxforda Down.. s ceros 119.16 35.02 Dorit Bom oa es i ee 123.63 37.55 PhropebyrG wie Ts es 136.79 46.84 Smee Oe eri es 111.10 29.09 No. 611] FERTILITY IN SOUTHDOWN SHEEP 667 The figures for the per cent. of lambs per ewe and the per cent. of twin-bearing ewes do not in all cases check each other, as records of certain ewes were available for the one column but not for the other. The writers tabulated birth frequencies in Shropshires, Cotswolds and Dorsets with the following results: TABLE VI Singles | Twins | Triplets SSRN i Breed r | | Í | No. | Per Cent.) No. | Per Cent. | No.| Per Cent. No.| Per Cent. Shropshire............ 10,585 69.41 | 4,561; 29.91 102 .67 |2| .01 2,148 67.75 956 | 30.22 57 1.88 y 22 Cotswóld 5,528} 79.24 |1,431 | 20.53 | 16; .23 Dorsets seem to have an exceptionally high percentage of triplets and quadruplets. FACTORS AFFECTING FERTILITY Heape (5) in a study of 122,673 breeding ewes, 413 Eng- lish flocks, suggests five physiological factors that may affect the hereditary expression of fertility. The most important factor according to him is the physical condi- tion of the ewe, which must be vigorous and healthy, espe- cially at tupping (mating) time. The second most important factor is the feeding of the ewe, especially flushing previous to breeding, and careful diet during gestation. The third factor in importance is the district; he cited the fact that the Suffolk in its native country pro- duced 60.46 per cent. of twins, while in Essex it produced only 42.87 per cent. The fourth factor in importance he found to be the age of the ewe; and the fifth, the season of year at which mating occurred. Carlyle and McConnel (3) at Wisconsin discuss time of mating and age of ewe, factors similar to those mentioned by Heape. In a study of twelve years of records of the station flock at the University of Wisconsin they found that ewes bred early in the season dropped a higher per- centage of lambs than those dropped late in the season, 668 THE AMERICAN NATURALIST [Vou. LI while ewes from three to six years of age seemed to be at the optimum breeding period of their life. Humphrey and Kleinheinz (6) found that two-year-old ewes pro- duced 141 per cent. of lambs and six-year-olds, 191 per cent. Possibly the writers do not understand the tables presented by them, but their calculations on the basis of the data there given would show the following averages at each age: TABLE VII Age Average No. per Birth No. Cases Corl seas ae wees 1.54 53 Bs a a es AT 1.56 52 4g. eee EOG Ee pO a | 52 TAE E E E 1.73 57 M E eae pe a a 1.92 24 EON aR ETRE R R 1.45 11 Bae CS tiie ay any oe a nes 1.00 2 BSS Pes ess aa 2,00 2 Pearl (9) made a biometric study of the fertility of a long-lived ewe whose breeding record was as follows: TABLE VIII Lambs Lambs April, 1806. oara oe aes a 1 ASLO cs ice town) SF ciee $ ree BOOT see i ss ae l CRT ea a O se eas 2 SOUS i cei ss 2 od 0. Ue Os grips EE pe 2 Aprl 3, 1800 usoen Ae 3 1819 i.o a a Gee 2 Mar, 20, BIO anla 3 IBAN i cise ahi. saree hs 2 CORT Bey yes ea ns 6 3 ETA ee eee hee 1 Jee eters Gs oa 4 3 DORE csinosan e + 1 TOAD ce PU Le ete es 3 SORE CU eR eka fhe ee 0 WIA Beery es aa 3 ROLE neso > hie a A 0 TOIR ote. eee 2 Witel oan 36 Assuming that the ewe was about one year old when the first lamb recorded was born, Pearl found that the mean point of the ewe’s effective breeding life was 8.57 years, that the median point was 8.17 years and that the modal breeding point, or the point of maximum fertility per unit of time, was at 7.34 years. Taking into account the seventeen years in which some young were born, the following constants regarding the number of lambs per birth were found: No. 611] FERTILITY IN SOUTHDOWN SHEEP 669 Mean number of lambs per birth, 2.12 lambs. f Standard deviation in number of lambs per birth, 0.76. Coefficient of variability in number of lambs per birth, 35.78 per cent. Marshall (8) found after a study of the lambing sta- tistics for various flocks of Scottish sheep for the years 1905, 1906, and 1907, that the percentage of lambs born was, as a general rule, highest among sheep which had been subjected to a process of artificial stimulation by means of special diet at the approach of the breeding sea- son. In some cases the number of lambs per ewes in the ‘*flushed’’ flocks was nearly 200 percent. Flocks which were run upon special pasture upon the approach of the tupping season generally produced a slightly larger per- centage of lambs than those receiving no sort of special feeding. Evvard (4) found among range ewes fed the same ration that the fourteen heaviest gaining ewes at time of breeding in his flock averaged 1.8 lambs; the four- teen medium gainers, 1.59; and the fourteen lightest gainers, 1.44. RELATION OF MAMMÆ To FERTILITY Alexander Graham Bell (2) conducted an experimental investigation on the relation of the number of mamme to fertility. An unusually high fertility among a flock of native sheep in Beinn Breagh, Nova Scotia, led Bell to examine the ewes in order to discover some distinguishing mark of the twin-bearing ewe. He found a certain num- ber of ewes with one to two supernumerary nipples in an embryonic, functionless condition. Of these abnormally nippled ewes, 43 per cent. had twin lambs, while of the normally nippled ewes but 24 per cent. produced twins. This apparent correlation between multinipples and in- creased fertility led to an extended series of experiments to ascertain whether by selective breeding, the super- numerary nipples could be made functional, and whether ewes with additional mammæ in a functional condition were more fertile than ewes with the normal number of nipples. ; 670 THE AMERICAN NATURALIST [ Von. LI No ‘difficulty was experienced in obtaining ewes that produced milk from six nipples. These multi-nippled sheep, however, did not prove to be more fertile than normally nippled sheep. In his 1912 paper (2) he states that the indications are that the six-nippled stock will ultimately prove to be twin bearers, as a rule, at ma- turity. METHOD OF OBTAINING DATA The source of the data in the present study was the American Southdown Record, the first twelve volumes being used to obtain cases of triplets, and volumes nine to twelve for twins and singles. The pedigree of each animal was reported into the third generation, recording the numbers of offspring at the birth of each animal. Some records on triplets were also taken from the Amer- . ican Shropshire Record, while Volume 25 was used to determine the ratio of singles, twins and triplets, Vol- umes 9-12 of the Southdown Flock Book for the same purpose, Volumes 12, 13, and 14 of the Continental Dorset Club Record, and Volumes 11 and 12 of the American Cotswold Record. ` RELIABILITY or Frock Book Dara Records of the number at a birth in sheep are probably highly accurate for such material, since there is no ob- servable tendency to discriminate in favor of, or against, recording offspring of multiple births, except the indirect one of lesser development in offspring from multiple births. This would not affect the reliability of the figures presented by the flockmaster, except perhaps to reduce slightly the proportion of multiple births registered. It may be safely assumed that the bulk‘of the records are accurate, barring clerical error. THE NUMBER AT A BIRTH as A Genetic INDEX Due to the physiological causes limiting the full expres- sion of the genetic fertility of an animal it is obvious that animals recorded as singles may be potentially twin or No. 611] FERTILITY IN SOUTHDOWN SHEEP 671 triplet bearers, or that ewes recorded as bearing twins may be genetically triplet producers or better. Hence it may be expected that not all single or twin bearers are alike in zygotic make-up with reference to fertility or that their breeding performance will fall into sufficiently well- defined categories to permit a rigorous Mendelian group- ing. The relation between a random lambing record and the average record of ewes was shown earlier in this paper to be high, hence a similar relation might be expected to hold for true genetic fertility, were it measurable, and a random record. Tar Data INVOLVED The Relative Influence of Sire and Dam.—Rietz and Roberts (12) found a mathematically significant effect of the sire on the number at birth as adjudged by the corre- lation between offspring and sires, although they do not find a similar relation between dams and maternal grand- sires. While the authors have not secured correlation co- efficients on this point, their averages may be so arranged as to throw some light on the same point. Using pedi- grees which were started from animals of single birth, the following comparison between sires and dams is pos- sible. TABLE IX RELATIVE INFLUENCE OF SIRE AND DAM ON BIRTH NUMBER, FROM PEDIGREES or ANIMALS OF SINGLE BIRTH No. Cases | Sire | Dam | Ave. No. Progeny | No. Cases | Sire | Dam | Ave. No. Progeny 1,872 1 1 -29 1,872 1 1 +29 925 1 2 .28 570 2 1 .25 14 1 3 .43 12 3 1 .50 5,570 2 1 .25 925 1 2 .28 2 2 .34 2 2 1.34 10 2 3 .20 6 3 2 Az 12 3 1 1.15 14 1 3 1.43 6 3 2 fet yi 10 2 3 1.20 Comparison of the records of single, twin and triplet sires mated to single, twin and triplet ewes in the pre- ceding table shows no particular influence of the birth rank of the sire, a fact which is confirmed in Table X, where the average performance of each is given. 672 THE AMERICAN NATURALIST [ Vou. LI TABLE X BREEDING PERFORMANCE OF THE MALES FROM PEDIGREES STARTED WITH SINGLE BIRTHS . Sires | Mean | Standard Deviation No. Casts Pdi pcan ea: | 1.2864 .00593 4668 2,811 e aa ilies 1.2776 .01031 4553 886 CET ORR | 1.3888 .07750 4875 18 The difference between the breeding performance of the singles and twins is 0.0088-+ 0.0119, which is, of course, not sufficient to be significant. It indicates either that the male has no influence on the number at a birth (the most probable supposition) or that singles and twins in the males are genetically similar. The difference be- tween the breeding performance of the triplets and singles is 0.1024 + .0777 and between the triplets and twins is 0.1112 + .07818, neither of which is significant. For the ewes the result is not particularly different. Table XI presents the result of this comparison. TABLE XI BREEDING Poroa OF THE FEMALES FROM PEDIGREES STARTED WITH SINGLE BIRTHS Dams | Mean | Standard Deviation | No. Cases a ene 1.26365 = 0.00631 .4635 2,454 A clepetsc - 1,29345= 0.00893 ' .4658 1,23 Bornin 1.33333 + 0.06490 A714 24 The difference between the progeny of ewes born singly and those born twins is .01349 + 0.1093; between singles and triplets is .05338 + 0.06252; and hetwont twins and triplets is .03989 + 0.0655. TABLE XII RELATIVE INFLUENCE OF SIRE AND DAM ON BIRTH NUMBER, FROM PEDIGREES or ANIMALS OF TWIN BIRTH No. Cases Sire | “Dam | Ave..No. Progeny | No. Cases | sire | Dam | Ave. No. Progeny 9906) ee (4 1.51 2805 | 1 | 1 1.51 124 | i 2 1.55 ts |. 3 BT 21 1 3 1.86 19 3 1 1.68 687 |} 2%) 4 1.57 kieo oid 1.55 roe | 2 1.56 468. | 2 | 2 56 1 | 2 | 8 1.60 tA eps 1.43 G i 1.68 eves ORGS far 1.86 rts | Ss 1.43 “wie 3 1.60 No. 611] FERTILITY IN SOUTHDOWN SHEEP 673 Table XII shows the relative breeding performance of the sires and dams in pedigrees started from twin births. Treating the sires in pedigrees from twin births as in Table X, Table XIII is produced. TABLE XIII BREEDING PERFORMANCE OF THE MALES PROM PEDIGREES STARTED WITH Twin BIRTHS Sires Mean Standard Deviation No. Cases Wii RT EA | 1.5296 = .00543 .51659 4,120 oeei eai a | 1.5682 = .00724 .49704 A 165 e E NE | 1.6154 = .06435 . 48650 26 The difference between singles and twins as sires is .0386 + .0091; between singles and triplets is .0858 + .0645; and between twins and triplets is .0472 + .0647. The difference between singles and twins is in this case significant, being about 4.2 times the probable error. Further consideration will be given this difference when the ewes are discussed. ~ Treating the ewes in pedigrees from twin births as in Table XI, Table XIV is produced. TABLE XIV BREEDING PERFORMANCE OF FEMALES FROM PEDIGREES STARTED WITH TWIN BIRTHS Dams Mean | Standard Deviation | No. Cases e EP E -1.2529-+.00581 0.51075 3,511 r EERE BON Sage Ok 1.5551 =.00827 0.51583 1,769 Bsa aes 1.7742+.05065 0.41811 31 The difference between singles and triplets is 0.2513 + .05098; between twins and triplets, 0.2191 + .05132; and Faea singles and twins is 0.0322 + .01011. Ewes from triplet births give significantly larger progenies than ewes from single or twin births, while ewes from ‘twin births give significantly larger progenies than ewes from single births, the last difference being 3.323 times . the probable error. It is interesting to observe that both 674 THE AMERICAN NATURALIST [Von. LI twin rams and twin ewes are significantly better breeders than singles. Just why this result is obtained here in the face of other contradictory data is difficult to understand. In order to combine the results of the two types of pedi- grees it was deemed advisable to utilize the ratio of 1:4.118 twins to singles discovered by examination of volumes 9 to 12, respectively, in order to have the normal relationship between twins and singles. This involved dividing the numbers of individuals in the twin group or multiplying those in the single group. In the first case errors would be increased, due to the elimination of cer- tain groups, while in the second case errors would be in- creased due to the exaggeration of differences between the random sample in the pedigrees begun from single births and the normal distribution of such a population. It was deemed best to use the second method, since it permitted the retention of the small groups, hence the ratio 1:4.118 was multiplied by the ratio 3,715:5,311, the numbers of individuals in the pedigrees from twin and single births, respectively, which gave the multiplying factor 5.887 for the pedigrees started from single births. Of course, this result is only suggestive; but it was impractical to record the additional 4,300 odd pedigrees necessary to get a true random distribution. Treated this way, multiplying Table V by 5.887 and adding to Table IX, Table XV is produced. TABLE XV No. Cases Sire Dam | Ave. No. Progeny| No. Cases Sire Dam lave. No. Progeny 18,82 | 1 1 1.33 13,826 | 1 1 133- 6,739 1 2 1.33 4,043 | 2 1 1.33 1,103 1 3 1.51 3 1 1.54 4,043 2 1 1.30 6,739 1 2 1.33 2,270 2 2 1.38 2,270 2 2 1.38 69 2 3 1.26 3 2 1.21 89 3 1 54 1,103 1 3 1.51 43 3 2 1,21 69 2 3 1.26 _ Treating the sires as in Tables VI and X, Table XVI is produced. No.611] FERTILITY IN SOUTHDOWN SHEEP 675 TABLE XVI BREEDING PERFORMANCE OF THE MALES GIVEN IN TABLE XI Sires | Mean . | Standard Deviation No. Cases y ERPE TOE | 1.3340 = .00229 .48658 | -20,668 PATTE | 1.3308 = .00403 47677 | 6,382 Dats Baste E TEN 1.3318 = .2908 -49533 | 132 The difference between singles and twins is .0032 + .0463; between singles and triplets, .0978 + .02916; and between: twins and triplets, .1010 + .02936. TABLE XVII BREEDING PERFORMANCE OF FEMALES FROM TABLE XV Dams Mean Standard Deviation No. Cases E aa e a | 1.3274 + 0.00243 : 0.48261 17,958 yas ees A | 1.3444 = 0.00345 0.48713 9,052 ere EA / 1.4128 + 0.02533 0.49234 172 The difference between singles and twins is found to be 0.0170 + 0.00422; between singles and triplets, 0.0854 + 0.02545; and between twins and triplets, 0.0684 + 0.02556. Several of the differences in Tables XV and XVI verge on significance, being at least three times the probable error. RELATIVE INFLUENCE OF MALE AND FEMALE IN GRANDPARENTS From the study of the relative influence of the sires and dams on the progeny it would seem fruitless from biometric grounds to look for transmission through one sex more than the other. Yet logically it would seem that the grandsire and grandam on the dam’s side would have a more potent effect on the birth number from the dam than would the paternal grandparents. Studies of this sort are available from the pedigrees. Perhaps the first concern is to determine the relation of the birth rank of the grandparents to that of the progeny. Table XVIIT presents this information. 676 THE AMERICAN NATURALIST [ Von. LI TABLE XVIII RELATION OF BIRTH FREQUENCIES IN GRANDPARENTS TO BIRTH FREQUENCIES or P ROGENY Grandsire | Grandam > liros Ave. Progeny Sumun s No. Cases l 1 1.6500= .00881 .5451 1740 į g 1.5 1.7041 .01488 6030 747 2 1 FO 1.7065=.01923 5717 402 2 2 2 1.7500+ .03328 5517 308 i 3 2 1.8095 = .08642 5871 21 3 J 2 2.0000= .07855 0000 6 z 3 2.5 1.6667 = .08297 .5634 21 3 2 25 1.7500 .14606 .4331 The difference between the average progeny from grandparents 1 g X 19 and grandparents 2 § X19 is .0579 + .02114. This is not three times the probable error, therefore the difference is not significant. The differ- ence between grandparents 1¢é X 19 and grandparents 1¢ 32 is .1609 + .08687. This also is less than three times the probable error, hence is insignificant. In fact none of the differences are significant. To determine whether birth rank in males or females among the maternal grandparents has effect on transmis- sion, they were compared in the same manner as the sires and dams were. The results for grandsires are: TABLE XIX RELATION OF BIRTH RANK OF GRANDSIRE TO BIRTH RANK OF PROGENY Birth Rank of Grandsire| Ave. Progeny Standard Deviation) No. Cases 1 | 1.6675=.00760 5643 2,508 2 | 1.7209 .01442 5779 731 EE T TT | 1.9000 = .06399 3000. - 10 The difference between singles and twins is .0534 + .0163; between twins and triplets is .1791 + .06559; and between singles and triplets is .2125-+ .06444. The difference between singles and triplets is 3.62 times the probable error, while the difference between singles and twins is 3.34 times its probable error. . Treating the dams in the same manner as the sires Table XX is produced: No. 611] FERTILITY IN SOUTHDOWN SHEEP 677 TABLE XX RELATION OF BIRTH RANK oF GRANDAM TO BIRTH RANK OF PROGENY Birth Rank of Grandam| Average Progeny | Standard Deviation | th Olen. 1 | 1.6615=.00737 | .5064 | 2,148 2 | 1.7177=.01239 | .5979 1,059 3 1.7381 +.06034 ‘5798 | | 42 The difference between singles and twins is .0562 + .01435; between singles and triplets i is .0766 + .06079; and obre twins and triplets is .0204 + .06159. The only significant difference is between singles and twins, which is 3.88 times the probable error. The probable errors involved seem to indicate little, hence a comparison by correlation of the maternal grand- sire and progeny, and maternal grandam and progeny was instituted. Table XXI presents the correlation for the maternal grandsire, Table XXII for the maternal grandam. TABLE XXI CORRELATION OF MATERNAL GRANDSIRE AND PROGENY Birth Rank Progeny Birth Rank Grandsire f 1 2 3 4 1 54 1,435 118 1 2,508 2 251 431 49 731 3 9 1 10 f 1,205 1,875 168 1 3,249 TABLE XXII CORRELATION OF MATERNAL GRANDAM AND PROGENY Birth Rank Progeny Birth Rank Grandsire pine d 1 2 3 1 811 1,254 82 2,148 y EEE TA 381 596 82 1,059 3 3 25 14 f. 1,195 1,875 178 3,249 The coefficient of correlation for Table XXT is .0496 + .0118, while for Table XXII it is .0382 + .0118. The difference between the correlations of maternal grandsire 678 THE AMERICAN NATURALIST [ Vou. LI and grandam is .0114 + .0167, a difference insignificant, hence one can not assume sex linkage. EXAMINATION OF SHROPSHIRE DATA FROM THE MENDELIAN STANDPOINT A number of Shropshire pedigrees were tabulated which were all started from triplet births. It had seemed from inspection that triplets might be genetically different from twins and singles, hence the pedigrees were tabulated to discover such a difference if possible. If the maternal grandparents affected the number at a birth from their daughter, then it was possible that certain differences might appear in the pedigrees indicating the genetic ef- fects. The results follow. TABLE XXIII RELATION OF BIRTH RANK IN OFFSPRING TO BirTH RANK IN DAM WHEN TH TERNAL GRANDPARENTS ARE SINGLES Offspring Dam 1 | 2 | 3 | 4 | Mean | Standard Deviation AEN 17 16 | 17 0 2.00 0.824 ee 12 10 15 1 2.13 0.894 eS Q a l 0 3.00 0.000 TABLE XXIV WHEN MATERNAL GRANDSIRE IS A SINGLE AND MATERNAL GRANDAM IS A TWIN Offspring Dam | 1 | 2 3 _ Mean Standard Deviation a ESITA ENE 15 9 15 2.00 0.873 AE TEOT 10 9 25 2.34 0.825 3.. 5 0 0 1.00 0.000 TABLE XXV MATERNAL GRANDSIRE A TWIN, MATERNAL GRANDAM A SINGLE 3 Mean Standard Deviation = aq QO oom wN 13 ; T 2.00 0.858 0 1.00 0.000 No. 611] FERTILITY IN SOUTHDOWN SHEEP TABLE XXVI MATERNAL GRANDPARENTS TWINS Dam | 1 | 2 3 | Mean | Standard Deviation by Ee 4 2 4 2.00 0.895 vae 9 9 9 2.00 0.817 desd tay 0 1 2 2.67 0.417 TABLE XXVII ; MATERNAL GRANDSIRE A TRIPLET, MATERNAL GRANDAM A SINGLE Offspring Dam 1 | 2 | 3 | Mean | Standard Deviation i E R E 0 0 0 0.00 000 Biiiecsverten 0 1 0 2.00 0.000 AES anger 0 0 0 0.00 0.000 TABLE XXVIII MATERNAL GRANDSIRE A SINGLE, MATERNAL GRANDAM A TRIPLET Offspring Dam | 1 2 | 3 | Mean | Standard Deviation Lio 0 0 | 0 0.00 0.000 ict deccs (ek 0 0 | 1 3.00 0.000 Lee mane 0 0 1 3.00 0.000 TABLE XXIX MaTERNAL GRANDSIRE A TRIPLET, MATERNAL GRANDAM A TWIN Offspring Dam | 1 | 2 3 | Mean | Standard Deviation a S 0 0 0 0.00 0.000 Dianna 1 0 0 1.00 0.000 SEE 0 0 i 0 0.00 0.000 TABLE XXX MATERNAL GRANDSIRE A TWIN, MATERNAL GRANDAM A TRIPLET Offspring Dam | 1 | 2 | 3 Mean | Standard Deviation ET tes 0 0 1 3.00 0.000 » Cee 0 1 1 2.50 0.500 Sass seeaesde 6. 0 0 0.00 0.000 680 THE AMERICAN NATURALIST [ Von. LI Since all the pedigrees were started from triplets the excess of triplets is so great as unduly to weight the ratios. Inspection of the ratios does not reveal any par- ticular difference in the progeny descended from a par- ticular pair of grandparents, whether the dam is a single, twin or triplet. Since also there seems to be no sex link- age involved it seemed desirable to combine similar matings from the standpoint of birth rank. The totals produced are presented in Table XX XI. TABLE XXXI SUMMARY OF TABLES XXIII TO XXX witH RESPECT TO BIRTH RANK OF MATERNAL GRANDPARENTS No. Offspring Birth Rank Maternal Grand- 1 2 3 4 Mean | Standard Deviation parents Both i ag single... 29 26 33 1 2.07 0.85851 One grand . seal AG 26 60 0 2.11 0.88983 Both grandpa coh twins...| 13 12 15 0 2.05 0.83516 One grandparent a pa 1 2 4 0 2.43 0.70855 Confirmation of the previous view that twins and singles are genetically alike, while triplets differ from either, seems to be found in Table XXXI. However, the differ- ence between triplets and the mating where one grand- parent is a twin is only 0.32+0.20. This is not three times the probable error, but by consulting Pearl and Miner’s (11) table it is found that the chances that the difference is significant are about two and a half to one. CONCLUSIONS- 1. In general sheep of a high birth rank tend to produce offspring of a high birth rank. 2. On the basis of the few data presented, the high- est record of a ewe appears to be a better selection stand- ard for high fertility than a random record. 3. The frequency of multiple births in sheep varies with the breed. , 4. Physiological factors may exert a marked influence No. 611] FERTILITY IN SOUTHDOWN SHEEP 681 on heredity, the most important factors being the vigor of ewe, the feeding of ewe, the age of ewe, the season and the region. 5. Apparently no relation exists between high fertility and additional mamme. . In pedigrees started from single births, the birth rank of the sire does not affect the birth rank of the prog- eny; in pedigrees started from twin births, the effect of high birth rank of the sire is only slightly significant (more than three times the probable error). 7. The effect of birth rank of ewe on the birth rank of progeny is the same as that of the sire except in the case of pedigrees started from twin births when it is slightly greater. 8. No evidence for a sex linkage of fecundity factors occurs in the pedigrees tabulated, as shown by a com- parison of the relative influence of progeny of the mater- nal grandam and the maternal grandsire. 9. Evidence from Shropshire triplet pedigrees suggests that triplets are genetically different from twins and singles, which two are probably genetically alike. LITERATURE CITED Ainsworth-Davis, J. R., and Turner, D. 1913. Fecundity of Sheep. In V Cong. Internat. Agr. Gand., Sec. 3, question 4, p. 5 Bell, Serag Graham. 190 Seg apa Sheep of Beinn Breagh. In Science, N. 8. 9, p A 1912. Sheep Srecding Experiments. In Science, N. S., Vol. 36, pp. Carlyle, W. E: ac McConnell, T. F. 1912. Some Observations on Sheep Breeding from the Experimenta- tion Flock Records. In Wisconsin Sta. Bul., 95, p. 19. Evvard, J. M. 1913. pien Data. Heape, Walte 1899. peers Barrenness and Fertility in Sheep. In Journal of Royal tip Soc. England, 3d series, Vol. 10, pp. 234-248. Humphrey, Geo. C., and Kleinheinz, Frank. 1907. Darlia on Sheep Breeding from Records of the University Flock. In 24th Annual Report, Agr. Exp. Sta., University of Wisconsin, pp. 25-41. 682 THE AMERICAN NATURALIST [ Vou. LI Jones, J. M. 1912. Sheep Breeding and Feeding Experiments. In- New Hamp- shire Sta. Bul., 163, pp. 24-28. Marshall, F. 1910. pijaiclogy of Reproduction. London, p. 598. phe Raymond. 913. Note ah pan ve Relation of Age to Fecundity. In Science, N. Vol. 37, pp. 226-228. 1917. The ‘Probable Error of a Difference and the Selection Problem. In Genetics, Vol. 2, No. 1, p. 78 Pearl, Raymond, and Miner, John Rice. 1914. A Table for es ep Probable Significance of ree i ants. In Papers m the Biological Lab. N 226. Maine Agr. Exp. “og p. 88. abe Pep et Hypes ma Breeds of Farm Animals. Boston, p. 391 (cites data | of Mansell’s). Rietz, H. L., and Riper, E. 1915 iii a Resemblance of Parents and Offspring with Respect to Birth T for Registered Shropshire Sheep. In U. S. t. ‘als Jour. Agr. Research, TA 4, No. 6, pp. 479-510. Youatt, Wm 1837. EREN Their Breeds, Management and Diseases. London, p. 508. LINKED QUANTITATIVE CHARACTERS IN WHEAT CROSSES DR. GEORGE F. FREEMAN ARIZONA AGRICULTURAL EXPERIMENT STATION Sıxnce wheat has but 8 chromosomes in the sexual cells and since the parents in the macaroni bread wheat crosses here discussed certainly differ in more than 8 visible char- acters, it was thought likely that a genetic linkage of some of these might be found. The following study is an endeavor to discover whether or not there is such a link- age between the shape of the head, 7. e., ratio of width of head (measured parallel to the face of the head which shows the sinuous furrow formed between the two rows of spikelets) to the thickness (measured parallel to the array of seeds in the spikelet) and the texture (trans- lucency or opaqueness) of the grain. Of the two parents here discussed No. 1, the macaroni wheat, had a much flattened head and very hard translucent grains, whereas the other parent, No. 35 (a bread wheat), had a nearly square head with soft opaque grain. The seed on the F, plants of this cross were all wrinkled and intermediate in texture between the two parents, 7. e., they were dull, being neither translucent like the macaroni parent nor opaque like those of the bread wheat parent. In order to make a quantitative expression of hardness the perfectly translucent grains of the macaroni parent were called 100 per cent. hard, the seeds of the F, 50 per cent. hard and those of the bread wheat parent were called 0 per cent. hard. Since in the seeds of the F, and F, plants every possible degree of intergradation oc- -curred between the characters of the two parents and since any form of classification adopted would be purely arbitrary, it was decided to make it the most simple pos- sible and place all of the variants into three groups by the following means: a grain that was approximately as hard 683 [ Vou. LI THE AMERICAN NATURALIST 684 39S | TPS POT | T CZ £9 |72 |88| SZ ST) Q| 1T F neg “ON amg 10S £96, Fo | Q Z |8 (OT F9 99 |LLIS8S|\80S|\LFT|61| A |9 e OC Xt 9UBIPIUL9}Uy 166 T 88 F Q iP i0% |T i03 i88 i$ IZT 9IT 90ZIETE86T OZI | F116 |E NTA peH | 9433 69'T € | 3 | F | $B! Se \OFT! OF |918 163338 H CZZIF8T ak 5 tr IG 1h g “1 Ad AT OPBIPIMISJUT Q Og TL I r 6 oč oT 8 ie ‘ON 91nd prey | £93 oF’ I jog pe (29 jo jos oz lat re rje - upp syuerd ‘ujm | 612 | 69Z eoz | orz |6ez|6 eet ost | 6zt | eor ost ort eer | eat lorn 66 | 6s | 62 | 69 | 69 yo ayoq | “oN MOL | oðvway |04309093 | OF | OSZ | 0 $I ost | ot OLT 09T OST | OPT oet | oz1 ip TOOL Oe | om | OFT eae SSANMOIH, QIGI ‘SNIVED JO TUALXH, GNV ava] ao sae NITMI NOMVIGUIOD “TIT WIAVL WOS] 991 ¢6° sr lip |99|or|#t| e] [7 ueg “ON emd wyog| Zeg I'l 1 |t |e i vat og lee los t9lšel or g) [fa ce XT ayerpouayuy | QZZI | 62°T 1| 1|# le | £0] T Lh |8h| 89 OBI! 96 jezttocioertor prisla | 8 | Ble XT prey £69 eo'r IT I/Z|3|4/98/87|8£ |29|¥9 | 88 007| 89/898 i84 OF 162 18 |8 13 E Ca et ayeIpomieyuy | Zeg eG" I TIT |9 |9 |1@|Lh) 89 89 | LE \6E 9 193 |63 | TL |I “ong “ON Cig peH | 866 LoL TIZ iF |£ |38|99|98 |12 9€ 12% ZLI |8 jg ‘TT “eT CON amd uep sued LA igg 69% 69g | GFZ 6gz|6zz |6Tg G60 |66T [68T |6241 69T |GET! FT GET Si 601! 66 | 68 | 62 | 69 | 68 Jo WIWIVTO | “ON THIOL | oßeway OLE |096 |097 | OFE | O£ | OZE | OTZ | 008 |O6T [OST |OLT | O9T | OST OPT OET |0ZT OTT OT | 06 | o8 | oz | 09 | os SSANMOIH, GTGT ‘SNIVUD JO TUAGXay, ANV avay 10 feson NIAALIA NOILVIANYOO “IT AIAVL 1S ral 8" Rts ie ia GE "ON aan opeipamiayuy | 6g 9¢°T f ie Siero wa t Aes XT peH | OST eo | F Z| || ol ar} 2] 8 [ON aang 60° | 66'T | 68°T | GL°T | 69T | GST | GFT | GET | 6z't | GIT | eot | 66" | 6g” | Gz: | 69" nis is: 00° | O6'T | O8'T | OL'T | 09T | OST | OFT | OFT | OZT OTT | oot | 06" | Og | oz | op upp Jo IWO ‘ON IOL TN ennL onen anay PUIHI jo n JIGI ‘“SNIVYD wo A4AIXA1, ANV avaj ao 4 SANOL L harmana NOILVIANI0O ‘I ATV No. 611] LINKED CHARACTERS IN WHEAT CROSSES 685 as the macaroni parent was called 100 per cent. hard; a grain as soft as the bread wheat parent was called 0 per cent. hard; all grains intermediate between these were called 50 per cent. hard. A plant which produced 60 per cent. hard grains, 30 per cent. intermediate grains and 10 per cent. soft grains would therefore be classified as follows: .60 X 1.00 + .30 X .50 + .10 X 0—.75 = 75 per cent. hard. Spotted grains (grains containing well-defined areas of opaque starch in an otherwise translucent grain) such as occur frequently in the macaroni wheat and also among the hybrids were treated as intermediate grains. After classifying the seeds of each plant in this manner all those which were over 663 per cent. hard were termed hard wheats, those from 334 to 66% per cent. hard were classed as intermediate and all less than 334 per cent. hard were classed as soft. In 1914 the two parents (Nos. 1 and 35) and F, of this cross were grown with the following results: The ratio, W./T. of the F, plants of this cross, is thus seen to be much nearer to the macaroni parent both in average and distribution than it is to the bread wheat parent. i Table II shows this same study made for the F, plants, grown in 1915. _ A uniform correlation between the hardness of the grain and the ratio of width to thickness of head is ap- parent in Table II. Table III gives the results of this same study for the crop of 1916. There is again apparent a marked correla- tion between the ratio of width to thickness of head and the texture of the grain. It will be noticed that both in 1915 and 1916 the soft (0 per cent. hard) hybrids had a higher average ratio W./T. than the sonora. It should, however, be remarked that the hybrid softs were also harder on the average than the pure number 35, but since 686 THE AMERICAN NATURALIST . [Von. LI TABLE IV DISTRIBUTION OF PLANTS WITH REGARD TO THEIR PERCENTAGE OF HARD GRAINS AND THE RATIO OF WIDTH TO THICKNESS OF HEAD Pure No. 1 in 1915 Pure No. 1 in 1916 Percentage of Hard Grains Percentage of Hard Grains 100| 95 | 90 | 85 | 80 | 75 70 | 65 60 | 55 | 50 |100 95 | 90| 85! 80 75| 70) 65) 60| 55| 50 96| 91 | 86 | 81 | 76 | 71 | 66 | 61 | 56 | 51 | 46] 96 91 | 86 81| 76 71| 66) 61| 56| 51| 46 244 | 240| 1 | 239 | 235 234 230 229 225 1 224 220 1 219 215 214 210 1 1 1 209 205 204 2001 111 Err ee 12 199 195 194 190| 1 1 { 2-a 1 189 > 185| 5 E AEA be Seer 2d 184 0) 1 (237418151) 418 21741 3 2 1 40P ET) 64.41) Gr Si 5|3/3/8/6/10/10/ 4| 1 ļ10|3|3/2|1j1 2 IFL II 4]. Sila Epa A k 1 7141060/198 9/11/10 11t 11513/5/215/1/1/1 3 STAIB aB |10| 1)'9| 31 BTS Eyal 5}413/2/ 4] 6/11|14| 4125| 4/7/5/5|6)4/7/4/4 2\1/4!15\6} 3) 4/10/83] 6/2\3\1\1 1 1}2/4|2| 8] 2) 4/8411, 9/5) {1|1/814| [211 No. 611] LINKED CHARACTERS IN WHEAT CROSSES 687 Pure No. 1 in 1915 Percentage of Hard Grains 100 | 95 | 90 | 85 96| 91| 86 | 81 80 | 75 | 70| es | 60 | 55 sa 61| 56 51 50 46 Pure No. 1 in 1916 Percentage of Hard Grains 100 96 95 |9085 30| 75| 70 65 91 |8681, |76)7 71| 66 61 60| 55 8S 56 51 80 1 | 7 1 7 2 2 2/1] they were opaque and graded insensibly into the condi- tion of pure No. 35 it was considered impracticable to make the arbitrary designation of degrees of hardness any more complex than the three groups used. It should be added, on the other hand, that a number of races were secured which were as soft or softer than the type of No. 35 and a number of races which, to all appeafances, were * fully as hard as pure No.1. The inheritance of hardness will, however, be reserved for discussion in a later paper. The question now arises as to whether this correlation is genetic or physiological. Might it not be caused by the simple fact that poorly filled (with starch) hard grains - will give rise to a more flattened head, than will plump (starchy, soft) grains by failing to fill up and distend the 688 THE AMERICAN NATURALIST [ Von. LI glumes? This question can be answered by studying the correlation between the flattening of head and hardness of grain in a pure race. Since all of the No. 35 was soft, this study could not be made for the soft wheat, but as there were many plants of the pure macaroni which pro- duced a greater or less proportion of soft grains, a com- parison was possible. That the plants producing a large proportion of hard grains on an average did not have more flattened heads than those which produced a greater proportion of soft grains is shown in Tables II and III by comparing the ‘‘hard’’ with the ‘‘intermediate’’ groups of pure No. 1. This is perhaps better shown in Table IV, where the distribution of the plants is made with regard to their hardness per cent. and the ratio of width to thickness of head. oo It should be noted from Tables II and III that there are numbers of individual plants with low ratios of width to thickness of head, but with high percentages of hard grains. It is difficult to see how this could occur if the low ratio was due simply to the lack of plumpness of the hard grain. .Moreover, races of hard macaroni wheats occur which have approximately square heads, and there are varieties of soft wheats (Little Club) with rather strongly flattened heads. In 1916 there were a few cases where hybrid races having low average ratios (W./T.) also were rather high in average per cent. of hard grains. All of the races in 1916 having ratios averaging as lew or lower than 1.35 and average percentages of 60 or more per cent. of hard grains are given in the following table: It is thus seen that all of the races which on the average violated the correlation in 1916 came from plants which , more or less markedly violated this same correlation in 1915. The cases given would indeed be hard to explain on a basis of violation of physiological correlations but if we are dealing with a genetic correlation, they may be easily explained on the ‘‘cross over’’ theory as used by Morgan for reversal of linkages in the characters of Dro- sophila. No. 611] LINKED CHARACTERS IN WHEAT CROSSES 689 TABLE V INHERITANCE OF NON-CONFORMITY TO USUAL CORRELATION BETWEEN ater ATIO W./T. oF HEAD AND AVERAGE PERCENTAGE OF HARD GRAIN Parents in 1915 Offspring in 1916 | w./T. Per Cent H. wW./T. Per Cent H. 132 93 98 1.34 88 169 1.10 50 1.24 73 219 1,20 50 1.35 ‘93 232 1,25 100 1.33 86 239 1.33 1,30 88 263 1.00 64 1.17 61 279 | 1.27 50 | 1.35 79 In+the opposite direction, 7. e., races which markedly violated the correlation by abies very high ratios (1.50 or more) but with low percentages of hard grains did not occur. Again it may be objected that an explanation of this correlation between shape of head and texture of grain as a genetic linkage, is incorrect because the linkage is not complete, 7. e., there is considerable regression. This objection may be fully met by the observation that both of the characters here concerned are quantitative and hence subject to fluctuation around a mean. Moreover, it ` is almost certain that both characters are genetically com- pound, i. e., each are the result of more than one factor, recombinations of which may markedly vary the quanti- tative visible expression of the characters. If, therefore, but a single factor for grain texture be linked with one of the factors concerned in the shaping of the head, there will result a partial ae of these two characters such as we find. The data here presented, therefore, seems to indicate that the two characters, hardness of grain and high ratio of width to thickness of head, which entered this cross together in the macaroni parent, tend to come out together in the segregates of the F, and F generations, 7. e., that there is a genetic linkage between one or more of the factors controlling the grain texture and head shape in the two varieties employed as parents. ON REVERSIBLE TRANSFORMABILITY OF ALLELOMORPHS H. TERAO THE IMPERIAL AGRICULTURAL EXPERIMENT STATION, TOKYO, JAPAN In genetical studies of variegation in plants, the fact has been observed occasionally that with a certain fre- quency a dominant allelomorph occurs in the correspond- ing recessive homozygote (De Vries,! Correns,? and Emerson’). In this paper the author presents a new in- stance of a similar phenomenon, which it is hoped may throw additional light on the subject. In certain pedigree cultures of the rice plant, Oryza satwa L., there happened to occur in 1912 families con- taining besides ordinary fertile plants a number of sterile plants. These sterile plants were normal in their growth, but showed a considerable barrenness at the ripening season. Some of them yielded no seed whatsoever, others bore a small number of normal seeds, and very few were mosaic forms with higher fertility. These families, two in number, each belonging to a different variety, were derived from single plants of the former generation, and were very uniform in other characters. From them the experiment was started. The rice plant, being a self-pollinated species, is con- venient material for breeding experiments. Although the experiments in this investigation were made largely from open-pollinations, the results obtained were always similar to those from experiments in which plants were artificially protected against accidental natural crossing. The observations of 1912 and 1913 are shown in sum- marized form in Tables I and II, a and b, and point to the following conclusions. Sterility behaves as a simple re- 1De Vries, H., ‘‘Die Mutationstheorie,’’ Bd. I, 1901, pp. 489-511; ‘Species and Varieties, their Origin by Mutation,’’ 1905, pp. 309-339. 2Correns, C., Berichte der Deutschen Botanischen Gesellschaft, Bd. 28, 1910, pp. 418—434. 3 Emerson, R. A., AMERICAN NaTuRALIST, Vol. 48, 1914, pp. 87-115; Genetics, Vol. 2, 1917, pp. 1-35. 690 No. 611] - ALLELOMORPHS 691 cessive to fertility, and the seeds resulting from partial fertility of sterile plants again give segregating families. In Family A, which shows an exceedingly slight fertility of sterile plants, the segregation ratio in the offspring derived from fertile individuals is quite close to expecta- tion, but in Family B which shows a considerably higher grade of partial fertility of sterile plants, the progeny of fertile individuals exhibit considerable deviations from the expected segregation ratio. 3 TABLE I THE SEGREGATING FAMILIES, A AND B, IN 1912 Partial Fertility of © oe Sterile Plants Fam. Total | Ratio per 4 Total | Fertile Spikelets Fertile Sterile | No. of | Steriles % | No. of ene Plants Plants Ind. | D R gg No. | 4 | ! ca. | Aries 36 13 49 26.53 2.94 1.06 9,000 2 0.02 Ss sealers 105 25 130 19:23 | 3.23 0.77 |14,941| 434 | 2.90 TABLE II THE FAMILIES DERIVED FROM FAMILIES A AND B (a) The Progeny of the Fertile Plants No. of Families f Ratio per3 ` Segregating Families E Uni- wire eae Total EN APA PUEN | greentng| rota | SEM | gating | Pixie | Stace [Number Sve Families | Families vid A. 10 22 32 0.94 2.06 | 1,068 346 414 | 24.46 B. 41 64 105 E17 1.83 | 4,874 | 1,301 te 21.06 (b) The Progeny of the Seeds on the Sterile Plants ber Ferg” | Smeg’ | Fore Piante | suerte ants | 3Pianiduls | Stenio 5 | Aon. | Se | 2 | 0 | ee L bE 24 401 115 516 | 22.29 These facts may be interpreted by the following hy- pothesis. The dominant and the recessive types con- cerned are assumed to be transformed by certain un- known causes into the other allelomorph. The recessive allelomorph which has made its appearance in Families A and B is assumed to have originated in the preceding 692 THE AMERICAN NATURALIST _[Vou. LI generation by the transformation of the dominant allelo- morph. This recessive state of the hereditary substance, however, has a tendency to revert into the original domi- nant state. Such reversion is especially likely to occur in vegetative cells, where each recessive allelomorph seems to be able to revert independently. Consequently, in reces- sive homozygotes the reversion generally will produce heterozygotic cells, either one of the two recessive alle- lomorphs being changed into the dominant. The hetero- zygotic cells thus formed will give rise to partial fertility in otherwise sterile plants. Again, the recessive allelo- morph in heterozygotic cells may be subject to similar reversion, and such reversion may occur both in the heterozygotic cells of sterile plants and in normal hetero- zygotes. Here, however, heterozygotic cells will be transformed into dominant homozygotic cells without visible effect on the plant concerned. The consequence of this reversion in the next generation will be that the proportion of the dominant segregates may exceed the theoretically expected figure. Finally, it may be assumed that between Families A and B there exists a difference in the reverting tendency of the recessive allelomorph, which necessarily will effect the differences in both the intensity of partial fertility of sterile plants and the devi- ations in the segregation ratio. In Table ITI the segregating families derived from the fertile plants of Family B are classified according to the magnitudes of the deviations in terms of probable errors. The true percentage for the recessive is assumed, in the one case as 25 per cent. (the Mendelian ratio), and in the other case as 21 per cent. (an arbitrary number). In comparing the two different frequency distributions made in this manner with the theoretical frequency distribu- tion, it is observed that while the frequency distribution of the deviations from 25 per cent. shows a considerable discrepancy from the theoretical, the latter fits the fre- quency distribution of the deviations from 21 per cent. rather closely, the goodness of fit being P = 0.915. Con- sequently, the ca. 4 per cent. deficiency of recessive segre- No. 611] ALLELOMORPHS 693 gates is a normal expectation and not an experimental error. TABLE III THE FREQUENCY DISTRIBUTION OF THE DEVIATIONS IN THE SEGREGATION RATIOS IN THE GROUP OF 64 SEGREGATING FAMILIES DESCENDED FROM FAMILY B OF THE YEAR 1912 Dev. /P.E. i —5 —4 —3 -2 —l 0 +i +2 +3 +4 +5 Total I } | | | | + | | | Experimental frequency (I). . iher 14; 17 15} | 2l Experimental frequency HD: aS a il 10 64 Bx pecsation:.G. os ee ee pdt 10.3 16.0, ed ACECHAR Note: In the apaina MERRTE (I) the true ai for re- cessives is taken as 25 per cent., and in (II) as 21 per cent. Such an aberrant segregation ratio seems to be a con- stant tendency all through the generations descended from Family B. This is shown in Table IV in which the ex- periments in the years from 1912 to 1915 are summarized. TABLE IV THE ABERRANT SEGREGATION-RaTIOS OBTAINED IN THE Years 1912-1915 | i | Years Fame, | plouts. | Inds. | Fertiles | Steriles | Ster. # | Dev. # | P.E. 4 yew. `- WE rie. vy 1 | Fertile 130) 105|. 25 | 19.23| 6.77} 2.55} 23 IMS “i 64 r 6,175| 4,874 | 1,301 | 21.06| 3.94 | 0.37 | 10.6 {b14 49h 40 “ 1,560] 1,207 | 353 | 22.63| 2.43 | 0.74 | 3.3 161g eT: 53 a 4,696] 3,732 | 964 | 20.52 |. 4.48 | 0.47| 9.5 Total....| 128 12,561| 9,918 2,643 | 21.04! 3.96 | 0.26 15.2 i a.: 24 | Sterile 516| 401 | 115 | 22.29! 2.71 | 1.21} 22 it. dk 34 ‘t 994| 779| 215 | 21.63| 3.37 |: 0.93 | 36 1018 23>: 19 a 684| 522| 162 23:68 | 1.32 iil 1.2 H Í Tota.: d 7 2,194| 1,702 | 492 | 22.43| 2.57 | 0.62 | 4.1 Again, in regard to the intensity of partial fertility of sterile plants, the descendants of Families A and B ex- hibited respectively relations similar to those seen in 1912. (Family A was not traced after 1913.) A count of fertile spikelets on sterile plants descending from Family B was made in 1914 on 281 plants bearing a total of 101,412 spikelets. In this count the number of fertile spikelets was 3,857, corresponding to 3.78 per cent. of the total number of spikelets. The latter figure may be regarded _as the average fertility of sterile plants in the progeny of Family B. : 694 THE AMERICAN NATURALIST [Vou. LI The fertile spikelets of sterile plants are generally scat- tered at random over the panicle, and each fertile spikelet may be regarded as representing a separate case of re- version; but in mosaic forms which show higher fertility and are of rarer occurrence, the reversion may have taken plaee in earlier stages of plant development, result- ing in larger fertile sections. Consequently, when the count of fertile spikelets is made with only the first type of sterile plants, a more correct value for the frequency of reversion may be obtained. The result of such a count on 902 panicles containing 93,635 spikelets is 1,858 fertile spikelets, i. e., 1.98 per cent. of the total number of spikelets. The mosaic forms appear in several different grades of partial fertility. In a panicle either one or more branches or even one half of the panicle can be highly or entirely fertile, the remaining part being absolutely or nearly ab- solutely sterile. Similarly, in a single plant some whole panicles can be entirely or highly fertile while others are of the ordinary grade of partial fertility. Furthermore, similar mosaic conditions were also observed in single » flowers of sterile spikelets. While all six anthers of a sterile spikelet generally bear none or but few pollen grains, occasionally flowers appear in which certain anthers contain a considerable number of pollen grains of normal appearance and others show the ordinary state of sterility. Hence it may be assumed that the reversion can take place at any stage of plant development. The partial homozygosity of heterozygotes, correspond- ing to the partial fertility of sterile plants, may be esti- mated in the following way. Assuming that the possi- bility of reversion at any stage of a plant’s life, similar to that observed above, may also occur in heterozygotic cells, then we may distinguish for convenience two differ- ent types of reversions; there is the reversion which will cause partial homozygosity within a single flower, and the reversion which will produce an entirely homozygotic spikelet or larger homozygotic sectant. Suppose then that the latter reversion will give to the heterozygote No. 611] ALLELOMORPHS 695 homozygotie (AA) spikelets in any part ‘‘x’’ of the total number of spikelets which is taken as a unit, and again that in the remaining (1 — x) part of the total number of spikelets, the other type of reversion will occur, turning some part ‘‘y”’ of the whole generative tissue taken as a unit from the Aa state to the AA state. For simplicity, however, we may substitute ‘‘x’’ for ‘‘y’’ in the above re- lation, because it seems presumable that a similar prob- ability of reversion may exist constantly all through the plant life. Such a plant will have the following consti- tution in regard to the generative tissue: x(AA) + (1—x)[x(A‘A) + (1 — x) (Aa)]. As the result of self-pollination, the progeny of such a parent plant will show the constitution: x(AA) + (1 —x)[4(1+ x)?(AA) + $(1 — x?) (Aa) +4(1 — x)?(aa)]. Applying arbitrary values to ‘‘x’’ in this formula, we shall get numerical relations among segregates. In Table V the results of such calculation are compared with results obtained by the experiments in 1913-1915. Thus we may find the average partial homozygosity of hetero- zygotes around 4 to 6 per cent., the average partial fer- tility of sterile plants being, as was already shown, ca. 4 per cent. TABLE V : CALCULATIONS ON DATA OF TABLE IV ; x % (AA+Aa) aa AA Aa ag eee 77.88 % 22.12% || 38.74% | 61.53% Sts See eae 78.57 21.43 39.69 60.31 © Gas shes sss ere ee ee 79.24 20.76 40.89 59.11 Observation, No. of Inds...... 9,918 2,643 94 . 1135 Parodtitegh > e oo a 78.96 % 21.04% 41.05% | 58.95% It has also been noticed that the sterility concerned is associated with an abnormality represented by the be- havior of chlorophyll at the ripening of seeds. While, at the ripening season, the chlorophyll in the fertile sections of the mosaic forms turns to yellow just as in ordinary fertile plants, the chlorophyll in the sterile sections still 696 THE AMERICAN NATURALIST [Vou. LI remains green. The fertile spikelets occurring in a small number on the otherwise sterile panicle appear on rip- ening as yellow spots scattered among green spikelets; the plants with both sterile and fertile panicles appear in the fall also as mosaic forms with green and yellow leaves. This feature of the sterile plants is in direct contrast to the behavior of the mosaic plants with the variegated and the entirely green leaves studied by De Vries and Correns. The observations in the foregoing pages seem to paral- lel those made by the authors cited at the beginning of this paper. In the present investigation, however, there was observed also the transformation of allelomorphs in the opposite direction, that is, the transformation of the dominant allelomorph into the recessive allelomorph, something scarcely mentioned in the investigations re- - ferred to above. The observations in this regard were in brief as follows. In the first place, the spontaneous occurrence of segre- gating families was observed again among the descend- ants of the families which had proved in the experiments already described to be constantly fertile. This suggests, just as did the occurrence in Family A and Family B in 1912, the probability of the AA cell changing into the Aa cell. In the second place, a constant tendency of the dom- inant allelomorph to be transformed into the recessive allelomorph. was observed in certain strains. In 1913, special attention was paid to such segregating families in which the excess of recessive segregates over the theo- retical expectation was particularly high. Although, as already noted, the variation among the segregating fam- ilies in 1913 with regard to the deviations from the reces- sive proportion might possibly have arisen from experi- mental errors associated with a certain probability of alle- lomorphic reversion from recessive to dominant, yet it was deemed not impossible that the very considerable excess of recessives exhibited by certain families might be caused by other reasons. This point was seemingly de- No. 611] ALLELOMORPHS 697 cided by the experiment made with Family B80 in 1913 (Table VI), since in this family there was noticed a con- stant tendency toward the allelomorphic transformation under consideration. TABLE VI THE SEGREGATION OF FAMILY B/80 AND Irs DESCENDANTS > | No. of Parent. a | oni Sterile aaa of se Families plants Brent Plants | Plants | Oye | Rocessives | 7 = | wa. 1 Fertile 991 69| 30| 30.30% | + 5.30% | 2.95% 1914.... 10 $ 1,020 Tal 293 28.73 + 3.73 0.9 1916... 5 4 435 309 126 28.89 + 3.89 | 1.40 1916:... 98 A 11,013) 7,832; 3,181 28.88 + 3.88 0.28 Total. . 114 Fertile gaa 8,937 | 3,630 28.89%-| + 3.89% |: 0.26% | | ; 1914 (a). 161 | Sterile 199, 147| 52 |- 23.62% | — 1.38% | 2.04% 1914 (b). 131 Sterile 100 5 95|- 95. oo% | oe 00% 2.92% aki 592 ie 548 32 516 94.1 1.25 1916....| 1202 | i 1,436) 99| 1,337| 93. ii E mar is | 0.77 Total. . 1923 | Sterile 2,084 136| 1,948| 93. 47% | | +68. 17% 0.64% 1 Derived from the family in 1913, i. e., Family B/80. 2 Derived from the group (d in aie? 3 Excluding the group (a) in In Table VI there is beside the ca. 4 per cent. ex- cess of recessives in the families derived from fertile parents, a remarkable excess of recessives in the families descended from the sterile parents in the group (b) in 1914. The sterile plant of this type could not be distin- guished from those which, as was shown in Table IV, gave segregating families with an excess of dominants in the intensity of the partial fertility as well as in the be- havior of chlorophyll at the ripening of the seeds. Con- sequently, it may be presumed that although these two types of sterile plants have the same genetical constitu- tion originally, the dominant allelomorphs resulting from the reversion of their recessive allelomorphs are of dif- ferent stabilities in the dominant state; that is, in the first type of sterile plants such dominant allelomorphs are very easily re-transformed into the recessive state, while in the second type the corresponding dominant allelo- morphs tend to remain in the reverted condition. 698 THE AMERICAN NATURALIST [ Von. LI Corresponding to the excess of recessive segregates, a deficiency of dominant homozygotes among dominant seg- gregates was also noticed. Among 153 families derived from fertile plants in the experiment above mentioned, 40 families were uniformly fertile, the remaining 113 fami- lies showing segregation. The former, therefore, is 26.14 per cent. of the total number of families, and shows 7.19 per cent deficiency from the theoretically expected percentage, 33.33 per cent., the probable error being + 2.68 per cent. In conclusion it may be stated that the allelomorphs concerned in this investigation are probably subject to reversible transformations, and that the probable fre- quency of the allelomorphic transformation may be prac- tically constant in a certain strain, and possibly may be different in different strains. As to the conditions under which such allelomorphic transformations take place, nothing is yet certain except that these conditions are of a hereditary nature. The manner in which different in- tensities of allelomorphic transformations are inherited will be the subject of further investigation. A word may be added here regarding the conception of dominance and recessiveness. Bateson’s theory of ‘‘presence and absence of factors’’ is sometimes under- - stood in the sense that the dominant allelomorph is re- garded as due to the real presence of an hereditary mate- - rial unit which is absent in the recessive allelomorph. Such a conception is not in full accordance with the idea of the reversible transformability of allelomorphs as de- seribed in this investigation. There is another possibility of the nature of allelomorphs. The dominant and the re- cessive allelomorphs may be supposed to represent two alternative conditions or phases of a single hereditary substance, somewhat resembling the chemical conception of polymerization. Consequently, the interchangeability between the dominant and recessive allelomorphs is not improbable theoretically. BUSSEY INSTITUTION, August 26, 1917 NOTES AND LITERATURE MUTATIONS IN DROSOPHILA BUSCKITI COQ.* Two mutations in eye color have appeared in my cultures of Drosophila busckvi. These mutations are of especial interest in that, as far as the writer has been able to learn, they are the first that have been recorded in this species. This is the eighth species of Drosophila in which mutations have been recorded, the other seven being ampelophila, repleta, confusa, tripunctata, virilis, obscura, and similis. The eye mutant which is brighter than normal has been ealled ‘*red’’ and the other which is darker than normal has been called ‘*ehocolate.’’ The normal eye of this species is darker than that of Drosophila ampelophila. The mutant red corresponds very closely to the normal eye color of ampelophila except that it is slightly brighter. Ridgeway’s ‘‘scarlet’’ (Plate I, color num-- ber 5, Ridgeway’s Color Standards and Nomenclature, 1912) corresponds most nearly to the eye color of this mutant. In the red eye the central fleck shows as a small round point, while in the normal busckii eye, it appears larger and less definite in shape. The red eye darkens with age and closely approaches the normal eye in color, but at its darkest stage it can be dis- tinguished from the normal in that it is less translucent. The chocolate eye is an opaque brown and presents none of the shiny appearance of the normal eye. The central fleck is invisible in newly emerged flies, but becomes more or less distinct as the fly ages. With age the color approaches normal, but always remains slightly darker. Flies over forty-eight hours old so nearly ap- proach normal that they are difficult to distinguish. A newly emerged chocolate corresponds most nearly to Ridgeway’s ‘*chestnut brown’’ (Plate XIV, color number 11, tone m}. The mutation red eye was first observed in November, 1916, and it seems probable that the original mutation occurred some- what earlier and was overlooked as several red-eyed flies, both ` males and females, were obtained from this cross. The original] stock had been collected about a month earlier in a tomato patch 1 From the Zoological Laboratory of Indiana Caen, Contribution © No. 157. 699 700 THE AMERICAN NATURALIST [Vou. LI near Bloomington, Indiana, and had been bred in the laboratory for two generations. The mutation chocolate eye was first observed in December, 1916, and here again the original mutation had probably been overlooked as in the cross where the mutation was first ob- served, several chocolate males and females appeared. The orig- inal stock in which this mutation appeared had been bred in the laboratory for three generations and was collected in the same tomato patch where the original red eye stock was collected. The stock in which the red eye appeared was collected on Sep- tember 19, 1916, and the stock in which the chocolate eye ap- peared was collected September 14 of the same year and these stocks had been bred as two separate strains when the mutations appeared. Tue Genetic BEHAVIOR or Rep Eye Some of the first observed red males were mated to virgin red females and the offspring of this cross were all red flies. This red stock has been kept going for several generations and has ‘given all typical reds. Both red males and red females were crossed with normal wild flies and in F, of each cross nothing but normal flies were found. Table I gives the results of the F, of these crosses. TABLE I F, oF Rep X Witp Cross Red 9X Wild g ' Culture Number Type of Mating | a aoan TOSE Eea = Leia 9 a 9 Normal) Red gaiu eased, 1 pair 317 | 322| 125| 119) 639| 244 og aparece pa aoa 305 | 313|. 109| 107| 618| 216 eee ERIE “ou 282| 100 557 | 198 ENAR TS “ou 326| 299 101) 625|: 1 a Mass 557| 552| 112|. 150|1,109| 262 pee SG i da 1,780.1 1,768 | 543 | 575 | 3,548 | 1,118 Red F X Wild 2 wo 340| 326| 101| 105| 666 e A mar 269| 293| 79| 88| 562| 167 pid hs PA “ou 231| 258}. 71 484 BM ars, “u 244} 250| 5 494| 145 rE a Mass 469| 466| 87| io0| 935| 187 POM anirai y a “ 228 4 1 Toth, inae. Be OE 1,818 | 1,816 | 487 555 | 3,634 | 1,042 Grand hii ice eis... 3,598 | 3,584 | 1,030 | 1,130 | 7,182 | 2,160 No. 611] NOTES AND LITERATURE ` 701 It can be readily seen that the red eye acts as a non-sex-linked recessive character with the red class falling a little short. Shortage of the mutational class frequently occurs. It seems to make no difference whether the red male or red female is used in the cross. In most cases when mass cultures were made, the ex- pected 3 to 1 ratio was less nearly approximated. The ratio of all the flies examined in the F, was 3.32 normals to 1 red. THE GENETIC BEHAVIOR OF CHOCOLATE EYE A pure stock of chocolate was obtained by mating some of the first observed chocolate males to their virgin chocolate sisters. This stock has bred true for several generations but since the eye changes so rapidly to a color approximating normal, in stocks where the flies are allowed to become more than twenty- four hours old, all gradations between the typical chocolate and normal will be found. The chocolate males and females were bred to wild normals and the F, flies were all normal. The re- sults of the F, of these crosses are shown in Table II. TABLE II F, OF CHOCOLATE X WILD Cross Chocolate 9 X Wild g Normal Chocolate Total | Total y of Mati y Choc- Culture Number ype ating 3 9 e 9 Normal inns cy Seapine oe Gaver ies aS 1 pair 12 43 95 475 138 aoas Ub ae ree 154| 146|° 49 | 60 Oy ce cs “ou 223| 181| 57 | 109 | 404| 166 EE A: | Mass 529| 468| 67 | 136 | 997| 203 Total. ...; «. 1,169 | 1,007 | 216 | 400 | 2,176) 616 Chocolate &' X Wild Q 408 1 pair 257| 252| 58 509| 124 Boe Caer re 8 45 | 61 448i SO Ee 338 “ou 343| 312| 58 | 82 | 655| 140 P due “ou ad 67|.17 | 17 | 150). 34 rT oo Mass 812| 667| 78 | 161 |1,479| 239 ry meee ae s 426| 440| 106 | 133 | 866| 239 Ga a e E 2,159 | ?,980.| 362 | 520 |4,139| 882 Grand total, i .| 3,328 | 2,987 | 578 | 920 |6,315 | 1,498 Chocolate eye also acts as a non-sex-linked recessive character with the chocolate class falling considerably below the expected 702 THE AMERICAN NATURALIST [Vou. LI number. The number of chocolate males, especially, falls low. The totals seem to indicate that this low number of chocolate males in comparison to the number of normal males is partially due to the fact that some of the chocolate males have been called normal, for in practically all of the matings, the number of normal males exceeds the number of normal females. This is unusual, for the writer has examined large numbers of wild Drosophila buscku and in a large majority of the cases the num- ber of females has been equal to or greater than the number of males. So it may be that the males approximate the normal color more rapidly than the females and since the flies were ex- amined only once a day, some of the chocolate males were mis- taken for normals. Some counts were made, examining the flies twice a day, to test this supposition and they indicated that better ratios could be obtained in this manner. But since the work was completed before this was realized, the difference to be obtained by twice-a-day counts was not thought to be of sufficient importance to require the repetition of the experi- ments. Also in these matings, the relative number of chocolates was lower where mass cultures were made. The ratio for all of the F, flies examined was 4.21 normals to 1 chocolate. THE GENETIC BEHAVIOR OF RED AND CHOCOLATE WHEN THEY ARE MATED TOGETHER Red males were crossed to chocolate females and red females to chocolate males and in the F, of each cross nothing but normal flies appeared. The results of the F, of these crosses are given in Table ITI. Here again the number of normal males is considerably above the number of normal females. This could be explained as before, that some of the chocolate males have been mistaken for normals, thus increasing the normal class and decreasing the chocolate class. Since the number in. the classes of red and chocolate each fell low in their respective crosses to wild, we can expect the number in these classes to be low in this cross. Ta- king this fact into consideration, the ratio can be considered a 1:2:1 ratio and gives indication of linkage between the two characters. No red-chocolate double recessives were found, therefore the two mutations may be interpretated as being -in the same chromosome. No. 611] NOTES AND LITERATURE TABLE III F, oF Rep X CHOCOLATE CROSS Red & X Chocolate 2 Culture | Type of | cco a | mee | preter | Knog Total | Total i Ch ence E fart ee | Jing | a | 9 | eae Tae £02 CI | 1 pair 237| 206) 95} 122 86 | 92| 443 217 | 178 ABR Spee f 149| 132} 70i- 56] 50l 48] 281] 126 452. i 21291 184| 78i 108} "73 83| '396| 186| 156 453 6 Hine o +8 207| 186 85} 70} 93| 393] 17 Tt ES e | te 246 | 238| 114| 129| 88 | 106| 484| 243| 194 a. | Mass 373| 360| 149| 153| 85 | 109| 733] 302} 194 Pee Riser V hemmed OY 1,306 | 592| 653| 458 | 531 2,730 1,245! 989 “Red 2 X Chocolate g aBtaralati d lpair | 232| 178] 21|. 22). 41 72| 4 410 0, 4 113 AE ae oo 173| 163} 62] 68] 48 nd 336 | 130 107 450 SS eee 214] 178| 79] 66) 66 = 145| 126 MODE is. roi gmi 1 179| 93| 99| 50 a 192| 108 6 aS 243) 249! 94] 117| 90| 94 les f211| 184 yy eee Saran | Mass 370| 326| 121| 156| 90| 129! 696| 277| 219 Totale. oorh sacs nas r LABI] A70) 6281 385 | 472 |2,691| 998| 857 Grand E E .. esse + |2,842 [2,579 |1,062 |1,181| 843 (1,003 (5,421 |2,243 |1,846 Since mutations have occurred in eight species of Drosophila it seems probable that mutations may be found in all the mem- bers of this genus. As to the frequency of mutations, there may be individual variation. The writer’s own experience would in- dicate that mutations occur less frequently in busckti than in ampelophila, for, during the same period in which the two busckii mutations were found, a smaller number of ampelophila were examined less critically and six mutations were found. LITERATURE CITED Metz, Chas. W. 1916. Mutations in Three Species of Drosophila. Genetics, I, 591- Morgan, Sturtevant, Muller, aud Bridges. 1915. The see anism of Mendelian Heredity, pp. xiii + 262. New York, Henry Holt and Co. Don C. Warren ALABAMA POLYTECHNIC INSTITUTE, N, ALA. 704 THE AMERICAN NATURALIST [Vou. LI SINGING MICE In November, 1916, Mr. B. S. York, of Ann Arbor, brought to me a “‘singing’’ house mouse that had been captured in his home. This mouse had been heard by members of his family for several weeks, especially late at night and early in the morning. Arrangements were made to carry on breeding ex- periments with it but it lived only two weeks. Singing mice have been recorded in a number of publications dating back many years. In 1912 Coburn! reported some work he had done with a female singer captured in December, 1911. This individual when mated with an ordinary mouse gave birth to five litters (thirty-three young). None of these were singers and no singers appeared in either the second or third genera- tions. Two other singing mice were described by Coburn in 1913.2 One was caught in the home of an Italian family in November, 1912, and the other was taken by a farmer in Mich- igan in March, 1913. Both of these were females. The Ann Arbor specimen that was brought to me also proved on dissection to be a female. Her song was similar to that re- ported by Coburn as follows: The sound is best described as a rapid whole-toned trill involving the tones c and d. .. . The quality of the tone resembled somewhat that of a fife or flute, but each tone ended with a slight throaty click. In every case the song could be heard at least 15 or 20 feet away. Many causes have been proposed for the presence of this ability to sing such as pregnancy, a diseased condition of the lungs or vocal cords, a parasitized liver, ete. There were no embryos or young in the Ann Arbor specimen and Dr. George R. LaRue was unable to find any parasites that could have in- duced the singing. : It has been suggested that since all of the singers captured thus far have been females, this characteristic may be sex-linked and due to some structural modification of the vocal apparatus. R. W. HEGNER UNIVERSITY or MICHIGAN 1 Coburn, C. A., Journ. Animal Behavior, Vol. 2, 1912, pp. 364-366. 2 Coburn C. A., Journ. Animal Behavior, Vol. 3, 1913, p. 388. THE AMERICAN NATURALIST Vou. LI. December, 1917 No. 612 THE GENESIS OF THE ORGANIZATION OF THE INSECT EGG. I PROFESSOR ROBERT W. HEGNER, ZOOLOGICAL LABORATORY, UNIVERSITY OF MICHIGAN 5. Interaction of Nucleoplasm and Cytoplasm There are phenomena that occur during the growth period that suggest how masses of cytoplasm that are differentiated both morphologically and physiologically may arise in the cortical layer of the insect egg. It has been suggested that ‘‘most of the differentiations of the egg cytoplasm have arisen during the ovarian history of the egg and as a result of the interaction of nucleus and cytoplasm; . . „°? and with this we fully agree, but our problem is to determine the nature of this interaction and in what ways it may take place. During every mitosis there is a more or less thorough mixing that involves the chromatin as well as other nuclear constituents, since chromatin-diminution is a normal his- tological process. Interchanges between nucleus and — cytoplasm, therefore, occur during the two multiplication periods that precede the formation of oocytes. Abun- dant opportunity is thus offered for factors in the chromo. somes to exert an influence upon the cell as a whole. A similar and probably even greater discharge of chromatic 27 Conklin, 1916, ‘‘ Heredity and Environment,’’ New York. 705 706 THE AMERICAN NATURALIST [Von. LI and other nuclear substances into the cytoplasm occurs during the maturation divisions of the egg, but this period may be neglected in this connection, since the organiza- tion with which we are concerned is already established before maturation takes place. Even when the nuclear membrane is intact, substances undoubtedly pass in and out of the nucleus much as they do through the cell mem- brane, and as in the latter, the nuclear membrane may change in permeability at different times, these changes being due to chemical processes taking place within the nucleus or in the cytoplasm. Such changes occur more often during periods of cell activity than at other times and thus we should expect pronounced interaction throughout the growth period of the oocytes. Besides gradual, and for the most part invisible, inter- changes of this sort there may be actual transference of visible masses of chromatin from the nucleus to the cytoplasm. These chromatin granules that escape into the cytoplasm have been called ‘‘chromidia’’ and are sup- posed to play a part in cytoplasmic differentiation. A peculiar process of interchange by means of sec- ondary nuclei is exhibited by certain insects, especially Hymenoptera.2® This process has been studied most _ carefully in the carpenter ant, Componotus herculeanus var. pennsylvanica (Fig. 11). At an early stage in the growth of the oocyte small vesicles containing a few granules of chromatin appear near the oocyte nuclei. These ‘‘secondary nuclei’’ appear to arise as buds from the primary nucleus, but no one has yet actually observed their formation in this way. It has also been suggested that they may be epithelial cells that have invaded the oocyte, but this seems very improbable. The writer has reached the conclusion that they consist of nuclear ma- terials that have been given off into the cytoplasm and have there become enclosed by membranes which give them a nuclear-like appearance. As the oocyte increases 28 Blockmann, 1886, Festsch. nat.-med. Verein zu Heidelberg; Buchner, 1913, Biol. Centribl., Bd. 33; Hegner, 1915, Journ. Morph., Vol. 26. No. 612] GENESIS OF ORGANIZATION OF INSECT EGG 707 in size the secondary nuclei increase in number until they entirely surround the primary nucleus, forming several layers. When the oocyte has nearly reached its full growth they begin to migrate from the group near the anterior end of the oocyte and become scattered through- Vv) J) a. aes SL ARI AD. 2h oS © 2.6 GG yes = @5 O o % 009%. Fie. 11. Secondary nuclei in the oocytes of the ne pte ant (a, b, c) and the “Blynienopterous gall-fly, Rhodites ignota (d). (Hegn 1915. a. Oocyte (o) shortly efter secondary nuclei (s) ating io appo = nurse cells. b. Older oocyte showing oocyte nucleus (0) ee by secondary nuclei (s). pp tion between oocyte and n hamber. . Part of a pHi eae oocyte inowtne follicular epithelium, yolk globules (ack) ena secondary nuclei. of an oocyte a Rhodites showing primary nucleus (large circle) and ies aie (small circles). out the egg, forming a rather oe layer a short dis- tance beneath the periphery. The further history of these bodies is not certain, but they undergo changes by which they lose their identity, since they can not be found in fully grown eggs. Their function is likewise prob- lematical. They may take part in the formation of germ- line determinants which probably occur in the eggs of 708 THE AMERICAN NATURALIST [Vou. LI this ant;?° they may aid in changing the substances fur- nished by the nurse cells into material available for the embryo ;*° or they may have something to do with the for- mation of yolk.*! It is also possible that they may con- ` trol differentiation in the peripheral layer of cytoplasm and thus provide a method of nuclear control of the or- ganization of the egg. The last hypothesis may be ob- jected to on the grounds that the secondary nuclei appear to be irregularly distributed and that they are known to occur in only a few species of insects. Another possible way in which the initial organization of the insect egg may arise is through the activities of mitochondria. The rather constant presence of these bodies in the cytoplasm of almost all types of cells indi- cates that they may be of considerable importance in the process of differentiation. If they take part in the gen- esis of egg organization they then may play the rôle attributed to them by certain investigators of being the cytoplasmic bearers of hereditary factors corresponding in this respect to the nuclear bodies of similar function, the chromosomes. The most striking differentiation in the cytoplasm of the insect egg is that which involves the germ-line deter- minants. As stated above, we do not know for certain in any case the origin of the peculiar cytoplasmic mass that contains these determinants, but a number of hypotheses have been suggested. In Miastor, for example, the fol- lowing is offered to account for the appearance of the ‘‘pole-plasm’”’ in the fully developed oocyte.*” It may be distinguished from the rest of the egg contents by its posi- tion at the posterior end and because of its affinity for certain dyes. It appears shortly before the maturation division is initiated, but no transition stages have been discovered—it has been either present or entirely absent in the preparations thus far studied. If we consider the history of this substance from the formation of the primordial germ 29 Hegner, 1914, Journ. Morph., Vol. 26. 30 Marshall, 1907, Zeit. wiss. Zool., Bd. 86, 31 Loyez, 1908, C. R. Assoc. Anat., 10 Reunion, Marseille. 32 Hegner, 1914, ‘‘Germ-Cell Cycle in Animals,’’ New York. No. 612] GENESIS OF ORGANIZATION OF INSECT EGG 709 cell to the growth period of the oocytes produced by this primordial germ cell, we may conclude that at the time the multiplication perio ends the pole-plasm has become equally distributed among the sixty- four oogonia. Then ensues the growth period during which the pole- plasm can not be distinguished. Later, however, just before matura- tion, pole-plasm substance reappears which is equal in amount to that contained in the primordial germ cell of the preceding generation or to that contained in all of the sixty-four oogonia which descended from that primordial germ cell. That is, the pole-plasm of the oocyte under discussion has in some way increased until its mass is sixty-four times as great as that of the oogonium before the growth period began. How this increase has taken place can only be conjectured. The pole-plasm in the oogonium may have produced new material of its own kind either by the division of its constituent particles or by the influence of its presence. The influence of a specialized mass of cytoplasm upon the chromatin is very well illustrated by the inhibition of chromatin-diminution in Miastor and Ascaris. In Miastor nuclear division is normal until at the four-cell stage one nucleus reaches the pole-plasm at the posterior end (Fig. 12, a, IV.). During the succeeding mitosis this nucleus, which is apparently under the control of the pole-plasm, does not undergo chromatin-diminution, whereas the other three do. One of the daughter nuclei resulting from the division of this undiminished nucleus remains entirely within the pole-plasm and is cut off from the rest of the egg with this specialized mass of cytoplasm as the primor- dial germ cell (Fig. 12, b). This nucleus always retains the full amount of chromatin; but its sister nucleus, which ` remains in the egg and is thus separated from the direct influence of the pole-plasm, undergoes diminution at the next mitosis. A similar segregation of specialized cytoplasm in the primordial germ cells occurs also in certain other insects and in copepods, but no diminution process has yet been discovered in them. In Ascaris, where chromatin-diminu- tion was first reported,** there is evidently a segregation of germinal cytoplasm at each cleavage division up to the sixteen-cell stage, when it is all confined in one cell, the 33 Boveri, 1887, Anat. Anz., Bd. 2. 710 THE AMERICAN NATURALIST [Vou. LI primordial germ cell. This cytoplasm, which is not vis- ibly different from the rest, as in Miastor, appears to in- hibit diminution in every nucleus that comes within its a, Longitudinal section through an egg of Miastor showing chro- Pe alt ee ge uclei I. and III. but not in nucleus IV. which has come under the influence of the pole-plasm (p Pl). Nucleus II. does not appear in this section Mp = chromatin that is cast off into cytoplasm. ne= nurse cells. pb= polat body. (Kahle, 1908.) . Longitudinal sete E an egg of Miastor, norae nenn germ cell (p.g.c.), nuclei undergoing chr Heng: ea (cMp), and the remains of chromatin cast out into the hie (cR), c= cytoplasm inakit Oy nurse cells above. (Hegner, 1914.) immediate influence as indicated by experimental studies on dispermic and centrifuged eggs.** In this respect it resembles the pole-plasm of Miastor. 34 Boveri, 1910, Arch. Ent.-mech., Bd. 30; Boveri, 1910, Festschr. R. Hertwig., No. 612] GENESIS OF ORGANIZATION OF INSECT EGG 11 6. Mendelian Factors and Cytoplasmic Organization The central biological problem of the present time is the method of evolution, and a knowledge of the mech- anism of heredity has long been recognized as necessary for its solution. The results derived from breeding ex- periments with the fruit fly, Drosophila ampelophila, have dominated the field of genetics for the past five years, but although of very great interest and importance, their evolutionary significance is not yet certain. To be of primary value from this viewpoint it is necessary to prove that new species may arise by means of Mendelian characters (mutations) such as white eye, miniature wing, club wing, ete. Since no one has ever been able to define satisfactorily what a species really is and hence what characters should be considered of specific value, this is a difficult problem. The definitions given by two of our foremost authori- ` ties, one a systematist and the other a geneticist, are as follows: The systematist writes :*° Forms of animals which present distinct assemblages of characters, in form, color and arrangement of parts under natural conditions, which are recognizable from descriptions and figures, should receive distinctive names and be catalogued, provided, of course, that the assemblage of characters includes all ontogenetic changes. If, in the examination of abundant material from different natural environments, we find these characters fairly constant, the forms may properly be called species; if not, varieties or races, The geneticist writes :°° Species may thus be distinguished by peculiarities of form, of num- ber, of geometrical arrangement, of chemical constitution and pro erties, of sexual differentiation, of development and of many other prop- erties. In any one or in several of these features together, species may be found distinguished from other species. The mutations that have appeared in Drosophila do not become recognizable until a late stage in the life his- tory of the individual, and are about the last characters 35 Williston, 1908, AMER. NAT., Vol. 42. 36 Bateson, 1913, ‘‘ Problems of Genetics.’’ 712 THE AMERICAN NATURALIST [Vou. LI to appear in the individual development. They for the most part affect the size and shape of the wings, the size, shape and color of the eyes, and the color of the body. If a systematist were asked whether these new races of Drosophila are comparable to wild species, he would not hesitate for a moment. He would call them all one species. If he were asked why, he would say, I think, “These races differ only in one or two striking points, while in a hundred other respects they are identical even to the minutest details.” He would add, that as large a group of wild species of flies would show on the whole the reverse relations, viz., they would differ in nearly every detail and be identical in only a few points.7 This point of view seems justified, since the foremost dipterologist in this country, a man who has named over one thousand species and genera, mostly of flies, says regarding the results of certain experiments carried on with Drosophila by one of his colleagues.** But I think it is absolutely certain—and I speak as an entomologist fairly familiar with flies—that it would be impossible to produce species of his sports even though they were bred for a thousand years.*° In talking over this species question with one who has had considerable experience in systematic work* it be- came clear that although as a rule only a few of the more conspicuously contrasting characters are selected for de- scriptive purposes, as a matter of fact the individuals of different species are often different in practically every morphological characteristic. One who is very familiar with these species will realize these differences at once, although many of them are of such a nature that they can not be described so that any one else will recognize them. There seems to be no difficulty, however, in finding numerous describable contrasting characters in Droso- phila, since at least fifty-nine are included in the descrip- tions of two recently named species*! that were selected 87 Morgan, 1916, ‘‘Critique of the Theory of Evolution.’’ 38 Dr, F. E. Lutz. 89 Williston, 1908, Par Ta Vol. 42. 40 Dr. Alexander G. Rut 41 Sturtevant, 1916, py "Ent. Soc. Amer., Vol. 9. No. 612] GENESIS OF ORGANIZATION OF INSECT EGG 713 at random, D. superba and D. projectans, and these char- acters relate to almost every part of the body. Many other differences would probably also be found between the physiological processes and general activities of the adults and between the morphological and physiological characteristics of the embryos, larve and pupx of the two species if they were compared from these standpoints. It has been shown that the factor for the character club wing affects not only the character that gives this muta- tion its name, but also other characters, for example the presence or absence of a pair of spines on the sides of the thorax, these being always absent when the factor for club wing is present.42 It is possible that the combina- tion of a number of such factors as that for club wing would ultimately satisfy the requirements of systematic entomologists and that new species could then be made up in the laboratory. Such mutations might therefore be of evolutionary value. If, however, these mutations fail to furnish characters of specific rank, or characters that may lead to the formation of new species, we must con- clude that they are not of evolutionary significance, and look elsewhere for the factors that are responsible for specific characters and that may undergo changes which lead to transmutation. Factors of this sort may lie in the chromosomes or in the cytoplasm, but they are probably the results of inter- action between chromosomes and cytoplasm. As pointed out above, interaction of this sort has abundant opportu- nity to operate during the germ-cell cycle. The cytoplas- mic differentiations resulting from the metabolic processes that culminate in the formation of an egg ready to undergo maturation are very striking in the case of insects, as in- dicated by observations and experiments on the eggs of chrysomelid beetles, and there seems to be no valid reason why the eggs of these beetles are different in their type and complexity of organization, both morphological and physiological, from those of Drosophila; for while we do- 42 Morgan, Sturtevant, Muller, and Bridges, 1915, ‘‘ Mechanism of Men- delian Heredity. ’’ 714 THE AMERICAN NATURALIST ` [Von. LI not know much about the growth of the egg and embryo- logical development of this genus of flies, we do know that these processes in certain other flies resemble those of beetles. If the Mendelian factors are located in the chromo- somes, it is evident that they may exert an influence upon the entire contents of the egg, (1) during the mitotic divi- sions of the oogonia, (2) during the so-called resting stages of the oogonia, and (3) during the growth of the oocytes. It is also clear that all of the factors carried by the chromosomes have an equal opportunity to interact with the cytoplasm and not alone those that remain within the egg after the elimination of chromosomes during maturation. The adult, however, that develops from the egg, whether fertilized or unfertilized, exhibits only those detailed characteristics whose genetic factors are sup- posed to be located in the chromosomes remaining in the egg after maturation, or in those that are brought in by the sperm. This seems to indicate that none of these factors has any permanent influence upon the egg organi- zation during the growth of the oocyte and until matura- tion is completed. It seems impossible to ignore the chromosomes or even to locate the principal factors of heredity in any other cell bodies. It may therefore be necessary to reconstruct our ideas of chromosome architecture and thereby aban- don the theory that these bodies consist of a linear series of factorial determiners for certain ferments and of noth- ing else. It may be possible to separate our hypothetical factors into two groups, (1) those responsible for such characteristics as the polarity, bilaterality and ‘‘pattern”’ of the egg, and (2) those that control mutations that ap- pear at a late period in the life history like those that are so abundant in Drosophila. Perhaps the latter may be anchored to the chromosomes as has recently been sug- gested ;*8 the main portion of the chromosomes might then represent the foundation for the factors responsible for the organization of the egg and the attached masses of 43 Goldschmidt, 1917, Genetics, Vol. 2. No. 612] GENESIS OF ORGANIZATION OF INSECT EGG 1715 ferments might constitute the factors responsible for the modification of embryonic, larval and adult characters— factors such as have been employed for experimental breeding purposes by most geneticists. According to this hypothesis it would probably be necessary to consider the main portions of each chromosome as sufficient for the production of an entire organism. The fact that the group of factors carried by any one chromosome in Dro- sophila controls characters that are not restricted to any definite part of the body gives weight to this assumption. Most geneticists are accustomed to deal with adult char- acters only, and on this account pay very little or no at- tention to the eggs, embryos and larve of the species they are experimenting with. But the eggs, embryos and larve contain all the factors for these adult characters, both those that are realized and those that are inhibited either by internal or external causes, and they may ex: hibit characters that make it possible to separate different lines although the adults may be indistinguishable. Fur- thermore, taxonomists have long recognized the value of embryonic characters as an aid k determining species. We should always be careful to distinguish between the parts of the egg that are of hereditary significance and those that are not. Thus the shell or chorion of the-silk- worm egg has been discussed under the heading of ‘‘cyto- plasmic inheritance,’’ whereas it is not a vital part of the egg, but, being secreted by the epithelium of the ovarian tube, is a well-defined characteristic of the adult female and its coloration, which follows the laws of Mendelian in- heritance,** is controlled by maternal factors. Such fundamental characteristics as polarity, sym- metry, and pattern, which are so clearly exhibited by the eggs of insects and certain other animals, are much more difficult to study than adult characters and are probably not so easily modified. If any or all of them are carried over from one generation to another in the cytoplasm we have then a real instance of cytoplasmic inheritance. Even if this is the case the chromosomes doubtless exert 44 Toyama, 1913, Journ. of Genetics, Vol. 2. 716 THE AMERICAN NATURALIST [Vou. LI an influence upon the cytoplasm during the oogonial and growth periods of the egg, and a study of the genesis of cytoplasmic organization may lead to data that will help us solve this difficult problem. If the polarity of the oocyte when recognizable is not inherited, i. e., if it is not transmitted to the primordial germ cells by the egg, and retained by the oogonia, it must arise de novo saat before or during the growth period. One observer*® has found that in certain beetles the position of the spindle remains, resulting from the differential divisions that precede the formation of the oocyte, indicates the polarity of the ultimate organism, but he does not tell us how this ‘‘polarité predifferen- tielle” is brought about. In all insects the end of the egg directed toward the head of the mother becomes the an- terior end of the offspring. This is also the pole of the egg lying next to the nurse cells or that is closest to the nurse-cell chamber. This relation between oocyte and nurse-cells may be the determining factor in the polarity of these eggs and, if so, would indicate that polarity here is due to environment. How this relation could influence the polarity may be explained by means of axial gradi- ents of metabolism, such gradients in this case being produced by greater external stimulation at the end near the nurse-cell chamber where nutritive substances are elaborated and added to the oocyte. By this theory of metabolic gradients, differentiation along an antero- posterior axis can be accounted for and further differen- tiations of a morphological and physiological nature would result from ‘‘chemical transportative correlation between the different parts.’’*® We should not lose sight of the fact, however, that these hypothecated physiological activities require protoplasm as a material basis and that their results depend upon the character of this protoplasm. If polarity is established at the stage suggested above, it follows a long series of nucleo-cytoplasmic interactions which have no doubt re- sulted in the differentiation and localization of numerous 45 Govaerts, 1913, Arch. Biol., Tome 28. 46 Child, 1916, Science, Vol. 43. No. 612] GENESIS OF ORGANIZATION OF INSECT EGG 717 kinds of cytoplasm. The appearance of a definite polar- ity might lead in some way to diffusion processes and the circulation of secretions resulting in further specializa- tions and localizations. One stage seems to initiate the next stage in the series of processes that accompany the visible changes in the growth and development of the egg, and the character of these processes is of course due to the specificity of the protoplasm. That the cytoplasm may exert a controlling influence upon the chromatin has been demonstrated in several in- stances. For example, we know that the chromatin- diminution processes during the early cleavage of both Ascaris and Miastor are controlled by the cytoplasm and that in these animals the germ-cell nuclei retain the full amount of chromatin because of the germ-cell cytoplasm they chance to encounter. Probably the peculiar distribution of the chromosomes at certain stages in the life histories of certain aphids, phyloxerans, and Hymenoptera is also controlled by the eytoplasm. In the aphid, Aphis saliceti, the first matura- tion division is visibly differential both as regards the chromosomes and the cytoplasm.“ The mitochondria congregate at one end of the dividing spermatocyte; this process is accompanied by a greater accumulation of cytoplasm at this end so that cell division results in one large cell containing all of the mitochondria and about two thirds of the cytoplasm, and one functionless small cell. The large cell also receives three chromosomes; the small cell only two. The peculiar maturation divisions in the males of the honeybee*’ and hornet,*® during which one ultimate sper- matogonium gives rise to only one spermatozoon instead of the usual four, may also be the result of cytoplasmic control. The cytoplasm may likewise be responsible for _the passing of a sex chromosome into the polar body dur- ing the maturation of the egg of certain aphids at the end of the summer season.” Such eggs must be fertilized ` 47 Baehr, v, 1909, Arch. fiir Zellf., Bd. 3. 48 Meves, 1907, Arch. mikr. Anat., Bd. 70. 49 Meves and Duesberg, 1908, Arch. mikr. Anat., Bd. 71. 50 Morgan, 1909, Journ. Exp. Zool., Vol. 7 . 718 THE AMERICAN NATURALIST [Vou. LI before they will develop, and always produce males. Many other peculiarities in the behavior of chromosomes that have been reported from time to time may also be due to the influence of the environment (cytoplasm), and there seems to be no reason why factors carried by the chromosomes should not be affected by the cytoplasm as well as are entire chromosomes. By the interaction of Mendelian factors with the cyto- plasm during the germ-cell cycle, it is even possible to explain the fact that ‘‘crossing over” occurs in the females of Drosophila, but not in the males.** In the latter, the spermatocytes do not pass through a pro- nounced growth period, and hence there is comparatively little nucleo-cytoplasmic interaction, and since the cyto- plasm carried by the sperm may be considered negligible, the factors borne by its chromosomes are not interfered with. In the female, however, there is ample opportunity for such interaction during the growth period, and factors at this time may be influenced by the cytoplasm or may influence the cytoplasm in such a way as to cause an ir- regular distribution of chromosomal factors. To the writer the following conclusions seem justified. The insect egg at the time of maturation is a mosaic of differentiated cytoplasmic areas predetermined to de- velop into definite parts of the embryo. This organiza- tion has resulted from the interaction of nucleus and cytoplasm during the germ-cell cycle. Such interaction * is taking place at all times, but is visible only when such processes as the protrusion of chromidia or chromatin- diminution occur. The many cases of cytoplasmic con- trol over chromatin behavior, and the apparent failure of the factors for the characters commonly used by geneti- cists to influence the egg organization, indicate the im- portance of more careful studies of the genesis of this organization. The importance of such studies is empha-. sized by the possibility that they may help toward the solution of the problem of the method of evolution. 51 Morgan, Sturtevant, Muller, and Bridges, 1915, ‘‘Mechanism of Men- delian Heredity.’? NEW FACTS AND VIEWS CONCERNING THE OCCURRENCE OF A SEXUAL PROCESS IN THE MYXOSPORIDIAN LIFE CYCLE. DR. RHODA ERDMANN OSBORN ZOOLOGICAL LABORATORY, YALE UNIVERSITY AND ROCKEFELLER INSTITUTE FOR MEDICAL RESEARCH, DEPARTMENT OF ANIMAL PATHOLOGY Tar classic observations of Balbiani, Bütschli and Thélohan on the myxosporidian development do not in- clude the occurrence of a sexual process which compre- hends the forming of a syncaryon in the life history of this protozoan group. Doflein (1898 and 1901) suggested two places in the life cycle of the myxosporidian where a caryogamy might probably take place. In the next period of investigations on myxosporidia the occurrence of this sexual process was stated by various authors, but they differ widely in the conception of the place in the life cycle in which the copulation occurs. Mercier, Awerinzew, 1 For a clear understanding of the question under discussion, it is neces- have a uniform nomenclature and to discard all terms which are only of Haese value. Noyaux du sporoplasma, noyaux du germe, noyaux sporoplasmiques, germnuclei, Amöboidkeimkerne were to be dis- carded and gametonuclei, noyaux des gamétes, Gametenkerne were to be sed. Instead of sporoplasma, sporoplasm or améboidkeim, only such ex- ressions as gametes, gametes and Gameten are admissible. Identical and adequate terms are capsulogenous cell, cellule en pS bar aut Valve cells, cellules valvaires and Schalenzelle sho r those cells which form the membrane of each single spore; cellule Venveoppe, envelope cells, and Hiillzellen should be used rage those cells which form brane of the fine ert In eases where only the nuclei of Pa cells are present terms noyaux d’enveloppe, envelope nuclei, Hiillenzellen- kerne agni $ substituted. If the cellular origin of the pansporoblast mem brane is not ascertained, envelope, membrane d “envelop ppe or Hülle ma be used. The terms Restkern, residual nucleus are eading. Somatic residual nuclei or somatische Restkerne should be used, if the definition on p. 679 holds true. Here a new name would be of great value eliminate the wrong analogies created by Doflein (1898, p. 309). The term ‘‘ reduce- tion aioe is only justified if the healers reduction of id cect has been ascertained. ; 719 720 THE AMERICAN: NATURALIST [Vou. LI Auerbach and Parisi try to show that a real syncaryon formation takes place at the onset of spore formation. Other authors, Keysselitz, Schroeder and Auerbach, be- lieve that only a plasmogamy can be pointed out at the beginning of spore formation, and that the union of the nuclei is effected either in the fully developed spore or in the young animal leaving the spore. The difference between these last two conceptions is theoretically with- out significance because the main part of copulation—the union of the nuclei—takes place at the onset of the new life cycle of the myxosporidian. Therefore it was of the utmost importance for decisive proof of this fact to find the copulation of the two gametonuclei inside the fully developed spore or in the young myxosporidian. Schroeder, 1909, observed the copulation of the two game- tonuclei in the spore; Auerbach, 1907 and 1910, found young animals of Myxidium bergense with one nucleus. I was able to demonstrate young Chloromyxum leydigi which were experimentally produced by placing the two- nucleated spores on gall plates (Erdmann, 1911). Here, after a treatment with intestinal secretions of the host the young animals leave the spore. They are at first bi- nucleate, later uninucleate. In my recent work, finished in 1913, which did not appear until 1917 in consequence of the war, I figured these young animals after fixation and staining. Also, Davis, 1915, though with some re- serve, presents young Spherospora dimorpha which have left the spore and show the fusion of their two nuclei. Later the separation of the syncaryon into its vegetative and generative components takes place. Georgevitch, 1914, presents the development of the young animal in Henneguya gigantea in—as it seems—changed pecu- liar conditions. The spore is still inside of the cyst of the big ‘‘tumor-forming’’ tissue-parasite. The binucleate form becomes uninuclear and then the usual vegetative multiplication of the nuclei begins, which leads up to a renewed spore formation inside of the tumor cyst. In Chloromyxum leydigi, a gall-bladder parasite, no such No. 612] MYXOSPORIDIAN LIFE CYCLE 121 complicated process takes place. As the young uninucle- ated forms develop we see an animal with three nuclei, all of the same size. Later multiplication of these vegetative nuclei and the formation of big syncytial masses occur. The plasmatic bodies of these vegetative animals contain two kinds of round corpuscles, ‘‘Reservekérper’’ and ‘‘Warbtrager.’’ In my publication in the Archw für Pro- tistenkunde, 1917, I give proof that the ‘‘Reservekérper’’ consists mostly of glycogen and I point out that the gly- cogenous contents are used up during spore formation. The vegetative animal can multiply either by division or by forming small vegetative gemmules (Erdmann, 1911). The fact that inside the animal vegetative propagative bodies can arise, was verified by Davis, 1915, pp. 354-355, in Spherospora dimorpha. Before the onset of spore formation a differentiation in the syncytial masses of Chloromyxum leydigi begins. We can distinguish parts in which the nuclei multiply and other parts where only the vegetative nuclei are seen widely scattered in the protoplasm. I called the first- mentioned areas ‘‘islands’’ (Erdmann, 1911) because in the living animal they rise above the surface of the vege- tative plasmatic body. They are distinguished by their pale color and in stained preparations by their large num- ber of small nuclei. At first all the nuclei in these islands are of the same size. Two nuclei with small cytoplasmic bodies approach each other and each cell divides up into a small and a big cell. The two small cells draw out in length and surround the two big ones, in this manner separating them from the other cells in the island. This quadruple group, two big cells and two small ones, is the starting point for the formation of the whole spore. The two big cells are gametocytes. These two gametocytes divide and form two gametes and two other cells which after a further division give rise to four cells—these four cells are the four capsulogenous cells. The whole spore contains, therefore, eight cells—four capsulogenous cells, two gametes and two cells which form the spore mem- brane. (Fig. 1.) 722 THE AMERICAN NATURALIST [Vor. LI I mentioned before that the glycogen which was found in the vegetative body is used up in spore formation. The membrane of the spore, the polar threads and the darkly staining structureless lumps, which have been seen by all authors inside the sporoblast, consist of glycogen and stain as well by chromatin as by specific glycogen stains. These lumps have been 0, 0. considered as ‘‘reduction nuclei’’ a o by various authors. They have A2 also been called ‘‘Restkerne’’ or ‘‘residual nuclei.” It may be emphasized for later discussion that they are glycogenous and 0 0 not chromatic in Chloromyxum On O On Op rris. Our knowledge of the sexual process in the myxosporidian life 0.0.0, 0... cycle since the investigations of See one 5s Keysselitz, Schroeder, and others le he or yrum tey: digi: Az and Bz are the spore has been completed by recent ar ee oe Eee authors: Georgevitch, 1914, ut ire ag nn ar dena Davis, 1915 and Mavor, 1916. : The best proof that no reduction takes place in the spore is given by Georgevitch (1914) and Davis (1915), who have been able to study the num- ber of chromosomes in their specimens before and after spore formation. Both authors agree that the number of chromosomes is not changed before and after this phe- nomenon. Davis figures six chromosomes in Spherospora dimorpha and Georgevitch, investigating Henneguya gigantea, finds eight. These two facts: first, that the number of chromosomes is not changed in spore forma- tion and, second, that the so-called “reduction nuclei”? inside the spore are really glycogenous bodies which are used up in forming the membrane of the spore, polar threads, and spore membrane of Chloromyxum leydigi, positively prove that the real reduction must occur at a different place in the life cycle of Spherospora, Chloro- No. 612] MYXOSPORIDIAN LIFE CYCLE 723 myxum, and Henneguya. The reduction consisting in the transformation of a diploid nucleus into a haploid or of a tetraploid one into a diploid can only occur at the begin- ning of the new life cycle after or before the union of the gametonuclei. The further facts which Davis, 1915, presents in Sphe- rospora dimorpha also tend to show that reduction could only take place at the above outlined place. The gametonuclei of this myxosporidian fuse together after it leaves the spore and form by one subsequent division two nuclei. One of these nuclei gives rise to all the cells of the later sporogenous body. The other of these two nuclei, distinguished by its size and structure, is the vegetative nucleus of the animal, the somatic ‘‘Rest- kern.” No other nuclei should be called ‘‘ Restkerne’’ except when they represent the nucleus or nuclei of the vegetative myxosporidian body, which does not play a part in spore formation. Such ‘‘Restkern’’ or ‘‘Rest- TABLE I oTilizelien e Occurrence of Envelope Saes Celis. Bomatio Residual | Cells, Pan- They Ave Author Species Nuclei in Sensu- sporoblast- Division ta ranne a| Lotaria ot | A R Loe Only Ti, OMmMOCrtee Nuclei 1, Awerinzew.| Myzxidium sp. One — 2 for each spore 2. ae aig Sere. None (p. 26) — of Birisi spora dimorpha One, settee two — j 4. Awerhione. P, E Art drepa Two — ie 5. Mayor...i: Ceratomyxa acad Two — g 6. Erdmann...| Chloromyxum leydigi Many — + Fa: Darit Spheerospora d ha, | Many — r polysporous form 8. Auerbach...| Myxidium bergense, None = s polysporous form ene yaa ch ha sacred in ono-,. di- an | BdBiled delaby ec orms only the propogative parts of the com- lete animal 9. Parisi........, Sphaerospora cavdata | No facts men-| Two y tioned 724 THE AMERICAN NATURALIST [Von LI kerne,’’ the origin and fate of which agree with this defini- tion, have been described by Awerinzew in Ceratomyxa drepanopsette, and in Myxidium sp., by Davis in Sphero- spora caudata, and by Mavor in Ceratomyxa acadiensis. In the disporous form, Spherospora dimorpha, two spores are found in the whole animal and the sporogenous body finally contains twelve cells—half of this number forms one spore (Fig. 2). These twelve cells are all 0,, 0; “a2 ©B2 0. 0,2 Osr 0312 a2 B21 On 2 Alt2 0: 0.4 Gi. Mogae 2.. Spherospora dimorpha (disporous form): An As and Ba Bz are the membrane-forming cells for each spore. Ay A12; Buz Bı are the two etes in each spore. A,,, Ay; Bin Bin are the two capsulogenous cells in each spore. division products of the one nucleus, the sister nucleus of which is the somatic ‘‘Restkern.’’ The cells which form the spore membrane of each spore, lying independently in the vegetative body, arise by a very late division. Chro- matic lumps which could be considered as ‘‘reduction nuclei’’ are lacking. It is easy to imagine why they are absent. Spherospora dimorpha lives in the urinary blad- der and has therefore a metabolism which may not afford opportunity for an abundant glycogen formation. Also, as mentioned before, no reduction of the chromosomes takes place. These facts could be easily ascertained in consequence of the large size of the cells and chromo- somes. Awerinzew’s description of spore formation in Ceratomyxa drepanopsette and Mavor’s of Ceratomyxa acadiensis have many features in common, but Awerin- No. 612] MYXOSPORIDIAN LIFE CYCLE 725 zew’s presentation differs from that of the other author in one important point. The cells A and B in Fig. 2 are said to be in Ceratomyxa drepanopsette the prod- ucts of a fusion of two cells. The reader has but to change the lettering of A in AC and of B in BC to see that the two*copule undergo changes identical with the two single cells in Spherospora dimorpha. Except this one point of difference—the beginning of spore formation—we note that the late division of the membrane-forming cell, the identical number of cells in each spore and the independ- ence of each spore in the myxosporidian body characterize both species of Ceratomyxa (Awerinzew and Mavor) and Spherospora (Davis). It appears highly improbable that in two different species of Ceratomyxide the basis of spore formation should be a copula (Awerinzew) or an univalent nucleus (Mavor). Summarizing the known description of spore formation of those myxosporidia, in which each spore is formed in- dependently of the other in the somatic body, and where no pansporoblast exists, we can demonstrate the following uniform features in all investigated forms. 3 1. Six cells or nuclei are used for the formation of each spore, when two polar capsules are present; eight, when four polar capsules are present. 2. The cells which form the spore membrane have a similar origin and are distinguished by the independence in which these cells develop as compared with the other constituents of the spore. Their mother cell is lying in a resting stage till the division of the gametocyte is finished, as described by Davis for Spherospora dimorpha, and by Awerinzew for Ceratomyxa drepanopsette. In Myzidium sp., where, according to the investiga- tions of Awerinzew, either one, two or three spores are lying independently in the myxosporidian body, there is a very late division of the cell, the divisional products of which form the spore membrane as recorded by this author (Fig. 3). Here the one gametocyte divides into two cells, one of which by a late division gives rise to the two- 726 THE AMERICAN NATURALIST [Vou. LI spore membrane-forming cells, the other forms the two capsulogenous cells and two gametes. We intentionally avoid, in Fig. 3, calling 0, the two chromatin ap lumps, which the author himself (p: 202) ° has 4] called ‘‘iiberfliissiges 0. A12 Chromatin”? (degenerie- o é rende Kerne), nuclei Az. A22 and do not add them 6 0 0 as such in the diagram Ati “M22 as Auerbach, 1912, p. re the spore ibrao Feria colis 28, did when discussing (valve cells). Am Ax are the gametes. Awerinzew’s investiga- Anz and Ay: are the capsulogenous cells. ; S ions. We consider these chromatin lumps only as formations probably of glycogenous nature and as being used during membrane formation. In Fig. 4 Auerbach’s conception (Type I) of how the spore formation is effected in Myxidium bergense is represented. Auerbach believes, as stated before, that either a plasmogamy (Type I) or a real copulation (Type Il) may be at the basis of spore formation. We BI will not discuss, for the 0, present, how the bigger and smaller cells which are seen at the begin- 0. 0 ning of spore forma- tion, arise. The latter an “Bi2 divides once and the 0 0 0 0 two divisional products Am At A122 Mysidium bergense (Auer- form the spore-mem- Fic, 4. n bach Type I); By and By are the spore brane-forming cells. The membrane-forming cells (valve cells). Anı other cell divides twice "4 A are the two gametes. Aus and ie Aim are the capsulogenous cells. to give rise to two gam- etes and two capsulogenous cells. The author does not especially mention the order in which these cells divide, No. 612] MYXOSPORIDIAN LIFE CYCLE 727 but nothing in his report contradicts the supposition that in Myxidium sp. and in Myxidium bergense there is a close analogy with Davis’s and Awerinzew’s observa- tions. He also observes the elimination of chromatin and ‘‘eventuell Bildung von Restkernen”’ (p. 24). These are not incorporated in Fig. 4 for the same reasons I pointed out for Myxidium sp. My observations in Chloromyxum leydigi show, further- more, that the cells which form the valves of the spore do not play a part in the development of the final contents of the spore, but are here in this form (comp. Fig. 1) the products of the first division of each gametocyte. In the case of Chloromyxum leydigi two gametocytes form by one division the two spore-membrane-forming cells; in the three other species, Myxidium sp. (probably Myxid- tum bergense), Ceratomyxa drepanopsette and Sphe- rospora dimorpha, one gametocyte forms the one cell, the division products of which are transformed into the valves of the spore. But in all four species these cells have the sole purpose of forming the spore membrane. After surveying the Figs. 1, 2 and 3, and having com- pared them, I have no doubt that the origin of the spore-membrane-forming cells is identical in the so-called monosporous and disporous forms. Advancing one step farther and taking into consideration those forms in which two gametocytes form the cells inside each spore (Fig. 1) we notice that in dealing with the origin and the position in the development of the spore, we have to add nothing. The spore-membrane-forming cells are distinguished by their early segregation from the game- tocytes and their non-entering into the series of those cells which are included in the spore. The only difference is that these cells do not divide further; if they did, we could easily construct the disporous type of Spherospora dimorpha (Fig. 2). The same holds true for those species which form two spores in one pansporoblast (Fig. 4) and where two gametocytes are observed at the basis of spore formation (Keysselitz, Schroeder). 728 THE AMERICAN NATURALIST [Vor. LI If we conceive these two cells in question (A, and B,) which form in Myxobolus pfeifferi (Keysselitz) the pan- sporoblast membrane to divide once more and the last di- vision inside the spore to be suppressed, we could have the type of Spherospora dimorpha. According to Keysselitz, in Myxobolus pfeifferi each of the two gametocytes together with the small cell ap- 0... 0,. Opn 0, qo oO. On 0.0, Oe Oa &. O, Oraa OP ten G. 5. General plan for polysporous disporoblastic forms. (Type ses Ag and Bz are the envelope-forming cells (pansporoblast- a Peon cells). An and A, and Bẹ» are the two gametes in each spore. The products of “the fourth division form valve tells Ann, Aisi, iha Bing, and cap- sulogenous cells Ajj, Aim Buw Biar ` proach each other and each divides up until six cells have arisen. Thus we have in all cases 14 cells, two of which have a different divisional capacity, for they stop dividing and the big cells form all the other cells which at the end compose the two spores. Their later fate is indicated in Fig. 5 and, though this does not. concern us in this discussion, we should like to emphasize the fact that two gametes are always present in a certain stage of spore development. We are convinced that the cells A, and B, represent genetically in the pansporoblastic forms that cell formation which in all monosporous, disporous and polysporous species gives rise to the spore membrane itself (Figs. 1 to 4). These cells or their nuclei were observed by Keysselitz, Parisi, Lo Gindice, Auerbach (Henneguya psorospermica), Georgevitch, and with cer- No. 612] MYXOSPORIDIAN LIFE CYCLE 729 tain restrictions by Mercier and Schroeder. This cell couple (A, and B,) should be called envelope cells or envelope cells nuclei when they do not fulfill their des- tin (Schroeder, probably Mercier). Even in those cases in which a pansporoblast membrane had not been dis- covered it might have been either overlooked or have -been in evidence at the. beginning of pansporoblast formation before the valves of each single spore had been developed. Later these take up the function of the Hiillzellen which make their retrogressive development plausible. These Hiillzellenkerne are neither Restkerne, nor reduction nuclei, nor somative residual nuclei. The term for residual nuclei of somatic nature (see definition, p. 679) has already been disposed of in monosporous and disporous forms and must be used in the same way in polysporous forms. In all forms which are polysporous and have many singly developing spores, the whole vege- tative body which is not used up in spore formation has somatic ‘‘restkerne’’ or residual nuclei. As I pointed out in Chloromyxum leydigi the vegetative animal may die after spore formation together with the ‘‘restkerne.’’ In this form the vegetative animal may prolong its life by forming internal buds, if it has reached a considerable size before and during spore formation. In all poly- sporous forms with pansporoblast, i. e., the disporoblastic forms, we have to be very careful when applying the name of somatic residual nuclei. In those species which are not tissue parasites and sometimes have cystlike formations which are surrounded by gelatinous envelopes, we may find somatic residual nuclei, because it seems improbable that the whole vegetative body is used up for spore for- mation. I believe this to be the case in Spheromyxum sabrazesi and Spherospora caudata. Where no residual vegetative nuclei were observed, the investigators may not have studied the whole animal, but only the propa- gative parts of it which have left the vegetative body (Parisi, Fig. 3). Keeping this point in mind, later in- vestigators may discover somatic residual nuclei in de- 730 THE AMERICAN NATURALIST [Vou. LI generating stages analogous to those found in the mono- sporous, disporous and the non-disporoblastic poly- sporous forms, among the debris of the dying animals inside the gall and urinary bladder, as I have shown in Chloromyxum leydigi. In the so-called ‘‘tumor-forming’’ disporoblastic polysporous forms, no facts are known - which show that somatic residual nuclei have been ob-. served. The beginnings of cyst formation, however, have never been studied, and it is only at this stage that somatic residual nuclei may be seen, and not after the eyst is crowded with sporoblasts and spores. But even in the fully developed cysts there may be degenerating somatic residual nuclei which have escaped observation. The facts which Weissenberg found in Glugea anomala and hertwigi—two Microsporidia—seem to support my suggestion. But the Hiillzellen or Hiillzellenkerne of the myxosporidia are never identical with somatic nuclei. Their undisputed place in the development of the myxo- sporidia will soon be clear. Before we proceed further in discussing the formation of the sporoblast membrane in those myxosporidian species in which the spores are formed in pairs inside one sporoblast, it may be recalled that several facts have been ascertained concerning the fully discussed spore-mem- brane formation in monosporous or disporous myxospo- ridian species: (1) The copulation of two gametes occurs during or after the two myxosporidia leave the spore. (2) No reduction takes place from the beginning of spore formation until the end, because the number of chromo- somes remains the same (Davis and Georgevitch). (3) The darkly staining masses of ‘‘restkerne,’’ ‘‘residual nuclei,” ‘‘reduction nuclei’? have been shown to be gly- cogenous and to be necessary for spore-membrane for- mation. (4) Membrane-forming cells or nuclei are set apart by different division intervals from the other cells of the sporoblast. Now from the above given summary of the latest facts in monosporous or disporous forms, it is clear that they No. 612] MYXOSPORIDIAN LIFE CYCLE 731 are strictly opposed to all the views which maintain that a copulation or so-called syncaryon formation precedes spore formation. But they are all in accord with the in- vestigations of all authors who have shown that there is no syncaryon formation, but merely a plasmogamy of two cells, without any copulation, at the onset of spore formation. When I wrote the second part of my investigations on Chloromyxum leydigi, 1913, I pointed out that the facts which were presented by Auerbach, Mercier and Parisi, as proofs of the occurrence of a synecaryon formation just before the onset of spore formation, are not quite con- vineing. Their figures can easily be arranged in such a manner that the supposed synearyon formation represents the division of gametocytes into a smaller cell, which in most all other known cases forms the membrane of the pansporoblast. (Compare Erdmann, 1917.) Itis not nec- essary to repeat here the attempted revision and rear- rangement of the figures of these authors. This same holds true for the synearyon formation in Myxidium ber- gense, Type II (Auerbach) and Ceratomyxa drepano- psette (Awerinzew). We will take it for granted that our views are correct as long as no new facts ascertained on smears—not sections—compel us to change our opinion. As mentioned above, all authors who have shown that no synearyon formation occurs, but that a plasmogamy of two cells without any nuclear fusion occurs at the onset of spore formation, can agree with us that the sexual proc- ess is going on at the beginning of the new life cycle. Auerbach and Parisi do not convince us that the figures which represent the so-called caryogamy can not be con- sidered as the dividing of the gametocyte in the two cells. In accordance with the facts and interpretations, Keys- selitz, Schroeder, Lo Giudice, Erdmann, and Georgevitch uphold the view that no merging of two cells or two cell pairs takes place to form the couples of cells which are later considerd as a quadruple group of the growing spore. By comparing the series of figures of all those drawings 732 THE AMERICAN NATURALIST [Vou. LI which are supposed to prove the merging of two cells, they can, as said before, be interpreted as the division of one cell into two. The larger of these cells, wrongly called macrogametocyfe according to the cell fusion theory, has divided and formed the cell wrongly called microgametocyte. This ‘‘microgametocyte’’ and its division products after one division, or these ‘‘microgametocytes’’ in cases where two gametocytes are observed at the onset of spore for- mation (Keysselitz, Schroeder, and Erdmann) now form, according to all known investigations, the pansporoblast membrane. Just as we could point out in disporous forms the uniformity of the origin of the spore-mem- _ brane forming cells (Figs. 2, 3 and 4) so we can do the same for the pansporoblast membrane and its nuclei in the following forms: Myxobolus pfeifferi (Keysselitz), Myxobolus ellipsoides (Lo Giudice), Spheromyxa sa- brazesi (Schroeder), Spherospora caudata (Parisi), Hen- neguya gigantea (Georgevitch), and Henneguya psoro- spermica (Auerbach); all following Fig. 5, provided we do not take into consideration the origin of the cells A, A, and B and B,. Keysselitz and Schroeder’s views, except one contradictory point, are exactly repre- sented by Fig. 5, but there are differences mentioned by the other authors. Still we make the generalization that there is one and the same plan of spore formation in the pansporoblastic myxosporidia though we know that facts are reported which do not fit in with our view. We hold the opinion that it is permissible to rearrange the observed facts, because all interpretations have been gained by piecing together and arranging facts according to the theoretical viewpoint of the authors. No continued observation of spore formation in the living animal has been possible. Also we are allowed to add facts ascer- tained in other species if the authors have only consid- ered sections and not smears. Sections are misleading because the whole quadruple group can not always be seen ~ on the same section and the origin of the small cell from No. 612] MYXOSPORIDIAN LIFE CYCLE 733 the big cell can not be traced without doubt. Therefore, most investigators have lately used smears to get a fuller and more correct view of the origin of the different cells from each other. It is astonishing how scanty the details appear when one considers the formation of the quad- ruple group in Auerbach’s, Lo Giudice’s, Parisi’s and Georgevitch’s presentations. Mercier’s Figs. 19-27, Plate 1; Auerbach’s Figs. 8a-15, Plate 2; Lo Giudice’s Figs. 29-34, Plate 1; Parisi’s Figs. 13-18, Plate 16; and George- vitch’s Figs. 32-35, Plate 1, do not show each single step of this important process. Connecting stages are missing. Therefore, one is allowed to interpret differently their presented facts, as I have done in Fig. 5. In Table II TABLE II I II Myxobolus pfeiffert Mercier. Spheromyxum sabrazesi Schroeder. - Cell A (copula) forms all other 12 As B, cells of the pansporoblast and the A, B, two ‘‘ Hiillzellenkerne.’’ All cells divide up to form the 12 pansporoblastiec cells and the two tt Hiillzellenkerne. ’’ Myzobolus ellipsoides Henneguya gigantea Henneguya psorospermica Lo Guidice. Georgevitch. Auerbach (Type I). A, An A, Ay Cells A, and A» do not form any of the other 12 cells of the pansporoblast. we can study the different opinions from the authors’ point of view. Lo Giudice, Georgevitch, Auerbach, Keys- selitz, and Schroeder are alike in interpreting that the two cells which develop into the pansporoblastic mem- brane or nuclei are separated very early from the other cells. They never intermingle with those cells inside the spore-membrane (except according to Schroeder). They can not, therefore, be microgametocytes and in conse- quence they have nothing whatsoever to do with a sexual phenomenon. This adds strong support to our view that 734 THE AMERICAN NATURALIST [Vou. LI the sexual process is at the beginning of the new life cycle. I do not wish to veil the great discrepancy between the conception of Mercier (Fig. 6) and the other authors 0 0., 0., On 0. Ou Omn Ore Ooa Oe Oon Oze On Ve OPiMateA Ben” tat clad bane 6. Myzxobolus pfeifferi, Mercier: Oi, and Coe are We cellos d’en- reloppe "oe = author, which represent the ja ii “ Restkerne ” or tion nuclei” of other authors. The products of the fourth snvidie form, oe each aes, the gametes, the valve cells and the capsulogenous cells. mentioned. All cells are products of a copula; and there is no setting apart of the g form- ing cells or nuclei, though later they appear at the accus- tomed places between the two spores (Plate I, Figs. 31, 32). I shall not risk an interpretation, but think a new investigation on this same.subect might be very promising and result in the desired uniformity. I think it highly probable that in all pansporoblastic forms the spore de- velopment follows the Keysselitz-Schroeder interpreta- tion: that two gametocytes form the basis of the spore formation. . But even if one believes that the quadruple group is not formed by two but by one cell pair, the prin- cipal point is not changed. It is indifferent for the the- oretical interpretation whether a segregation of the sec- ond cell pair from the first takes place, and both then form the quadruple group, or whether two cell pairs approach each other and form the quadruple group. The pedog- amy is merely a closer one in the first case. To summarize: In the observed myxosporidian species No. 612] MYXOSPORIDIAN LIFE, CYCLE 735 with pansporoblast (exception, Mercier, Myxobolus pfeif- feri), the first division products of the gametocyte or gametocytes form the pansporoblastic membrane or, if degenerated, its nuclei. This division is a heteropole di- vision and forms highly chromatic small cells or nuclei which never intermingle with the cells inside the pan- sporoblastic membrane (exception, Schroeder). On the basis of these facts, we need only state that the heteropole division has no connection whatsoever with a reduction division, as Keysselitz tentatively suggested. This conspicuous division produces the Hiillzellen or Hiillzellenkerne, and it may not be impossible that in the case of Spheromyxa sabrazesi these small chromatin cells do not intermingle with the others and divide up, though the author mentions it on page 366. They originate, ac- cording to Schroeder’s second interpretation, in the same manner as most authors describe, but divide together with the big cells until twelve cells are present in the pansporo- blast; they then take their accustomed place inside the sporoblast membrane and are easily recognizable. This apparent exception merits further investigation. We maintain our conclusion that the Hiillzellenkerne or the Hiillzellen are identical with the spore membrane forming cells of the non-disporoblastic polysporous forms. They have the following features in common: they are the first division products of the gametocytes; they do not intermingle with the other cells inside the spore; they form the envelope, in one case of the spore, in the other of the pansporoblast. They are neither somatic residual nuclei nor Restkerne nor reduction nuclei. They are cells which have a tendency to degenerate in some disporo- blastic forms when their functions are taken up at an early period by the valve cells or spore-membrane-form- ing cells. It remains, as the last part of our discussion, to deal fully with the significance of those darkly staining masses which have been described as ‘‘Reductionskerne’’ or ‘*Restkerne’’ ‘‘inside the sporoblast.’’ In the following table we give a short survey of the known facts. 736 THE AMERICAN NATURALIST [Vou. LI A survey of Table III brings out clearly certain facts. When darkly Author 1. Awerinzew .. 2. Auerbach ... 4, Awerinzew .. By ever. i543. e 6. Erdmann .... Ty PRED bis ee K 9. Auerbach . staining masses are observed inside the TABLE III Occurrence of Chromatic Bodies In - side the Spore Interpreted as Species uction .Myzxiđdium spec. ........... Seldom distinct small nuclei, generally ‘‘Zwei ziemlich grosse Chromatinkiigel- chen’’. (p. 201). . Myxidium bergense ....... Diffusion of chromatin or formation of two ‘‘rest- kernartigen Gebilden’’ : (p. 20). Spherospora dimorpha ,,..Formation of round chro- matie bodies. .Ceratomyxa drepanopsette.. Infiltration of chromatic small hodies into the cyto- plasm before the spore membrane includes the gametocytes after the sup- posed copulation. Note here the later Bigi of spore membra Ceratomyxa acadiensis .... Formation of coina chro- 10) which are later re- . Chloromyxum leydigi ...... Poriation of several large, i de eeply staining bodie ha disappear after a ore membrane is formed. .Spherospora caudata ..... ‘Formation of small, deeply bodies before the supposed copulation (Fig. 15). Spherospora dimorpha, polysporous form ....... No diffused infiltration of chromatin observed, also no formation of round chromatic bodies. ..Myxidium bergense, polysporous form ....... Formation of two ‘‘restkern artigen Gebilden’’ or dif- fusion of chromatin. | No. 612] MYXOSPORIDIAN LIFE CYCLE 737 pt © . Keysselitz ... Myxobolus pfeifferi ....... One to four round chro- H ee ): « Mercier ....%. Myxobolus pfeifferi ....... Diffusion of small chromatic bodies into the cytoplasm (Figs. 33, 34) after the m bo . 22, 2 . Lo Guidice ... Wyzobolus ellipsoides ..... Several round chromatic, deeply staining bodies, which are not observed after Lae Anenii ae : is formed (Figs. 42). - Schroeder ...-§nhwromyxum sabrazesi ... Chromatic, deeply aan pi w brane is formed. (Comp. Figs. 30, 32 with Figs. 33, 34. . Auerbach .... Henneguya psorospermica .. foul i To judge after Figs. 6 to 18, extrusion of a chro- matic body in cytoplasm. (Note, only sections to fuel oO . Georgevitch ..Henneguya gigantica ...... Four deeply staining chro- matic bodies called by the author degenerated nuclei. spore, they disappear after the spore membrane is formed. It is proved that in Chloromyxum leydigi they are of glycogenous nature as well as the spore membrane itself and the polar threads. In some cases their number is ir- regular. These chromatic lumps may be products of nuclear division, but the true chromosomes have not been found. Those authors (Mercier, Awerinzew, Parisi) who believe they have shown a synearyon forming, have also observed an extrusion of chromatin immediately after the union of the supposed micro- and macro-gametocyte. In Mercier’s case a second diffusion of round bodies is shown inside the spore which corresponds with the facts ob- served in other species. Ceratomyxa drepanopsette 738 THE AMERICAN NATURALIST [Vou. LI (Awerinzew) has only an extrusion of chromatin before the synearyon formation, while in Myxobolus ellipsoides (Parisi) it occurs immediately after this phenomenon. These exceptions in the series, t. e., that inside the spore no chromatin diffusion is observed, may be due in the case of Parisi to a limited number of studied forms and in the case of Awerinzew to the fact that the spore membrane in Ceratomyxa is formed very late. Yet these exceptions do not prevent the final statement that the darkly staining chromatic masses in the spore are not reduction nuclei, or restkernartige Gebilde, but play an important part in the development of the spore membrane. The whole trend of our critical review leads up to the following conclusions: 1. Reduction in myxosporidia has thus far not been dis- covered. 2. The so-called reduction nuclei inside the spore are chromatic or glycogenous masses, which serve the spore- membrane formation. 3. The so-called residual nuclei of the disporoblastic forms can not be considered as identical with the somatic residual nuclei of the mono-, di- or poly-sporous non- disporoblastic forms. They are the functionless nuclei of the envelope cells of the disporoblastic forms. 4. The envelope cells can by their origin only be com- pared with those cells in the mono-, di- or polysporous nondisporoblastic forms which later give rise to the valve cells. 5. The somatic residual nuclei are well-defined in mono-, di- or poly-sporous nondisporoblastic myxosporidia. Their analogy has not thus far been discovered in disporo- blastic polysporous forms. LITERATURE (For literature before 1910 compare Auerbach, M., ‘‘ Die Cnidosporidien,’’ Leipzig, 1910). Anite M. 1912. Studien über die Myxosporidien der norwegischen Seefische und ihre Verbreitung. Zool. Jahrb. Abt. f. Syst., Vol. 34, pp. 1-50. No. 612] MYXOSPORIDIAN LIFE CYCLE 739 Davis, H. S. 1916. The Structure and Development of a Myxosporidian Parasite of the Squeteague, Cynoscion regalis. Journal of Morphology, Vol. 27, pp. od 346. Erdmann, Rh. 1911. Zur Lebensgeschichte von Chloromyxum leydigi, einer mean aie we hrs aie 4 (Teil I). Arch. f. Protistenkunde, Vol. 24, pp. 149- Erdmann, Rh. sai ene a TE und a esa zur ander ae ae (Teil II). Arch. f. Protistenkw ; Geo enik, y 1908. Sur le cycle évolutif chez a SE oie C. B. A Bo Ea T: pn p. 190. ee a "1915. Etude du cycle évolutif chez les Myxosporidies. rch d. Zool. exp. et gen., T. 54, pp —409 Lo ara P. 1911. Sullo Soe del Mynobolss ellipsoides Thel. Riv. mens. Pesca Idrobiologia, Anno 6 (13). Lo Giudice, P. 1912. Studi sui Cnidosporidi. Pavia, pp. 1-88. Mavor, I. W. 1916. On the Life History of PP TEER acadiensis, a New Species of Myxosporidia from the Eastern Coast of Canada. Contrib. Zool. Lab. of the Museum of Compara. Zoology at Harvard 78. Nemeczek, A. 1911. Beiträge zur Kenntnis der Myxo und Microsporidien der Fische. Archiv. f. ieia Bd. 22, pp. 143-169. Parisi, B. 1910. Spherospora caudata n. sp. Zool: Anzeiger, Bd. 36, 253-2 pp- : Parisi, B. 1913. Sulla sphaerospora caudata. Atti della Societa Italiana di Scienze Naturali, Vol. 51, pp. 1-11 * In consequence of the war I did not see my own reprint printed in the Archiv für Protistenkunde so it is impossible to add volume and page reference. EVIDENCE FOR THE DEATH IN UTERO OF THE : HOMOZYGOUS YELLOW MOUSE? HEMAN L. IBSEN AND EMIL STEIGLEDER Curnot (1905) and Castle and Little (1910) have pre- sented conclusive evidence that yellow mice are always heterozygous and hence cannot be made to breed true. Their combined results show that when yellows are mated together the proportion of yellows to non-yellows in the offspring is almost exactly 2:1 instead of the usual 3:1 ratio resulting from the mating of heterozygotes. Castle and Little seem justified on this account in assum- ing that the homozygous yellows are not viable, especially since the size of litter from the yellow X yellow mating is markedly smaller than that obtained from yellow X non-yellow or non-yellow X non-yellow matings. Until quite recently no attempt had been made to de- termine embryologically the actual fate of the homozy- gous yellows. Since the present investigation was begun, however, Kirkham (1917) has published a preliminary statement of the results of such a study. His results, presented only in abstract, show that of 69 embryos from yellow X yellow parents, 26 or 37.8 per cent. were de- generating. For ‘‘non-yellows’’ he used albinos.2 Of 1 Papers from the Department of Experimental Breeding, Wisconsin Agri- cultural ee aad = No. 11. Published with the approval of the Director of the 8 [The problem a ne fate of the homozygous yellow mouse was under- taken at my suggestion during the summer of 1916 by Mr. Steigleder and the experimental work on which the present paper is based was done en- tirely by him. As he was, however, unable to complete the problem the accumulated material and records were turned over to Dr. Ibsen, who has checked all records with the preserved embryos and is alone responsible for the tabulation, interpretation and presentation of the results—L. J. Cole.] 2 This selection of albinos for ‘‘non-yellows’’ was unfortunate, since they apparently were not tested genetically, and hence may or may not have carried the factor for yellow. From the fact that the proportion of dead embryos was markedly different we may accept Kirkham’s assump- ae 740 No. 612] DEATH IN UTERO OF MOUSE 741 the 84 embryos from albino parents, only 2, or 2.3 per cent., were degenerating. This makes it seem quite prob- able that the homozygous yellow zygote develops for a time and then dies. In our study no attempt has been made to investigate the very early stages, as was done by Kirkham, but a large number of embryos have been obtained from non- suckling females pregnant from 13 to 19 days.* In all 688 embryos have been examined. These have been ob- _ tained from (1) yellow females mated to yellow males, (2) yellow females mated to non-yellow males (choco- lates), (3) non-yellow females (chocolates) mated to yel- low males, and (4) non-yellow females mated to non- yellow males. In this last mating most of the parents were self blacks. During the investigation two distinct types of dead em- bryos were encountered, (1) those in which development had ceased shortly after implantation, corresponding to those described by Kirkham, and (2) a few which seem to have developed normally till about the thirteenth day and then died, presumably because of overcrowding in the uterus. These latter were characterized by their dead, yellow appearance and smaller size as contrasted with the living pink color and larger size of the normal embryos. The first kind has been designated ‘‘dead em- bryos A’’ in the tables, while the second kind is classed as ‘‘dead embryos B.” We are primarily concerned with ‘‘dead embryos A,’’ and it is to be understood that ref- erence is to this type unless specifically stated otherwise.* Similarly, by ‘‘living embryos’? we mean those which tion that they were really genetically non-yellows, though this brings the argument dangerously close to reasoning in a cirele. In our work a number of albinos were mated both to yellows and to non-yellows, but since the genetic constitution of the albinos was not sufficiently established in most ases, the embryos from these matings are not included in our tabulations. 3 The normal duration of gestation in the mouse is about 21 days, but it is often less. 4 We are indebted to Dr. Alva Wilson for sectioning for us dead embryos from each of the four types of matings. Microscopical examination of these has verified our previous conclusions as to their character. 742 THE AMERICAN NATURALIST [ Vou. LI were obviously alive when the mother was killed. As the females were killed by chloroforming, all embryos were usually dead by the time they were examined. The following tables give the results obtained. Each of the first four tables represents one of the four types of mat- ings described above. In a few instances there has been some uncertainty as to the exact stage of gestation at which the embryos were removed, but in such eases this has been determined approximately by means of weights of the living embryos. TABLE I YELLOW 9X YELLOW ¢ r Average Size of Litter Stage of Ges- No. of Livi: Dead Dead Percentage tation Litters Dauro ae hes cas ae energy Living Total Embryos | Embryos 13 days: 5 33 14 29.8 6.6 9.4 days: onii 6 50 9 15.3 8.3 9.8 15 days. oes: 4 26 10 1 27.0 6.5 9.3 Todays: ri 42 13 2 22.8 6.0 8.1 Beaks 5 25 4 13.8 5.0 5.8 18 days...... 5 27 13 32.5 5.4 8.0 IO:dayaic es: 1 4 100.0 4.0 ‘SOtGh =. se 5 33 203 67 3 24.54 6.15 8.27 TABLE II YELLOW Ọ X NON-YELLOW ¢ Average Size of Litter Dead Dead Percentage Stage of No. of Living Gestation Litters Embryos agri T e arr Living Total Embryos | Embryos Saays. ENA 5 34 12 26.1 6.8 9.2 E a A 6 56 1 1.8 9.3 9.5 days. . 4 35 4 10.3 8.8 9.8 16 days. ..... 2 14 8 36.4 7.0 11.0 17 dave, 2. <. 3 25 1 3.8 8.3 8.7 1S days co 3 17 1 1 5.3 5.7 6.3 19days. sa; 1 10 10.0 10.0 Total ..... 24 | m 27 1 12.33 | 7.96 | 9.13 The data presented agree in the main with Kirkham’s, but our percentage of dead embryos A, from the yellow X yellow mating (Table I) is considerably less than his, while our percentage from the non-yellow X non-yellow No. 612] DEATH IN UTERO OF MOUSE 743 mating (Table IV) is somewhat higher. Theoretically, if we assume that a certain proportion of embryos o other gametic composition die from unknown causes in mice of all colors and that all the homozygous yellows are TABLE III NON-YELLOW Ọ X YELLOW ¢ re | ? | Average Size of Litter Stage of No. of Living Be se aan | Soret Gestation Litters | Embryos | Embryos | Embryos | host | Tiving | Total Embryos | Embryos tI dayo., 1 2 1 33.3 2.0 3.0 ee POT 1 6 6.0 6.0 PES EE PO 2 15 7.5 75 Ip dave. : is: 2 20 1 a d 4.5 10.0 11.0 17 days. 2 11 5. 5 IS dam.. iig 1 8 1 11.0 8.0 9.0 19 days...... Total cos, 9 62 3 1 4.55 6.89 To TABLE IV NON-YELLOW ? X NON-YELLOW ¢ | i TIENER Average Size of Litter Stageof | No.of Living Dead Dead Dead Gestation Litters | Embryos ur ver Emon Embryos | Living Total A Embryos | Embryos IS days. 0% 2. 3 22 3 12.0 7.3. 8.3 14 days...... 3 28 1 3.4 9.3 9.7 15 days. 2... 16diys. 3 Z 7.3 8.0 17 dat... 3 19 6.3 6.3 18 days...... 3 26 r: -10.3 8.7 9.7 19 dave. e 1 4 4.0 4.0 LO oe 16 121 7 2 5.38 7.56 8.13 represented as dead embryos, then the difference between the percentages of dead embryos from the yellow X yel- low mating and the non-yellow X non-yellow mating should be approximately 25 per cent. In our results the difference is 19.2 per cent., while in Kirkham’s it is 35.5 per cent. Neither of these is especially close to the ex- pected percentage. If, however, his results and ours are combined the difference is 23.0 per cent., which is close to expectation. If instead of comparing yellow X yellow with non-yel- low X non-yellow only, all matings other than yellow 744 THE AMERICAN NATURALIST [ Vou. LI X yellow are combined for this purpose, it is found that the dead embryos in these combined matings constitute 8.9 per cent. of all embryos present. Subtracting this from 24.5, the per cent. of dead embryos in the yellow X yellow mating, the difference is only 15.6 per cent. When all these results are combined with those of Kirk- ham the difference is still only 19.4 per cent., which is con- siderably lower than that obtained by using only the classes considered by him. An attempt will be made farther on to explain this deficiency. Tables II and III, representing the reciprocal crosses of yellow X non-yellow, show marked contrasts in sev- eral respects. In the yellow ? x non-yellow & mating (Table II) 12.3 per cent. of the embryos were dead, while in the non-yellow Ẹ X yellow ¢ mating (Table III) the percentage is only 4.5 per cent. For both matings the percentage of dead embryos theoretically should be the same as in the non-yellow X non-yellow mating (Table IV), since in neither case is there the possibility of any of the offspring being homozygous yellows. ‘The per- centage in Table III is approximately the same as in Table IV, but in Table IT it is much higher. It is well known that yellow females tend to take on more fat than females of other colors. There is a possibility that this physiological difference may also in some way influence the production of a greater number of dead embryos A. This is offered merely as a suggestion, since there is no direct evidence for or against it. In discussing the stages of pregnancy in which the dead embryos are to be found, Kirkham makes a statement which does not accord with our observations. He says: ‘‘No degenerating embryos have been found in either white [albino] or yellow mice pregnant more than sixteen days.’’ He believes complete resorption of the degen- erating embryos has taken place by the end of the six- teenth day. Table V, which is a summary of all our four types of matings, indicates that in our material there is no marked decrease in the percentage of dead embryos toward the end of pregnancy. . No. 612] DEATH IN UTERO OF MOUSE 745 TABLE V RELATION OF DEAD EMBRYOS TO STAGE OF PREGNANCY FOR THE Four TYPES OF MATINGS Stage ot Gestation e ee TT a ot poet E E E T ee 30 121 24.8 E e Va PE a A ees Ri 11 151 73 SO GSEs £55 hee he 14 91 15.4 1G GRUB io ee ee te 22 125 17.6 Le AGE crt i koe awn cng 5 (eS alk Raa ary AR arene 18 97 18.6 Moe n TE A SAO 4 18 22.2 His ee a | 104 688 15.12 As previously stated, Cuénot, and Castle and Little noted that the average number per litter is less from yel- low females mated to yellow males than from any other type of mating. Our results, summarized in Table VI, bear this out. Here it will be noted that the average num- ber of living embryos is less for mating 1 than for any TABLE VI LITTER SIZE FOR THE DIFFERENT TYPES OF MATINGS Average Number of Embryos per Litter No. of i ETUS E Nng Litters | Living Em- Dead Embryos bryos A and B Total 1. Yellow 9 X yellow o'......4...:| 33 6.15 2.12 8.27 2. Yellow 9 X non-yellow co’....... 24 7.96 1.17 9.13 3. Yellow g -yellow F ....... 9 6.89 0.44 7.33 (2 and 3 combined)............. 33 7.67 0.97 8.64 4. Non-yellow 2 X non-yellow o.. 16 7.56 0.57 8.13 (2,3 pe 4 combined) .......... 49 7.36 0.84 8.47 of the other types of matings singly or combined. It is evident that the average number of living embryos per litter represents those which would in the regular course of events have been born alive. In the yellow X yellow mating the average is only 6.15,° while for the other mat- ings combined it is 7.63. The average number when the dead embryos are also included should be approximately the same in both cases, and this proves to be true, being 8.27 for yellow X yellow and 8.47 for all other matings. 5 Two of the litters in this mating were made up entirely of dead em- bryos A and ue es had no living embryos. So far as living embryos are concerned their size would be 0, and they have been included as such with the other litters produced by this mating. 746 THE AMERICAN NATURALIST [ Vou. LI During 1916-17, the period at which the embryological study was carried on, numerous living litters of mice were born in the laboratory. Miss Sarah V. Jones, who was also working with mice, has generously furnished data from her own breeding experiments, and these have been incorporated with ours. Some of the mice in her experiments were black-and-tans, but as Dunn (1916) has shown that these are a form of yellow and are also always heterozygous, they have been classified as yellows. Table VII shows the average size of litters for the various types of matings. TABLE VII AVERAGE SIZE oF LITTERS Born Durine 1916-17 (Figures in parentheses indicate the number of litters.) Type of Mating Average Number of Mice per Litter At Time Color was Type of Mating Recorded iy Bellew OM yellow. g totes as 5.36 (140) 4. 55(121) 2. Yellow F X non-yellow du... ....s0%s 6.21(88) a §,63(71) 3. Yellow d X non-yellow F ............ 7.02 (50) 5.95(46) (2 and Seombined} 25005 220008 6.51(138) 5.76(117) 4. Non-yellow ? X non-yellow ¢ ........ 6.88 (42) 5.49 (37) (2, 3 and. 4 combined). .....6. 2. arau 6.59 (180) 5.69 (154) It will be seen that there is a deficiency of litter size from the yellow X yellow mating here as in the em- bryological material. Theoretically the average size of litter from yellow parents should be 75 per cent. of that from any of the other matings. Both Cuénot, and Castle and Little, however, have found it to be above 80 per cent. Our results are in accord with these findings.® Table VIII shows the percentages found by the various investigators. In order to make them comparable with the others our results as given in the table are for the living litters only and do not include the data from the embryological investigations. 6 Miss Durham (1911) found very little difference in average litter size between the offspring from yellows X yellows and the offspring from mat- ings where at least one of the parents was a non-yellow. She says: ‘‘Only mice which lived long enough to have their colors determined are included in these averages.’’ It seemed possible to us that she had not found any No. 612] DEATH IN UTERO OF MOUSE 747 TABLE VIII PROPORTIONATE SIZE OF LITTERS FROM YELLOW X YELLOW AND YELLOW X NON-YELLOW MATINGS (Figures in parentheses represent number of litters.) Average Size of Litter atio of * Yello Authority x Yellow” to‘ 'Yel- x YellowX Yellow | Yellow xNon-yellow low XNon-yellow ” Guia OO a aiana 3.38(50) | 3.74(50) 90.37 % Castle and Little Sedo eS 4.71(277) 5.57(325) 63 % bsen and Steigleder, 1917...... 5.36(140) 6.51(138) oe 33 % The most striking fact brought out in this table is that as the average size of the litters increases the ratio tends to decrease and therefore to approach 75 per cent. It would seem from this that if one could secure a race of mice having a high enough average per litter the the- oretical percentage could be obtained. Using Table VIII as a basis for computation, such a race should average approximately 10 young per litter in the non-yellow x yellow mating and consequently about 7.5 per litter for the yellow X yellow mating. It is not Prone that a race of this sort exists. Various theories have been advanced to explain this unexpectedly large litter size in the yellow X yellow mating. Two suggested by Castle and Little will be con- sidered here. Both take as their starting-point that ‘‘the perishing of a pure [homozygous] yellow zygote makes possible the development of a certain number of other fertilized eggs.’’ The explanation follows: ‘‘Two ways may be suggested in which this might come about. First, more eggs may normally be liberated at an ovulation than there are young born subsequently. In that case, failure of some eggs to become attached to the uterus may make the chances greater that the remainder will become at- difference because the size of litter had not been recorded at birth. Our records have been gone over and the litter size found at the time the colors were recorded. Some litters did not live long enough to be recorded in this manner and hence could not be included. By referring to Table VII one ean see that our results do not at all agree with Miss Durham’s. As a matter of fact, in our material the percentage relation is lower than when the litter size was computed at birth, being almost exactly 79 per cent. when ‘‘ yellow x yellow’’ is compared with ‘‘non-yellow X yellow.’’ 748 THE AMERICAN NATURALIST [ Von. LI tached, or the perishing of some may make the chances greater that the rest will successfully complete their de- velopment. Or secondly, the production of a relatively small number of young at one birth may lead indirectly to more free ovulation subsequently, and so to the pro- duction of a large litter at a second birth.’’ It will be seen in the first place that the above theories are based on false premises. It was not known at the time these theories were proposed that the homozygous yel- low zygote does not perish in the sense of disintegrating and finally di ing. It merely ceases to develop after a certain stage has been reached, and then remains more or less stationary till parturition. It might still be maintained that since these undeveloped zygotes take up very little room there would still be the possibility for the ‘‘other’’ zygotes to develop. In that case the average number of total embryos (including dead embryos A and B) should be greater in the yellow X yellow mating than in the non-yellow X yellow mating. Our data, presented in Table VI, indicate that this is not the fact. The other theory that ‘‘the production of a relatively small number of young at one birth may lead indirectly to more free ovulation subsequently, and so to the production of a larger litter at a second birth’’ still remains a possibility so far as our data are concerned. There are several other possibilities which seem worthy of consideration. Instead of looking for causes that tend to increase the size of litters from yellows X yellows it may be profitable to determine if possible causes that may decrease the size of litters from the non-yellow X yellow mating. The first and most obvious possibility is that overcrowding in the uterus may have this effect by causing the death of some of the embryos. However, none of our results bear this out. In the first place there are proportionately just as many dead embryos B (whose death is probably caused by overcrowding) in the one type of mating as in the other, and in the second place this would mean that for our race of mice, having a high average litter size, there should be a proportionately No. 612] DEATH IN UTERO OF MOUSE 749 large death rate due to overcrowding, and this would tend to increase the percentage relation between ‘‘yellow x yellow’’ and the ‘‘non-yellow X yellow’’ instead of de- creasing it, which actually is the case. We know that dead embryos B do not materially de- crease the litter size in the non-yellow X yellow mating, but on the other hand we have clear evidence that dead embryos A have a decided effect in this respect. The only manner we could postulate in which this could affect the litter-size percentage relation for the two contrasted types of matings would be to assume that in the yellow xX yellow mating dead embryos A are due almost entirely to the fact of their being homozygous yellows, while in the non-yellow X yellow mating separate agencies are at work producing an appreciable number of deaths. If a more careful examination of dead embryos A should re- veal rather easily. distinguishable differences this ex- planation could be tested. There is still another explanation which so far as we know has never been suggested, and which has more evi- dence in its favor than any of the others proposed. When yellows are mated to yellows it is to be expected that some of the litters, especially if they are small, will consist en- tirely of homozygous yellows. Since these do not com- plete development, the entire litter will be composed of dead embryos A and consequently there will be no living embryos born at parturition. Such a litter will there- fore be of 0 size. In our embryological investigation two . litters consisting entirely of dead embryos A were found in the yellow X yellow mating and none in the other types of matings. Even with these two litters of 0 size in the yellow X yellow mating the average size of litter as com- pared to the litters of the non-yellow X yellow mating was not sufficiently low to bring the percentage relation down to 75 per cent. It is, however, 80.18 per cent., which is appreciably lower than the percentage obtained from litters of animals allowed to give birth to their young. (See Table VIII.) In the latter case it would naturally be impossible to detect the litters of 0 size. AS 750 THE AMERICAN NATURALIST [ Vou. LI In our embryological study there were 33 litters in the yellow X yellow mating. Of these, as previously stated, two consisted entirely of dead embryos A. This means that for 31 litters containing living embryos there were two that did not, or 6.45 per cent. If we assume a like 35 T T T T T T T T T T T 30 w A. 25F eee UN Fi í VAA his “a g %7 Pi 2 20} kod X 7S 4 7 O y IY ETI N, koma b , (A @ S ® N © i5} ¥ x) 2 4 5 oa Re = 10k N Š Ya x <4 = X; N if N Pa 5 serii A * ‘ bee pe eee Noes D er See efe i L L i fi fi l l L korbi 0 i g 3 8 9 10 il L 5 6 7 Size of litters Fig. 1, percentage of litters of 0 size with our 140 normally born litters from ‘‘yellow X yellow” (Table VIII) the total number of litters would be increased to 149 and their average size would be 5.04 instead of 5.36. Taking 5.04 as a basis for comparison with 6.51, the average size of litters from the non-yellow X yellow mating, we find that the percentage relation is 77.42 per cent. This is reason- ably close to 75 per cent., the expected percentage. In this connection it may be well to call attention to a fact, previously stated, that as the average litter size increases the percentage relation tends to approach 75 per cent. This is exactly what one should expect. For with in- creased size of litter there would be fewer litters of 0 size, and the averages as actually found would more arsed represent the true averages. Curves (Fig. 1) have been constructed oiin the frequency of the litter sizes for certain of the types of matings. The data upon which they are based may be found’ in Table IX. A careful survey of these curves, No. 612] ` DEATH IN UTERO OF MOUSE 751 however, does not seem to lead to any very definite con- clusions as to the special fitness of any of the various theories discussed in the preceding paragraphs. It will be seen that in the yellow X yellow mating there are pro- portionately more litters of small size and fewer of large size than are to be found in the other matings. This is what one should theoretically expect on all the theories. TABLE IX LITTERS Born Durine 1916-17, CLASSIFIED ACCORDING TO NUMBER OF YOUNG IN LITTER | Number of Litters | No. in Litter Yel. $ X Yel. 7 bree Ogee be ae N athe Y Non.- T rae x 1 2 3 3 2 3 21 5 1 1 4 25 11 3 4 5 25 13 4 f 6 24 18 9 5 {T 25 11 6 12 8 10 10 12 3 9 3 il Bi 4 10 3 3 5 5 il 2 3 1 Total. 140 88 50 42 Summing up, we may say that all of our evidence tends to confirm the conclusion of Castle and Little that in mice homozygous yellow zygotes are produced in the yellow X yellow mating, but that these zygotes fail to de- velop normally after implantation in the uterus. Why this should be so is not evident and our investigation has not thrown any light on this point. Itis possible a careful microscopical study of the embryos which die early might reveal some abnormality of development which would account for their failure to survive, but it is not probable that it would be of such a simple nature as the analogous cases of death of homozygous recessives lacking chlo- rophyll in corn and some other plants. It seems more probable that in mice there may be a ‘‘ lethal factor,’’ similar to those so well known in Drosophila, which is so closely linked to the factor for yellow that they are prac- 752 THE AMERICAN NATURALIST [ Vou. LI tically at the same locus and there is consequently no crossing over. According to the investigations of Little (1915) an en- tirely similar condition obtains in regard to a dominant white spotting factor (W) in mice, which, like yellow, ap- pears never to occur in a homozygous condition. Whether the homozygous individuals in this case also die at an early stage and might be found as dead embryos has not yet been determined. Little has, however, demonstrated (1917) that the two factors are independent in heredity, and that litters from yellows carrying W, mated inter se, average only three per litter (10 litters), while similar yellows mated to ww non-yellows have litters averaging 5 per litter (9 litters). Although the numbers are small the percentage relation for the two matings, 60 per cent., is quite close to the theoretical expectation, 56.3 per cent. LITERATURE CITED Castle, W. E., and C. C. 1910. On a Modified pepe Ratio among Yellow Mice. Science, N. S., Vol. 32, pp. 868-870. - Cuénot, L. ; 1905. Les races pures et leurs combinaisons chez les souris. (4™° note.) Arch. Zool. Expér. et Génér., 4° Series, Vol. 3, Notes et Revue, pp. exxiii—cxxxii. Cuénot, L. 1908. Sur quelques anomalies BARR des proportions Mendeliennes. (6° Note.) Arch. Zool. r. et Génér., 4° Series, Vol. 9, Notes et Revue, pp. vii-xv. Dunn, L. C. 1916. The Genetic Behavior of Mice of the Color Varieties ‘‘ black- and-tan’’ and ‘‘red.’’ AMER. Nar., Vol. 50, pp. 664-675. Durham, F. M. 1911. Further Experiments on the Inheritance of Coat Colour in Mice. Jour. Genetics, Vol. 1, pp. 159-178 Kirkham, W. B. 1917. Embryology of the Yellow Mouse. Proceedings of the Amer. Zoologists, Abstracts. The Anatomical Record, Vol. 11, pp. 480-481, Little, C. C. 1915. The Inheritance of Black-eyed White Spotting in Mice. AMER. Nat., Vol. 49, pp. 727-740. Little, C. ©. 1917. The Relation of Yellow Coat Color and Black-eyed White Spot- Ang of Mice in Inheritance. Genetics, Vol, 2, pp. 433—444. SHORTER ARTICLES AND DISCUSSION ON THE FAUNA OF GREAT SALT LAKE In a recent number of the AMERICAN NATURALIST! appeared a paper entitled ‘‘ Notes on the Fauna of Great Salt Lake,’’ by Dr. Chas. T. Vorhies. From observations made by the present writer in the region at the mouth of Bear River, Utah, during the summer and fall months from 1914 to 1916 inclusive, infor- mation is available that supplements in part the data given by Dr. Vorhies. Bear River, the largest of the three main tributaries of Great Salt Lake, breaks up into a series of channels at its mouth and forms a great delta at the northern end of Bear River Bay. Im- mediately below the mouth of the river the waters of the bay are freshened by the incoming river water. Conditions here vary greatly however from day to day, and at present heavy salt water frequently comes up as far as Slaughter Island at the lower part of the marsh area in the river delta. Below this point the sur- face water coming from the river may be fresh while at a depth of a few inches a stratum of brine may overlie the mud. On calm days this overlapping proceeds for long distances. The prevailing summer winds however are from the south and south- west and these drive the salt water in toward the marsh nearly every afternoon. It is common belief that the Southern Pacific cut-off has served as a dam to separate Bear River Bay from the main lake (cf. Vorhies, p. 494). For a considerable distance this causeway is made up of trestle work allowing free interchange of water from either side. Although tests of the density of the water were not made, the writer is certain that the difference in content of salts between the water on the north and south sides of the cut-off is slight while water sufficiently saline to enable the life characteris- tic of the Lake to flourish is found at least twelve miles above the cut-off and within four or five miles of the point at which the main ae of Bear River opens into what is known as South Ba 1 Vol. LI, No. 608, August, 1917, pp. 494-499. 753 754 THE AMERICAN NATURALIST [ Vou. LI Brine shrimp, Artemia fertilis Verrill, occurred at this point in enormous numbers and adults and larve of alkali flies, Ephydra gracilis Packard and E. hians Say, were abundant. E. subopaca Loew was less common. The brine shrimp were gathered in great masses, and took advantage of the slightest depressions in the mud as shelters against the ever-fluctuating currents. Thousands frequently gathered in the lee of the boat during periods of observation by the writer. In May and June adult Ephydra were found on mud, laid bare by water receding from the high spring levels, in which alkalis were rising through surface evaporation. On these areas the flies formed dense masses several feet square. The insects were busily probing or kneading the mud with their proboscides so that the surface was heavily pitted or stippled with small de- pressions that were visible at a distance of several feet. The greater part of these insects were Ephydra gracilis. The statement made by Dr. Vorhies (p. 498) that ‘‘enemies play no part in keeping down the numbers of Artemia, or of Ephydra in the larval stage’ is not corroborated by observa- tions of the present writer. After the first of September each year shovelers, Spatula clypeata (Linneus), began to congregate in the bay below the mouth of Bear River, and by October 1 thousands of these ducks were present. The birds lay in great banks on the open water, and it was not unusual to see such flocks that were at least two miles long and from one quarter to one half a mile broad. The shovelers were feeding almost en- tirely upon Artemia fertilis and larve and pupæ of Ephydra, and were crammed with them constantly. Usually this species of duck is not a good table bird but individuals shot here were all exceedingly fat, and the writer found them excellent eating. These ducks remained in fall until the fresh water bays were covered with ice. Another species of duck, the lesser scaup, Marila affinis (Eyton), came into this region from the north be- tween October 2 and 12 each year, and by October 20, was abun- ant. These lesser scaups also frequented the lower bay, and, like the shovelers, fed to a large extent upon the brine shrimp and the immature stages of the alkali flies. At dusk on October 14, 1914, flocks of these ducks were observed from Promontory Point passing from Bear River Bay southwest past Fremont Island in the open lake. As there is no fresh water feeding ground in that direction it was assumed that they were going out No. 612] SHORTER ARTICLES AND DISCUSSION 755 to feed at some favorable locality in the lake. Later in October the other ducks were joined by American goldeneyes, Clangula c. americana Bonaparte, while from observations it was certain that the green-winged teal, Nettion carolinense (Gmelin), at times fed upon this same food. The number of crustaceans and fly larve destroyed by these birds must be enormous. In addition to these ducks great flocks containing thousands of Wilson’s phalaropes, Steganopus tricolor Vieillot, and northern phalaropes, Lobipes lobatus (Linneus), are found on the salt water during migrations, where these birds likewise feed upon the brine shrimp and the fly larve and pupe. During October and November flocks of eared grebes, Colymbus nigricollis cali- fornicus (Heermann), were found on the lake along the cut-off where their food must have been taken from the same supply as none other suitable is found. It may be mentioned here that a considerable number of shovelers and many thousand eared grebes winter on Owen’s Lake in California, where saline condi- tions are similar to those in Great Salt Lake, and where a simi- lar fauna is found. Avocets, Recurvirostra americana Gmelin, and black-necked stilts, Himantopus mexicanus (Miiller), also fed upon Artemia and Ephydra at the mouth of Bear River, and no doubt these animals furnished food to other shore birds. Definite data on this point is not at hand, however, as all of the shore bird stomachs collected there have not yet been examined. It is probable that the ring-billed gull, Larus delawarensis Ord, and California gull, Larus californicus Lawrence, also take this same food at times. From the facts outlined above it will be seen that the toll taken by birds from the brine shrimp and alkali fly larve and pupex during the course of a season constitutes a mass of indi- viduals almost beyond comprehension. The digestion of food by the birds concerned is always a rapid process, and with soft- bodied animals like the brine shrimp a considerable mass would be consumed each day; and the same is true of the larve and pupe of the alkali flies. The immense number of these creatures in the waters of the lake must be attributed to the large number of offspring produced rather than to an absence of enemies. ALEXANDER WETMORE. BIOLOGICAL SURVEY, WASHINGTON, D. C. 756 THE AMERICAN NATURALIST [ Vou. LI FUSION OF ‘‘RHINOPHORES”’ IN CHROMODORIS! THERE have been found during the present spring nine spec- imens of the nudibranch Chromodoris zebra Heilprin which form a series exhibiting an interesting gradation in the degree of coalescence of the ‘‘rhinophores.’’ The animals were each of average adult size, 10-12 em. in length. In none of these cases was there any evidence that the structural variations had resulted from injury. In the period over which these individ- uals were obtained there were also collected about 1,000 normal specimens of the same species. These figures give, however, no precise idea of the relative frequency of ‘‘rhinophore’’ variation, because a larger number of specimens had been collected in previous years without any occurrence of these variations being observed. aa) (4) (2 | (a) Fies. 1-6. Outlines of anterior ends of Chromodoris zebra Heilprin, show- ing increasing degrees of fusion of the “rhinophores.” Fig. 5, frontal view; the rest, dorsal aspects, g. la, the normal condition; Fig. 1b, variation in the edges of the “rhinophoral’’ collars of three individuals The bases of the two ‘‘rhinophores’’ of C. zebra are, as in other Dorids, normally surrounded by well developed individual cylindrical collars. The distal termination of a collar is usually circular in outline, but occasionally pointed at one side (Fig. 1). In two specimens the ‘‘rhinophoral’’ collars were closely approximated, after the fashion outlined in Fig. 2. Three spec- imens were found in which the ‘‘rhinophoral’’ collars, and the depressions into which the ‘‘rhinophores’’ are separately re- tracted when stimulated, had completely fused (Fig. 3). In ~ 1Contributions from the Bermuda Biological Station for Research, No. 75. No. 612] SHORTER ARTICLES AND DISCUSSION TST these animals the two ‘‘rhinophores’’ themselves were separated by their normal distance of about 1 em. The next step in ‘‘rhin- ophore’’ fusion is illustrated in Fig. 4, one example having been collected. In another specimen the ‘‘rhinophores’’ were found to be closely united at the base (Fig. 5), while in the remaining two specimens that exhibit fusion of the ‘‘rhinophores’’ (Fig. 6) the process of coalescence had been pushed much further, a single stalk, giving rise at its free end to two short diverging projections, representing the normal pair of ‘‘rhinophores.’’ ‘rhinophore’’ of C. zebra is locally stimulated by being touched, it is retracted within its pocket, the basal collar usually contracting over it, while the companion ‘‘ rhinophore’’ on the other side of the animal is usually not contracted. In other words, the ‘‘rhinophores’’ are, with reference to their re- traction, subject to independent bilateral control. . The process oe retracting the ‘‘rhinòphore’”’ consists of two phases—the ‘‘rhinophore’’ is itself contractile, and it is in addition pulled down into its pocket by the action of muscles situated at its base. With the fused ‘‘rhinophores,’’ even in such eases as that illus- trated in Fig. 6, the independent bilateral control of the organs is pre ved. If one tip be stimulated, that side of the compound ‘‘rhinophore’’ is contracted, the other (unless the stimulation be severe) remaining inert. Under slightly stronger stimulation applied to one tip of a compound ‘‘rhinophore,’’ the contraction of the organ itself is immediately followed by the traction of muscles upon the same side of the base of the double “‘rhinophore,’’ resulting in a bending of the whole structure toward the point of excitation. . The reactions of the abnormal specimens therefore support. the view that these abnormal ‘‘rhinophores’’ have been pro- duced by a process of fusion, probably rerniang from the orig- inal close approximation of ‘‘rhinophoral’’ Anlagen. Two cases have been available for experiment in which one of the normally placed ‘‘rhinophores’’ possessed a divided tip; these divided-tip ‘‘rhinophores,’’ superficially not unlike the single median struc- ture above described, gave no evidence of independent control for the two tips, both parts contracting together when one tip was irritated. It would appear that the development of the collar surround- ing the base of the ‘‘rhinophore’’ is directly dependent upon the growth of the latter structure; in every case there was a close 758 THE AMERICAN NATURALIST [ Vou. LI correspondence between the bulk of the - Fhinopbhoran:) and the dimensions of the collar or collars. W. J. CROZIER AGAR’S ISLAND, BERMUDA NOTE ON THE HABITAT OF GEONEMERTES AGRICOLA! THE terrestrial nemerteans include a small number of species, all belonging, apparently, to one genus, but widely scattered over the world. They occur conspicuously on islands, some of which are well removed from any large mainland. The origin of these land nemerteans is a matter of some interest, and sev- eral suggestions have been made relative to the manner of their evolution. One of these terrestrial nemerteans, Geonemertes agricola (W.-S.), was found at Bermuda by v. Willemoes-Suhm (1874). The anatomy of this species was subsequently de- scribed in detail by Coe (1904), who gave some attention, also, to the habits of the worm. These observers, as well as Verrill (1902), agree that G. agricola is to be found ‘‘only along the shores of mangrove swamps and on the adjacent hillsides’’ (Coe, p. 566). Coe found it ‘‘not only above high-water mark but also for some distance along a zone which is covered for a short time each day with sea water,’’ but noted that the intertidal in- dividuals were ‘‘as a rule smaller than those living in the soil which is a little above the reach of the tide, but in earth which is nearly saturated with salt water.’’ Standing bodies of fresh water are absent in Bermuda. Coe consequently held that this particular species, at least, repre- sents a land nemertean which has almost certainly been derived directly from a marine ancestor, and not, as Montgomery (1895, p. 483) had argued for the generality of land nemerteans, from a fresh-water form. During the past several years I have repeatedly encountered G. agricola in a type of habitat which is significantly different from that recorded for this nemertean by the observers just quoted. In the neighborhood of every large or small mangrove ‘‘ereek’’ or swamp which I have examined, the worm has been found, in relatively considerable quantities, well below low-water mark even at spring tides. The species occurs in the localities 1 Contributions from the Bermuda Biological Station for Research, No. 76. No.612] SHORTER ARTICLES AND DISCUSSION 759 listed by Coe, but is also common among the masses and sheets of matted sea weeds (Laurentia, Valonia, Halimeda, and asso- ciated plants) which cover the bottom of Fairyland Creek. Specimens were also obtained from under rocks situated a few feet beneath low-water level, in muddy bays bordered by man- groves, such as Tucker’s Bay in Harrington Sound. The indi- viduals collected in these places embraced white forms, some with a tinge of pink, others decidedly pink (as in the ‘‘pale’’ form figured by Coe, 1904, Pl. 1). They were 30-60 mm. in length, and some contained embryos. In June and in July young Geonemertes were gotten among sea weeds in Fairyland Creek; these were 6-12 mm. in length. They were identified principally through the microscopic examination of the stylets and other organs. The stylets and stylet basis in these young specimens were of the juvenile type for this species, as figured by Coe (1904, Pl. 25, Figs. 21, 24, 25). These young specimens were in some cases pure white, in others tinged with ‘‘smoky brown.’’ I found no pinkish specimens less than 30 mm. in length. The observations upon the specimens of this species inhabiting salt water indicate, as Coe concluded from his study of the land- living individuals, that liberation of the young occurs in June and in July. My largest examples of G. agricola from the water were obtained in the spring. Large specimens of G. agricola are negatively phototropic, the ocelli occupying the region of the body most sensitive toward light. They orient away from the light with diagrammatic pre- cision. This response leads to their being found, during the day, under stones and about the roots of alge. It is hardly possible to credit the view that G@. agricola has extended the variety of its habitats during the brief time since Coe’s studies were made (1903); it is therefore necessary to believe that this species of nemertean is not only terrestrial in the proper sense, but truly marine as well. There seems no good ground upon which to distinguish and separate the indi- viduals found respectively on land, in the intertidal zone, and definitely in the sea water. The terrestrial ‘‘variety’’ may then be regarded as having originated, perhaps not so very long ago, from the form which is undoubtedly marine—unless one is pre- pared to believe that, introduced as a terrestrial form, it has at some time secondarily taken to the sea after a protracted evo- 760 THE AMERICAN NATURALIST [ Vou. LI lution as a land animal. This case seems to have some resem- blance to that of a grapsoid crab at Bermuda, Sesarma ricordi M.-Edw., of which a terrestrial variety has been described by Verrill (1908, p. 328). It is my impression that the larger marine specimens of G. agricola are less hardy, more easily caused to fragment by handling, than are those taken on land. This may, however, be merely a physiological consequence of differences in habitat, which could be exhibited within the life- history of a single individual. I have not been able to keep the salt water specimens alive after abruptly transferring them to damp earth. The young individuals, however, are quite hardy, and seem capable of enduring this treatment for several days at least. These observations add further, and possibly final, weight to the argument that some, at least, of the land nemerteans have proceeded directly from ancestors inhabiting salt water. W. J. Crozier AGAR’S ISLAND, BERMUDA REFERENCES Coe, W. R. 1904. The Anatomy and Development of the elena Nemertean (Geonemertes agricola) of Bermuda. Proc. Bost. Soc. Nat ist., Vol. 31, ka 531-570, pl. 23-25. se Poiainin Biol. Sta., No. 4.) gapti Te ti dr. 1895. T Derivation of the Freshwater and Land Nemerteans, and d Questions. Jour. Morph., Vol. 11, pp. 479-484. Verrill, A. E. 1902. The Bermuda Islands. (Repr. from Trans. Conn. Acad. Sci Vol. 11, with changes.) New Haven, x + 548 pp., 38 pl., inde. 1908. Decapod Crustacea of Bermuda. I. Brachyura and Anomura. eir Distribution, Variations and Habits. Trans. Conn. Acad. Arts and Sci., Vol. 13, pp. 299-474, pl, 9-28. Willemoes-Suhm, R. V 1 On a Taia Widhük found in the Bermudas. Ann. and Mag. Nat. Hist., ser. 4, Vol. 13, pp. 409—411, pl. 17. NOTES AND LITERATURE SUNSPOTS, CLIMATIC FACTORS AND PLANT ACTIVITIES VARIATIONS in solar radiation, if of sufficient magnitude, should be followed by variations in terrestrial climatic conditions. In the absence of long series of determinations of the heat radiated by the sun, meteorologists have turned to variation in the num- ber of sunspots as a possible factor underlying variation in the climatic factors. This involves the assumption that a period of many sunspots differs from a period of few spots in heat and light radiation. If such climatic factors as heat and precipitation be closely re- lated to the number of sunspots, the number of sunspots should be a factor of importance in determining plant activities. In. recent years attempts have been made to correlate growth, espe- cially that recorded in the annual rings of trees, with number of sunspots." Since any attempt to relate growth phenomena to sunspot number presupposes relationships between climate and sunspot frequency, the botanist should be interested in the attempts of the meteorologist to ascertain the relationship between solar and terrestrial atmospheric phenomena. The purpose of this review is to call attention to certain recent discussions of this subject. For a review of earlier literature, the reader must refer to Hann’s ‘‘Handbook of Climatology’’ and to the careful discus- sion from the biological side in Chapter XIX of Huntington’s ‘‘ Climatic Factor.’’? Heretofore, those who have discussed the interdependence of terrestrial and solar phenomena have been content to plot curves for the two phenomena and to determine the existence of relation- ship between them by similar trends in the two curves. The sources of error in such a method are very great. Fur- thermore, it gives no quantitative measure of intensity of rela- tionship. Such a measure can only be secured by means of some correlation or contingency coefficient. 1 Douglas, A. E., ‘‘A Method of Estimating Rainfall by the Growth of Trees.’’ In Puer Huntington’s ‘‘The Climatic Factor,’ pp. 101-121. 2 Huntington, E., ‘‘The Climatie Factor as Illustrated in Arid ame £ Pub. Carn. Tast. Wash., 192, 1914. 761 762 THE AMERICAN NATURALIST [ Von. LI Walker, the director general of observatories for India, has made a great advance in the investigation of the possible rela- tionship between the number of sunspots and meteorological phe- nomena by applying the modern methods of correlation to the problem I have tabulated his three series of correlations for a large number of stations widely distributed over the earth, with the results given in the accompanying table. | Frequency of Correlations Intensity of Correlation Sunspots and Sunspots and Sunspots and Rainfall Temperature Pressure — .59 to — .53 — 2 — — .52 to — .46 1 1 2 — .45 to — .39 5 6 2 — .38 to — .32 6 i 3 — .31 to — .25 9 10 8 — .24 to —.18 17 12 8 — .17 to —.11 15 15 8 — .10 to — .04 21 14 15 — .03 to + .03 27 16 11 + .04 to + .10 13 A 8 +.11 to +.17 18 5 10 +.18 to + .24 9 4 8 + .25 to + .31 T 1 6 +.32 to +. 6 — 2 +.39 to + .45 — — -— + .46 to + .52 1 — -— Total number of stations. 151 97 91 Remembering that correlation is measured on a scale of — 1 to -+ 1, this table shows at once that there is no uniformity for the globe as a whole in the correlations between number of sunspots and either of the climatic factors considered. Instead coeffi- cients for some stations are positive while those for others are ` negative. Thus as far as the data available go, they indicate that in some regions rainfall, temperature or barometric pres- sure are higher in periods of larger numbers of sunspots, whereas in other regions they are lower. The magnitude of the coeffi- cients is generally low.* Over thirty per cent. of the constants 3 Walker, Gilbert T., ‘‘Correlation in Seasonal Variation of Weather,’’ IV-VI. Mem. Ind. Met. Dep., 21: 10-12, 17-118, 3 world maps, 1915. 4 These correlations are based on such small samples that for their full interpretation the theory of the distribution of small correlations now being developed by ‘‘Student,’’ Soper, Fisher, Young, Cave, Lee and Pearson must be considered. Nothing brought out by the work of these writers, which will be reviewed later, invalidates the general correctness of the conclusions reached here. No. 612] NOTES AND LITERATURE 763 lie between + .10 and —.10. The probable errors of the con- stants are of about this magnitude. The average values of the three sets of correlations are: For sunspots and rainfall, 151 stations, r == — .0175 For sunspots and temperature, 97 stations, r = — .1360 For sunspots and pressure, 91 stations, r = — .0331 While these averages are exceedingly small, all are negative in sign, indicating that for the globe as a whole lower rainfall, tem- perature and barometric pressure are associated with greater numbers of sunspots. The same relationship is apparent if only the 76 stations for which all three relationships have been caleulated be considered. The averages are: For sunspots and rainfall, r = — .0349 For sunspots and temperature, T= — .1534 For sunspots and pressure, T = — .0486 If in obtaining the average correlations the constants for the several stations are weighted with the number of years for which records are available the averages, indicated by the bars, are: For sunspots and rainfall, 151 stations, r = — .0103 For sunspots and temperature, 97 stations, r = — .1243 For sunspots and pressure, 91 stations, r = — .0387 Thus, however calculated, the averages indicate generally negative values of the correlation coefficient for all three rela- tionships. The same fact is brought out if the coefficients are classified according to sign only. Thus: Number of Sunspots Frequency of Posi- | Frequency of Zero | Frequency of Nega- and: — tive Correlation Correlation tive Correlation SOTA i cs ks ew 70 i 80 Pempermture 6506. bce wk 18 4 75 is Fae Daa ay en iar haan aa 38 3 50 For all three relationships the negative correlations are more numerous than those which are positive in sign. In a brief review it is quite impossible to give in detail the meteorological considerations discussed in the original memoirs. Furthermore, the most of these do not directly concern the botan- ist. The conclusions of practical biological importance to be drawn from Walker’s investigations seem to be the following: 764 THE AMERICAN NATURALIST [ Vou. LI a. The relationship between the number of sunspots and the annual record of terrestrial meteorological phenomena is very slender indeed. It is so slight that at the present time it is im- possible to assert on the basis of the data of any one station alone that any relationship at all exists. Thus, as far as they go, these data hold out very little hope to the biologist of being able to correlate plant activities with sunspot number, unless light in- tensity be the means of solar influence. . For rainfall and barometric pressure the correlations are aiaa low. They average practically zero, but are apparently on the whole negative in sign. c. The correlation between number of sunspots and terrestrial temperature is the most consistent and substantial of the three. The coefficients average about — .14. Thus years of larger num- bers of sunspots are in the long run years of lower, not higher, terrestrial temperature." These results are directly opposed to the theories which seem to have prevailed among many writers. J. ARTHUR HARRIS 5 Possibly, as Walker suggests, superheating in equatorial regions may raise the temperature in the upper air but lower that at ground level. The temperatures in which the botanist is primarily interested are, however, which may influence the film of vegetation which covers the globe. INDEX THE NAMES OF CONTRIBUTORS ARE PRINTED IN SMALL CAPITALS, Alcohol and White Mice, L. B. NIcE, 596 LLARD, a A., Troman and Synchron e Rhythm Allelomorphs, Multiple, pee Modi- fying Factors in oats to 8 lection, H. S. JENNINGS, 301; versible iesen omoabiity of, H. ERAO, 690 Amphibia, a ge ap nba = rth rica GRE 311 FR Ae Literature, Sources of, Roy L. Moopts, Awn, The Inheritance of the Weak, in Ave S, H, Love and A. C. FRASER, 481 BAUMBERGER, J. P., Solid Media for rearing Drosophila, 447 Behavior, Synchronism and Syn- chronic Rhythm in, H. A. ALLARD, 438 Biocharacters as Separable Units of Organic Structure, HENRY FAIRFIELD OSBORN, 449 e Facts and Bud-varia- i E. M. East, 129; Signifi- vst of Animal Coloration, W. LONGLEY, 257; Enigmas and the Theory of En e Action, [LEONARD THOMPSON ‘TROLAND, 321 Biometrie Studies on the Somatic and Genetic Physiology of the Sugar Beet, J. ARTHUR HARRIS, gir ag and Mendelian Inheri- A.C. and A. L. Hae Bripges, Catvin B., the Variable Force Sy petheas of Crossing Over, 370 Bud-variation, E. M. East, 129 CastLtE, W. E., Piebald Rats and Multiple Factors, 102 Chromodoris, Fusion of Rhinophores in, W. J. Crozier, sa pe ee sa 3 Œnothera ANNE M. Lutz, 375 present Somatic, in Mutants. and Hybrids, | CLAUSEN, R. E. and T; H. G SPEED, rilate Pause Diter ences versus Reaction Ny coi a in seat 31, Cock ae A, snore Foss ssil F Fish- PAE S, 61 oe Animal, W. H. LONGLEY, Control, DIUNA, Field Tests OSEPH GRINNELL, 115 Correlation p ierg al and the Prob- f Varietal Differences in Disease wee J. ARTHUR Harr Crosses, ‘heen, aE of the ree a Lovs and ree Wheat, Linked _ Quantitative “Characters GEO F, FREE 683 PE ina al the Variable Force Hypothesis of, -CALVI B. BripeEs, 370 Crown Gall, ng sages of Prunus to, CLA apes TON O, SMI be bay ZIER, ions of IER, J. Migrat Tropical pa 377; Mul- Fission E olo- Geonemertes agricola, Daxc HAKOFF, VERA, Differentiation in the Developing Organism, 419 Datnges, L. L., The Flora of Great 9 DAVENPORT, CHARLES B., The Per- ity, Heredity and Work of Otis Whitm Dean and Eastman’s "Bibliography of Fishes, OLIVER P. Deer-mice, The Réle of TAR aa in the Formation of a e of, F. . SUMNER Differentia tion n by Se tion and Environment in the veloping a , VERA DANCHAKOFF, Dimidium neama, Mutation in, 8. O. Mast, 351 Distributional Control, JOSEPH oe NELL, Data, Erlis L. oo Sh 765 766 THE neal Mar Sere, Media for rear- ing, J. P. BAUMBER A Ksg relations in, Don C DUNN T m Nucleus and Cyto- N, as Vehicles of Heredity, 286 EAST, E. M., The Bearing of B ooN Facts on Bud- Satsnus, eithin and Dean’s ak bo se of Fishes, OLIVER P. Hay, Ecology of the Protozoa, TAN £ Egg, Insect, The Genesis of the Or- ganization of the, Ropert W. EGNER, 641, 705 — Action and Biological Enig- mas, LEO HOMPSON TRO- sae, 321 ERDMANN, RH Sexual Process in the Paana ie Life Cycle, 719 Evolution, from the pini pari of a Geneticist, A. FRANKLIN SHULL 361; and ’ Rats, A. °C. and A, K GEDOORN, 385 ‘Factor,’ ? The Different mer of the, Howarp B. FROS Factors, Multiple, and Pisbeld Pac, W. E. Cas Modifying, ENN geval Inheritance of, in Sheep, ARD N. WENTWORTH and J. E. nan Fish-seales, Fossil, European, T. D.. A. Fishes, The Migration of, Davin ARR JORDAN, 1 Fossil Fish-seales, European, T. D. i d H. H. Love, The Tokat ianio of the Weak Awn in Avena Crosses, Paor GEORGE clea gree ak titative Charac Crosses, 683 i ag E Mendelian Class, RAY- PEARL e Different tures, 429 Gates, R. R., The Mutation Theory and the Species Concept, 577 Gene, The Theory of the, T. H. Morean, 5 513 AMERICAN NATURALIST [ Vou. LI sa oe of the Daan of ond Insect Egg, Ropert W. HEG Genetics versus Paleontology, Wo. EGOR iaoea agricola, Habitat of, We d. ees Germ Plasm of ’ Œnothera, ROBERT T. HANCE, 567 GoopsPEED, T. H. and R. E. CLAU- SEN, Mendelian Factor Piter- sangen versus Reaction puer in Heredity, 31, Great Salt L Lake, Fauna pr CH T. VORHIES, ” 494; Pr Tipton WETMORE, 753; Flora of by L DAINES; 499 REGORY, WM. K., The Coal Meas- ures Amphibi a of North America, 311; Genetics versus Paleontol- gy, 622 piisa wl JOSEPH, Field Tests of Theo concerning Distribu- Fratii Conival, 115 HADLEY, papisa The Case of Tri- chomonas, 2 GEDOORN, A. A and A. L., Blend- n Inheritance, j 385 Pl IS, J. ARTHUR, The Applica- tion "of Correlation Formule to ietal Differ- ences in Disease Biometric Studies on the Somatic ic Physiology of the . Sugar Beet, 507; Sunspots, Cli- matic Factors and Plant Activi- ties, 761 HAUSMANN, LEON a The Ecology of the Protozoa, 15 Hay, OLIVER P., Dean and Tat man’s Bibliography, of caine 383 NER, ROBERT W., The Genesis of the Organization of the Insect Egg, 641, 705; Singing Mice, 7 Heredity, Mendelian Factor Differ- ences versus ion ne poor rests in, Te E. CLAUSEN, 92: Nodos and Cy ee as Vehicles of, L. Holothurians, ern falc by Fis- in, W. J. Crozier, 560 T HEMAN I. and EMIL STEIG- Death in Utero ae the Ho- mony gous Yellow Mouse, nbreedin De grees RAYMOND PEARL, 545; PS S Kinship, A Single No. 612] Numerical Measure of the Total Amount of, RAYMOND PEARL, 636 Inheritance, Mendelian, and Blend- g, ; = DOORN, 189; of ae. Weak Aw H. Love and A. C. Fra aiei "481; of Fertility in os Epwarp N. W: J. B. Sweet, 662 Isolation, The Rôle of, in the For- mation of a Race of Deer-mice, . B. SUMNER, 173 JENNINGS, H. 8S., Modifying Fac tors and Multiple ‘Allelomorphs and the Results of Select ibn, 301 JONES, DoNALD F., Linkage in Lyco- on 60 RDAN, DAVID STARR, The Migra- “Hoa of Fishes, 186 oe Degrees of, The Measure- and Numerical, oe of, ens PEARL, LIN wage E. W., Linkage in Maize Linkage in wr bie i bate DONALD F. JONES, 608 Linked ine eset ative Characters in Whea osses, GEORGE F. FREE- MAN, fasa. LITTLE, C. ©., Multiple Factors in Mice and Rats, 457 LONGLEY, W. H., The Selection Problem, 250; The Biological Significance of "Animal Coloration, 257 Love, H. H. and A. C. Fraser, The Inheritance of the Weak Awn in Cer Avena Crosses, 481 LUTZ, eye M., Characters Indic- ative of the Num ber of Somatic Chromosomes present in Ginothera Mutants and Hybrids, 375 Mast, Ci >» asa ge in Didinium Nas 851 Manan, Factor Differences versus ; SEW Inheritance and and A. L. HAGE- RIG sa 373; Blending, A. ©. DOORN Mice, and Rats, Multiple ~—. in, ©. ©. LITTLE, 457; White, d Al cohol, L. B. Nice , 596; Presta ROBERT EG „704 MICHAEL, ELL s L., a apoE Pea ical Prediction "tro tional Data and its Verification, | 572 INDEX 767 et hiere s, DAVID STARR 6; of ee oe Nudi- "ithe Theory of the Morghclogian! Prediction from Dis- tributional Data and its Verifica- 2 tion, ELLIS L. MICHAEL, aw ee Factors and Piebald Rats . CASTLE, AINTER, pt, in Dros mene busckii ‘oe ee . WARREN Mutants and Hybrids, (¬hera, Number of Somatic Chromosomes present in, ANNE 375 NIc, L. B., ‘tase Effects of Alcohol on White Mic , 59 Notes and Literate, 311, 383, 507, 6 Nucleus and Cytoplasm as Vehicles 86 ’ Periodic ; Shoreward Migrations, W. J. CROZIER, 377 (Enothera, Mutants and Hybrids, r of Somatic Chromosomes present in, ANNE M. Lutz, 375; An Attempt to modify the Germ Plasm ntan ugh the Germinating Seed, RoserT T. Hance, 567 Organism, Developing, Differentia- tion by rai, a and Environ- ment the, VERA DANCHAKOFF, 419 ot E of the Insect Egg, . Genesis of the, Roperrt W. HEG- 641, 705 ER OSBOR? ; HENR Y FAIRFIELD, Biochar- acters as obeak Units of Or- ie § 44! H PAINTER, THEOPHILUS S., A TA Mutation of Piophila casei, 3 06 Paleontology versus Genetics, Wa. K. GREGORY, 622 Panulirus argus, Regeneration in, Eoi . WALTON, 308 RAYMOND, The Selection Problem, 65; The Probable Er- r of a Mendelian Class Fre- noT, 144; Studies on aaea- : The Measurement and Nu- merical Exp.ession of Degrees of Kinship, 545; A Single Numer- ieal Measure of ot Total Amount of Inbreeding, 6: — “cultares, Howard B. Potea u The Rôle of Isola- tion in the Formation of a Race of Deer-mice, F. B. SUMNER, 173 j i, A Wing Mutation in, THEOPHILUS $S. PAINTER, 306 CLAYTON O. SMITH, 47 Quantitative Characters, Linked, in Wheat Crosses, GEORGE F. FREE- MAN, 683 _ ‘ow and ee Fac- C 102; and Evolution T C. and ryd L. HAGE- , 385; and Mice, Multiple Pastors in, ©, C. 457 wk O, Wai in Panulirus argus, C, W. N, Resistance of Prunus to Crown Gall, CLAYTON °0. SMITH, 47 oe Problem, RAYMOND grr sW“ He LONGLEY Faetors Southdown, Inheritance of Seep, Sou in, Epwarp N. so ia WORTH and J, > THE AMERICAN NATURALIST Ga i | “102, 186, 238, 301, 370, tur aoa, | Whitman, Charles (Von. LI TEIGLEDER, EMIL and HEMAN IBSEN, 1 ology of the, J. ARTHUR HARRIS 7 SUMNER, F. B., The Rôle of Iso- pe in the Formation of a Nar- Localize r- icin (Peromyseus), 1 Sun Spots, Climatic Faa y Plant Activities, J. ARTHUR tility in Southd wn Sheep, 66: Synchronism and nike so tiny thm Behavior, H. A. ALLARD, 438 Terao, H., On Reversible Trans- formability of Allelomorphs, 690 ger ene poe Pose ae of Al- lelomorp i The Cas Oie of, PHILIP ADLEY, 209 TROLAND, LEONARD THOMPSON, Bi- ological Enigmas and Enzyme Action, 321 Vormis, CHas. T., The Fauna of Great Salt Lake, 494 WALTON, A. o, Regeneration in Panulirus 308 Warren, Don C., Mutations in Dro- sophila busekii Coq, 699 WENTWORTH, Epwarp N. and J. B. Inheritance of oe. ai in Southdown wn Sheep, ETMORE, ALEXANDER, Fauna a Great Salt 753 les Otis, Personality, Heredity and and Work ‘of, CHARLES VENPORT, Wing Mutation’ in in Piophila casei, THEOPHILUS 8. PAINTER, 306 WRIGHT, Sewatt, The Probable Error. of Mendelian Class Fre- quencies, 3 n a ee ae