Skip to main content

Castle, William Ernest

Castle, William Ernest

(b. Alexandria, Ohio, 25 October 1867; d. Berkeley, California, 3 June 1962),


Castle was the fourth of six children born to William Augustus and Sarah Fassett Castle. Of modest means, Castle’s father had been a schoolteacher in Johnstown, Ohio, before turning to farming. Castle received a B.A. degree from Denison University in 1889, and after teaching for three years entered Harvard University, where he received a second B.A. in 1893, an M.A. in 1894 (under C. B. Davenport), and a Ph.D. in 1895 (under E. L. Mark). After several years of teaching biology at the University of Wisconsin (1895–1896) and at Knox College, Galesburg, Illinois (1896–1897), he was appointed instructor in biology at Harvard (1897–1903), assistant professor (1903–1908), and professor of biology (1908– 1936). From his retirement in 1936 until his death, Castle was research associate in mammalian genetics at the University of California, Berkeley. Castle was married in 1896 to Clara Sears Bosworth, who remained his constant companion until her death in 1940. The vitality that Castle brought to his work was somewhat masked by an external appearance of reserve and formality, yet he was strongly liked by his students and respected by his colleagues.

Castle’s interest in natural history began early. As a farm boy, he collected wild flowers and learned to graft trees and to reconstruct skeletons of animals. At Denison, Clarence J. Herrick, an enthusiastic teacher of geology, zoology, and botany, introduced Castle to Darwin’s theory of natural selection, a subject that fired the young man’s interest in biological ideas. Academic emphasis at Denison was on classics, however, and Castle majored in Latin. After graduation he taught classics at Ottawa University, Ottawa, Kansas, from 1889 to 1892. In the fall of that year he entered Harvard University to pursue further his interests in natural history. He received his Ph.D. with a thesis focusing on a cell-lineage study of the ascidian Ciona intestinalis. This work was directed toward the elucidation of certain disputed questions concerning the mode of origin of the primary germ layers of tunicates. Carefully and methodically tracing the embryonic state of every cell from first cleavage through late gastrula. Castle took issue with the prevailing view of the origin of the chordate mesoderm. He concluded correctly that the mesoderm in Ciona and other primitive chordates originates from pouches in the infolded endoderm of the gastrula, in a manner similar to that in the echinoderms. Ciona is a hermaphrodite in which individuals are self-sterile (i.e., sperm of an individual will not fertilize eggs from the same individual). In the course of his studies Castle also discovered that self-fertilization is prevented not, as had been supposed, by ripening of sperm and eggs at different times but by physiological incompatibility between the gametes. This phenomenon had already been observed in certain flowering plants but never before in animals; Castle speculated that such a block to self-fertilization was probably chemical in nature.

Before 1900 Castle described himself as an “experimental evolutionist.” While he continued work on some developmental problems in invertebrates, he also began to study the role of heredity in determining the sex ratio in mice and guinea pigs. As a result of breeding experiments and his reading of Darwin, Weismann, De Vries, and Bateson, Castle was prepared to grasp the significance of Mendel’s work when it was rediscovered in 1900. From that time on, the direction of Castle’s research was clear; he became an enthusiastic Mendelian and devoted the remainder of his career to investigating either aspects of Mendel’s laws or the relationship between Mendelianin heritance and evolution. For example, within a year or two after becoming familiar with Mendel’s work, he attempted to apply the concept of dominance and recessiveness to the inheritance of sex (1903). Somewhat complex, his interpretation involved a subsidiary assumption of selective fertilization (i.e., only certain sperm will fertilize certain eggs) and did not throw much light on the problem of sex inheritance. At the same time, however, Castle was carrying out a long series of experiments (1901–1906) on inbreeding and outbreeding in the fruit fly Drosophila. Castle was the first to use this organism for any extensive laboratory breeding experiments, and it was through his work that T. H. Morgan’s attention was initially called to Drosophila.

In 1951 Castle noted that four general questions had motivated his work between 1900 and 1920: (1) How generally applicable are Mendel’s laws? (2) Is Mendel’s assumption of the purity of gametes true? (3) Are the germ cells and somatic cells basically different, as Weismann asserted? (4) Can the fundamental nature of genes be modified by selection? The first question was attacked in a series of breeding experiments with guinea pigs between 1903 and 1907. Castle showed that the inheritance of coat colors followed strictly Mendelian lines, a conclusion that was verified by the work of Bateson, Punnett, A. D. Darbishire, Doncaster, C. C. Hurst, and Cuenot on a variety of other organisms. This evidence convinced Castle that Mendel’s conclusions could be generalized throughout the animal and plant worlds.

Unlike William Bateson, however. Castle was by no means a complete convert to all the Mendelian assumptions. For example, Castle tended to answer the second question posed above in the negative. He believed that it was possible for two contrasting genes in a heterozygote to affect or contaminate each other in such a way that neither gene expressed itself in the same manner in subsequent generations. Thus, as a corollary. Castle maintained that it was possible to produce permanent blends of contrasting characters, a conclusion that was somewhat substantiated by his studies of inheritance of mammalian coat colors. Castle’s belief in the modifiability of gametes was also the basis of his view that permanent genetic change can be effected through selection.

In 1909 Castle, with the expert surgical help of his co-worker J. C. Phillips, showed dramatically the validity of Weismann’s distinction between germ and somatic tissues. Castle and Phillips transplanted the ovaries from a pure black guinea pig into a pure white guinea pig whose own ovaries had been removed. The transplant took, and in three subsequent breedings the female produced only black offspring. Thus, the genetic composition of the ovaries was unaffected by their being in the body of a different genetic type. This experiment represents one of Castle’s most significant and unequivocal contributions.

The question that dominated much of Castle’s life between 1900 and 1920 was the modifiability of Mendelian factors by selection. In June 1906 Hansford MacCurdy, along with Castle, completed a study of the inheritance of color patterns in rats. MacCurdy had shown that the darkly pigmented, uniform coloration pattern characteristic of wild rats was dominant over the piebald or “hooded” pattern. Hooded rats are white with a colored area (darkly pigmented) on the head and a narrow line down the back. A hooded rat crossed with a wild rat produced all wild F1. In the 2F the ratio of wild to hooded was 3:1. The hooded individuals, however, showed a much greater variability as a result of this cross. This suggested to MacCurdy and Castle that the recessive hooded gene was modified by existing in the hybrid along with genes for normal, wild pigmentation. Using this variability, Castle, with MacCurdy and J. C. Phillips, carried out a series of selection experiments on hooded rats between 1907 and 1914. In one line (called “plus”) Castle selected for an increase in the extent of the hooded pattern, and in another line(called “minus”) for a decrease in the hooded pattern. Selection was effective and ultimately yielded individuals far beyond the limits of the variability of the original parent series. After thirteen generations, selection in the “minus” strain produced almost solid white rats, while selection in the “plus” strain produced almost totally pigmented individuals. Castle wondered next whether such changes were permanent—i.e., whether they would disappear rapidly as soon as selection was relaxed. After an equivalent number of generations of backward selection (using sixth-generation “minus” strain rats) Castle found that the individuals produced were still much less highly pigmented than the parents from which he had originally started in the forward selection. In an important paper of 1914, “Piebald Rats and Selection,” he pointed out that selection permanently modified genes by bringing them to reside in germ cells with other genes, which contaminated them. Castle’s colleague at the Bussey Institution (the Harvard graduate school of biology at Jamaica Plain, Massachusetts), E. M. East, suggested that the concept of “multiple factor” might also account for these selection results. However, Castle rejected this notion on the ground that he had already obtained a nearly all-black race of rats without having witnessed any significant reduction in variability. He reasoned that if selection simply increased or decreased the number of modifying factors bearing on the extent of hoodedness, then he should expect to see the limit of variability decrease as selection in any one direction proceeded.

Castle’s conclusion, striking as it did at a fundamental assumption of Mendelian genetics (i.e., purity of the gametes), drew sharp criticism from T. H. Morgan’s group at Columbia, principally from A. H. Sturtevant and H. J. Muller. As a result of these criticisms Castle repeated his selection experiments between 1914 and 1919. They suggested to Castle that the hypothesis of modifying factors was the most consistent with all of his various results. In a paper of 1919 (“Piebald Rats and Selection; A Correction”) Castle conceded gracefully to the criticisms of the Morgan school. Characteristically, as L. C. Dunn remarks, Castle made the correction in a seminar attended by his students and colleagues, some of whom had also disagreed with him on this matter.

Castle’s concession put to final rest the assertion of certain workers in the late nineteenth and early twentieth centuries that selection itself created new and specific variability in the germ plasm. The Danish botanist Wilhelm Johannsen had challenged this idea in 1903 and 1909 with a series of selection experiments with plants. However, until the work of Castle and his colleagues the idea had not been sufficiently investigated in animals. Aside from confirming Johannsen’s results with animals, Castle’s experiments with hooded rats also gave strong support to the increasingly important multiple-factor hypothesis in genetics.

Also in 1919 Castle posed another challenge. Questioning the foundation of the chromosome theory—linear arrangements of genes and the process of genetic mapping—Castle argued (in Proceedings of the National Academy of Sciences) that a nonlinear, branching model was more in agreement with some of the breeding data. Sturtevant, Bridges, and Morgan replied heatedly that Castle had failed to understand the key to the linear theory: the phenomenon of double crossing-over. Ultimately he was forced to concede on this point as well. Nevertheless, Castle’s challenge forced the Drosophila school to reexamine its fundamental assumptions in light of the evidence. While Castle did indeed, at the time, seem to misunderstand the basis of chromosome mapping, others shared his confusion and were enlightened by the explanations necessary to resolve the controversy.

Castle’s work in the last forty years of his life was devoted to three problems: (1) the inheritance of quantitative characters, such as size, in mammals: (2) the construction of genetic maps for small mammals, such as the rat and the rabbit; and (3) the inheritance of coat colors in large mammals, such as the horse. Once he had understood and accepted the idea of linearity in chromosome structure, Castle carried out a number of breeding experiments to determine the genetic maps of several common mammals. His work and that of his student L. C. Dunn contributed significantly to the understand ing of mammalian genetics. Questions about the inheritance of quantitatively varying characters, such as size, had arisen considerably earlier in Castle’s career, especially in relation to his advocacy of blending inheritance. In work carried out between 1929 and 1934 Castle, in collaboration with P. W. Gregory, showed that difference in size between different races of rabbits was due not to permanent blends between parental factors for size but, rather, to different rates of development of the fertilized egg. This rate was clearly determined by a genetic factor (or factors) in the sperm and was inherited in a Mendelian fashion.

After his retirement Castle became research associate in mammalian genetics at the University of California, Berkeley. In these years, partly as a relief from the arduous breeding experiments, he became interested in the inheritance of coat colors in various breeds of horses. Data for his analyses could be obtained from studbooks, and his work on this subject contributed to a more sound knowledge of breeding lines in horses. Castle’s last published paper (1961)concerned the inheritance of coat pattern and gene interaction in the palomino breed. Significantly, this study brought to full circle Castle’s interest in mammalian coat color inheritance, a subject that he had begun investigating sixty years before.

Outside of his strict interest in experimental genetics, Castle maintained throughout his career a strong concern for eugenics, the application of breeding data to human inheritance. His first publication on this subject was Genetics and Eugenics (1916). In the 1920’s and 1930’s, as racial and immigration matters drew heavily on eugenic findings. Castle spoke out frequently. He argued strongly for objective evaluation of the roles of heredity and environment in drawing conclusions about human physical, psychological, or intellectual characteristics. Castle felt that there was an important place in biology for sound eugenic ideas but, like many biologists at the time, he believed that doctrines of racial superiority and segregation, based on supposed genetic facts, were unfounded and detrimental to social progress.

Through his lecture courses at Harvard, Castle introduced many undergraduates to the problems of genetics and evolution. His primary influence, however, was on graduate students; in twenty-eight years at the Bussey Institution he trained over twenty doctoral candidates. At the Bussey, Castle was responsible for mammalian genetics and E. M. East for plant genetics. However, as L. C. Dunn remarks, there was little formal division between the fields. Students of East’s could, and frequently did, confer with Castle and vice versa. Among Castle’s most important students were J. H. Detlefsen, C. C. Little, Sewall Wright, L. C. Dunn, Gregory Pincus, and P. W. Gregory. At the same time East trained such students as R. A. Emerson, Edgar Anderson, Paul Mangelsdorf, and E. R. Sears. Between them Castle and East had a significant and very widespread influence on American genetics.

Castle always enjoyed the individuality of research and opposed the strong organization of science. Similarly, he disliked bureaucracy and avoided it as much as possible. Warm and generous, Castle was interested in the individual in science, not the organization. His door was always open to his graduate students, but he spent little of his active career in any organizational capacity. Nevertheless, Castle took on administrative duties when necessary. For example, he helped to found the American Breeders Association (1903) andto reorganize it as the American Genetics Association (1913), and to found the joint section on genetics of the American Society of Zoologists and the Botanical Society of America (1922) and to reorganize it as the Genetics Society of America (1932). He also was vice-president of the American Genetics Association, chairman of the joint section on genetics (1924), vicepresident of the eastern branch of the American Society of Zoologists (1905–1906), and president of the American Society of Naturalists (1919). In addition, Castle was a member of the editorial board of the Journal of Experimental Zoology from its founding (1904) until his death. He also helped to found the Journal of Heredity (1913) as a new form of the American Breeders’ Magazine, and with ten colleagues he founded the journal Genetics in 1916.

Castle received a number of academic honors. He was elected a member of Phi Beta Kappa at Harvard (1893), the American Academy of Arts and Sciences (1900), the American Philosophical Society (1910), and the National Academy of Sciences (1915). In addition, he was awarded an Sc.D. by the University of Wisconsin and an LL.D. by Denison University in 1921, and the Kimber Genetics Award of the National Academy of Sciences, the first time this award was given (1955).

Castle was important to genetics in the twentieth century in three ways: (1) His strong support of Mendel’s work between 1900 and 1910 provided (in part) in America the same service that Bateson provided in England. While later Mendelians, such as T.H. Morgan, ultimately made more of the theory than Castle was able to do, his interest and work carried the young science through the early days of skepticism and doubt. (2) Castle helped to extend to mammals the conclusions that other Mendelians had reached with insects or plants. (3) Castle immediately saw, after 1900, the relationship between Mendel’s work and the Darwinian theory of selection. Although for many years he argued for what came to be a discredited view, his foresight did much to turn the attention of biologists to the importance of understanding natural selection in genetic terms.

In many respects, however, Castle’s work was out of the mainstream of genetics after 1915. He approached problems of heredity essentially through breeding experiments and never entered into the study of the physical basis of heredity, which led Morgan and his school to establish the chromo some theory. While this may have been to some extent a result of Castle’s long-standing association with the practical breeding tradition of the Bussey Institution, it was also a function of his personality. He had an aversion to complex experiments and abstractions. Breeding results were definite, while the theoretical constructions of the chromosome theory, brilliant as they were, seemed to him symbolic. He was frequently unable to see to the heart of an abstract matter and as a result often argued on the losing side of a controversy. Nevertheless, his practical common sense, enormous energy, and open-mindedness were important in establishing the Mendelian theory on a firm basis for a number of kinds of organisms.


Works by Castle include “On the Cell Lineage of the Ascidian Egg. A Preliminary Notice,” in Proceedings of the American Academy of Arts and Sciences, 30 (1894), 200– 216; Heredity of Coat Characters in Guinea-Pigs and Rabbits, Carnegie Institution of Washington Publication no. 23 (Washington, D.C., 1905); “Color Varieties of the Rabbit and of Other Rodents: Their Origin and Inheritance,” in Science, 26 (1907), 287–291; Selection and Cross-Breeding in Relation to the Inheritance of Coat Pigments and Coat Patterns in Rats and Guinea-Pigs, Carnegie Institution of Washington Publication no. 70 (Washington, D.C., 1907), written with H. MacCurdy; “A Successful Ovarian Transplantation in the Guinea-Pig, and Its Bearing on Problems of Genetics,” in Science, 30 (1909), 312–313, written with J. C. Phillips; “The Effect of Selection Upon Mendelian Characters Manifested in One Sex Only,” in Journal of Experimental Zoology, 8 (1910), 185–192; Piebald Rats and Selection: An Experimental Test of the Effectiveness of Selection and of the Theory of Gametic Purity in Mendelian Crosses, Carnegie Institution of Washington Publication no.195 (Washington, D.C., 1914), written with J. C. Phillips: Genetics and Eugenics (Cambridge, Mass.. 1916); “Is the Arrangement of the Genes in the Chromosomes Linear?,” in Proceedings of the National Academy of Sciences of the United States of America, 5 (1919), 25–32: “Piebald Rats and Selection, A Correction.” in American Naturalist. 53 (1919), 265–268; “The Embryological Basis of Size Inheritance in the Rabbit,” in Journal of Morphology, 48 (1929), 81–93, written with P. W. Gregory; “Race Mixture and Physical Disharmonies,” in Science, 71 (1930), 603– 606;“Further Studies on the Embryological Basis of Size In-heritance in the Rabbit,” in Journal of Experimental Zoology, 59 (1931), 199–210, written with P. W. Gregory; “The Genetics of Coat Color in Horses,” in Journal of Heredity, 31 (1940), 127–128: and “The Palomino Horse,” in Genetics. 46 (1961), 1143–1150, written with W. R. Singleton.

A more detailed biography is L. C. Dunn, “William Ernest Castle, October 25. 1867-June 3, 1962,” in Biographical Memoirs. National Academy of Sciences, 38 (1965), 31–80.

Garland E. Allen

Cite this article
Pick a style below, and copy the text for your bibliography.

  • MLA
  • Chicago
  • APA

"Castle, William Ernest." Complete Dictionary of Scientific Biography. . 23 Sep. 2018 <>.

"Castle, William Ernest." Complete Dictionary of Scientific Biography. . (September 23, 2018).

"Castle, William Ernest." Complete Dictionary of Scientific Biography. . Retrieved September 23, 2018 from

Learn more about citation styles

Citation styles gives you the ability to cite reference entries and articles according to common styles from the Modern Language Association (MLA), The Chicago Manual of Style, and the American Psychological Association (APA).

Within the “Cite this article” tool, pick a style to see how all available information looks when formatted according to that style. Then, copy and paste the text into your bibliography or works cited list.

Because each style has its own formatting nuances that evolve over time and not all information is available for every reference entry or article, cannot guarantee each citation it generates. Therefore, it’s best to use citations as a starting point before checking the style against your school or publication’s requirements and the most-recent information available at these sites:

Modern Language Association

The Chicago Manual of Style

American Psychological Association

  • Most online reference entries and articles do not have page numbers. Therefore, that information is unavailable for most content. However, the date of retrieval is often important. Refer to each style’s convention regarding the best way to format page numbers and retrieval dates.
  • In addition to the MLA, Chicago, and APA styles, your school, university, publication, or institution may have its own requirements for citations. Therefore, be sure to refer to those guidelines when editing your bibliography or works cited list.