the melanic series, the grade of intensity of hair pigmentation in the offspring does not exceed that of the darker parent. The only exceptions appear in the "brown" and "dark brown" parentage, where a small percentage of children are represented as of the next darker grade. Many more such cases were in our original records, but wherever the question was asked of the recorder whether the hair of the child, 4, exceeded in darkness that of the darker parent, B, the reply was almost without exception negative. Samples of hair were asked for and these never proved darker in the children than in the parents. A common source of error lies in not disconnecting the effect of a slight grayness in the parent. In one sample of hair from a mother that was reported lighter than the daughter the gray hairs were carefully picked out, when it appeared that the natural color of the hairs of the mother and child were as like as possible. Consequently one is justified in laying no stress on the 11 children out of 600 (less than 2 per cent.) in which the hair color was returned as darker than that of the parents particularly as despite efforts these returns could not be confirmed. ents may be assured that their children will eventually have hair as dark as the darker parent or of a lighter tint, but not darker. Consequently, parents with flaxen or yellow hair will have children all alike and like themselves in this respect. But parents with black hair may I have children with flaxen hair or with light brown hair

It follows, then, that par

N.

Red Series.

or (because of the masking qualities of black pigment) with red hair.

What is the relation of this principle to the law of alternative inheritance? The latter is only a special case of it. When characters A and B are crossed the more intense character appears in the offspring-the less intense character is recessive-the "heterozygous" children do not exceed the more intense parent. If, now, two such heterozygous persons be mated, one fourth of their offspring show the recessive condition, which by hypothesis is of a lower grade than that possessed by the parents; the remainder of the offspring may attain the grade of their parents; but they will not exceed that grade. This principle of the non-transgressibility of the upper alternative inheritance only, but also for blending inheritance-indeed, it seems to be of universal applicability.

An exception to this rule is exhibited by some heterozygous forms. The cross of a high-combed fowl and a low-combed is a fowl with one intermediate grade of comb. Two heterozygous combs in the parentage throw, inter alia, high combs. Not all cases of heterozygous forms constitute exceptions to this law of the non-transgressibility of the upper limit, and human hair color seems, even in the heterozygous condition, to follow the law. The workings of the principle are veiled in some cases of cryptomeric characters, i. e., built up of hidden factors.

Certain important consequences flow from this principle. These one of us has pointed out in a brief communication to Science.3 If the progeny stands on the average in respect to a character at a lower grade than the parents then, if inbreeding is practised, the two parents of the next generation will probably have this character at a lower level than their parents and will produce children having the character less well developed than they have it themselves. If inbreeding be practised for several generations it is clear that in some, at least. Vol. XXVIII. pp. 454-455, October 2, 1908.

of the children, the character in question would probably become quite degenerate. Since the note in Science was published we have read a paper by Feer, to which our attention was called by the title "Der Einfluss der Blutverwandschaft der Eltern auf die Kinder." This paper comes very near to our point of view. After showing that retinitis pigmentosa and congenital deaf-dumbness are the diseases most closely associated with inbreeding the author concludes that they are not so much inheritable diseases in the usual sense as inheritable diseases of degeneration, and depend on degeneration of the embryonic ectoderm. It seems clear from such data as Feer adduces that our general thesis will hold true for many human characters-that inbreeding does not cause them to degenerate, but having a tendency to degenerate, inbreeding will prevent any recovery and, in addition, will hasten the downward tendency from generation to generation. The only way to avoid progressive degeneration is to bring in (usually necessarily from outside) blood with the tendency to produce the characteristic in a welldeveloped condition.

Combining now the results of the three studies on eye color and hair color and form made by us," it appears that two parents with clear blue eyes and yellow or flaxen straight hair can have children only of the same type, no matter what the grandparental characteristics were; that dark eyed and haired, curly-haired parents may have children like themselves but also of the less developed condition. In the latter case what the proportions of each type will be is, for a fairly large family, predictable by a study of the immediate ancestry.

4

CARNEGIE INSTITUTION OF WASHINGTON,

DEPARTMENT OF EXPERIMENTAL EVOLUTION,
COLD SPRING HARBOR, N. Y.,

December 1, 1908.

Separate, Berlin, 1907; also "Jahrbuch für Kinderheilkunde," Bd.

LXVI.

G. C. and C. B. Davenport, "Heredity of Eye-Color in Man," Science, N. S., XXVI, pp. 589-592, November 1, 1907; "Heredity of Hair-Form in Man,' ," AMERICAN NATURALIST, XLII, pp. 341–349, May, 1908.

A MECHANISM FOR ORGANIC CORRELATION1

G. H. PARKER

PROFESSOR OF ZOOLOGY, HARVARD UNIVERSITY

THE year 1909 is notable for its many historical associations. It is not only the fiftieth anniversary of the publication of "The Origin of Species," but it is also the centenary of the birth of Charles Darwin and of the publication of Lamarck's "Philosophie Zoologique." To the American its associations with Lincoln are precious memories. But it is not to these historical matters that I wish to refer. Science ever looks forward, not backward, and it is on certain modern aspects of the movements centering about the problem of evolution and especially on those connected with the name of Darwin that I wish to speak.

Although biologists have been familiar with Darwin's theory of natural selection for almost fifty years, it must be confessed that they are only at the threshold of the problem of evolution. That species have arisen by transmutation is now universally admitted, but how transmutation has been accomplished remains at present one of the unsolved riddles. The Lamarckian factors, though possible, must be set down as still unproved. Natural selection, so far as observation and experiment go, seems to play a real part in transmutation, but the extent of its application is still a matter of much uncertainty. Even the recently advanced mutation theory, on which hopes at one time ran high, is coming to assume at best a supplementary rôle. In fact it is evident that the most serious efforts of the past have failed of full accomplishment and it seems likely that the process of transformation is not exclusively dependent upon any single principle, but is of great complexity involving in all probability a consid

1Read before the Boston Society of Natural History, February 12, 1909.

erable number of factors. Of these factors we are only beginning to get glimpses. I believe they will come into clearer view only as we progress in the solution of general biological problems. It is my intention to bring before you very briefly one of these lines of progress and to point out its possible bearing on the problem of evolution. You are all doubtless familiar with the claim that Darwinism or natural selection is at best only a partial or insufficient factor in evolution. Its actual workings seem to be concerned with the elimination of only the most poorly adapted members of any stock; it is a process that is not closely enough adjusted to call forth those slight but constant differences which every systematist recognizes as the distinguishing marks of a species. To quote from a recent criticism:

Every student of systematic zoology or botany has a keen realization

of the fact that a majority of the distinguishing characters which he recognizes in the various species . . . that come under his eye are of a sort that reveal to him no trace of particular utility.

I

Ac

For this reason it is believed that these characters could not have been produced through natural selection. hope to show you, however, that we can make the admission implied in this quotation, to the effect that specific characters are not necessarily useful, and still be able to explain their occurrence and fixity through Darwinism. A general outline of this proposition has already been given by Plate in his consideration of correlation. cording to this author the development of a specific character of no special use may take place through correlation, that is, through that unknown law of growth by which an indifferent organ may be so bound up with or related to a useful organ that it, the indifferent organ, is perfected along with the useful organ as this latter is it is conceivable that a specific character, even though developed or specialized through selection. In this way useless, may arise at least indirectly through natural

selection.

It is to be noted that Plate's conception of

the mechanism of correlation is not detailed; in fact, he