A. STATEMENT OF PROBLEM.

In man, various mammals, and some birds two or more adjacent fingers are sometimes intimately connected by an extension of the web that is normally a mere rudiment at their base. Such a condition is known as syndactylism. A good introductory account of syndactylism is given by Bateson (1904, pp. 356-358). Taking a number of cases of syndactylism together, he says: "A progressive series may be arranged showing every condition, beginning from an imperfect webbing together of the proximal phalanges to the state in which two digits are intimately united even in their bones, and perhaps even to the condition in which two digits are represented by a single digit." He also calls attention to the fact that in the human hand "there is a considerable preponderance of cases of union between the digits iii and iv;" while in the foot the united digits "are nearly always ii and iii." The matter of syndactylism in birds has a peculiar interest because of the fact that among wading and swimming birds syndactylism has become a normal condition of the feet, and, moreover, just this feature is one that has become classical in evolutionary history, because Lamarck thought it well illustrated his idea of the origin of an organ by effort and use.

Concerning the cause of syndactylism little can be said. Both in mammals and birds the digits are indicated before they are freed from lateral tissue connections. The linear development of the fingers is in part accompanied by a cutting back of this primordial web, in part by a growth beyond it. In syndactylism growth of the web keeps pace with that of the fingers. From this point of view syndactylism may be regarded as due to a peculiar excessive development of the web.[5] In some human cases adhesions of the apex of the appendage to the embryonic membranes has stimulated the growth of the interdigital membrane, resulting in syndactylism. But it would be absurd to attempt to explain syndactylism in general on this ground. The more "normal" forms of syndactylism, as seen in poultry, still want for a causal explanation.

Most of the cases of syndactylism whose inheritance is about to be described arose in a single strain of fowl and can, indeed, be traced back to a single bird. This ancestor is No. 121, a Dark Brahma hen described in a previous report.[6] It was only in the search for the origin of the exaggerated forms of syndactylism observed in some of her descendants that an unusually great extension of the web in her feet was noticed. The syndactyl condition of my birds did not, thus, arise de novo, but had its origin antecedent to the beginning of the breeding experiments. In addition to this main strain a slight degree of syndactylism has appeared among some of my Cochin bantams.

The types of syndactylism which have appeared in my flock form a rather extensive series. First, (A) the single web, which, in my specimens, always occupies the interspace between digits iii and iv. This is the same interval which is most apt to show the web in syndactylism of the human hand, and, it is suggestive to note, it is this interval that is filled in those wading birds that have the single web only between the toes (e.g., Cursorius, Glareola, Vanellus, Squatarola, Charadrius, Limosa, Machetes, Himantopus); second, there is (B) the double web, one-seventh as common, which always occupies the interspaces between the digits ii-iii and iii-iv.

On another basis, the syndactyl feet may be classified as: (a) toes adherent, web small in extent, and (b) toes distant, web broad. I have found the narrow web only between digits iii and iv. It is one-eighth as common as the broad-webbed type. The broad, double web approaches closely to the type found normally in swans, geese, and ducks.

Finally, the syndactyl feet may be classified as: a, straight-toed, or , curve-toed. Class a is to class in frequency as 2:1. In the typical curve-toed syndactyl foot the web between iii and iv is complete to the nails of each; in fact, in extreme cases the nails of the two toes are more or less fused together. From the fused nails the middle toe, being the longer, passes in a curve to the distal end of the metatarsus. The D-shaped interspace between the curved iii and straight iv toe is filled with the web. In other cases the nails are merely approximated and the middle toe is slightly curved. In three instances (4 per cent of all) the outer toe (iv) is curved toward the (straight) median toe (class ´).

As stated, the polydactyl offspring trace back their ancestry to No. 121; her feet both show the double, broad, straight-toed type (Bba). We shall attempt in the following paragraphs to trace the heredity of her type of polydactylism and of the others that have subsequently arisen.

B. RESULTS OF HYBRIDIZATION.

In taking up the results of breeding experiments to test the method of inheritance of syndactylism, it will be best first to give in a table all pens in which the character showed itself, with the frequency of the different types of foot in them (table 23).

The history of the syndactyl strain begins with No. 121 ? and in the matings 1 to 8 are given the results of crossing together some of her progeny derived from a normal-toed father. This father was either No. 8a or 1a, both full-blooded Tosa (Japanese Game) fowl and without suspicion in either soma or offspring of syndactyl taint. There is no record of trace of syndactylism in the progeny of 121 × 8a (or 1a); but a slightly developed condition of syndactylism may very well have been overlooked by me in this F₁ generation (as I had never thought of such an abnormality), even as I at first overlooked the syndactylism visible in No. 121. But when these F₁ hybrids were mated together (pen 627, serial Nos. 1 to 8) I got, in the different families, from 10 per cent syndactyl offspring down to none at all.

At first sight the suggestion arises that, if inheritance is at all Mendelian, the normal condition is dominant and that the heterozygotes throw again, in pen 627, the syndactylous condition. If this hypothesis were true it would follow that syndactyls bred together should, sometimes at least, throw, even in large families, 100 per cent syndactyl offspring. But only 2 families, Nos. 30 and 34, have yielded 100 per cent syndactyls, and these contained 2 and 1 offspring, respectively; so they are not significant. On the other hand, there are numerous matings of 2 extracted normal-toed parents that have produced only normal-toed offspring (families Nos. 14, 15, 21, 22, 23, including 119 individuals). Consequently the conclusion is favored that normal-foot is recessive and syndactyl-foot dominant, and this shall be our working hypothesis.

On our hypothesis, No. 121 is probably a heterozygote. Mated with the recessive normal, expectation is 50 per cent heterozygous, showing syndactylism; the remainder normal-toed. But dominance is here, as in polydactylism, very imperfect. For this reason and because it was not looked for, no syndactylism was noted in the first hybrid generation. The offspring prove to be of two sorts, however. No. 180 ? is a pure recessive, and in 8 matings with as many different sisters of his he got 184 normal-toed to 1 syndactyl. These same sisters, mated to another brother, No. 242, in some cases gave 9 per cent and 10 per cent syndactyl. No. 242 is, consequently, probably a DR and, mated to DR sisters (which constitute according to expectation about one-half of all) gives some DD's, part of which constitute the 9 to 10 per cent of syndactyls. Of course, 25 per cent DD is to be expected; the difference gives a measure in this instance of the imperfection of dominance in the "extracted" as well as "heterozygous" condition.

Matings 9 to 15 (pen 747) are instructive in comparison with the foregoing case. Both parents are derived from pen 658, which contained as breeders a heterozygous Dark Brahma male (No. 146) and various females of non-booted races far removed from suspicion of syndactylism; expectation being an equal number of DR and RR offspring. In pen 747 No. 1888 ? acts like a DR, and so do the hens in matings 9 to 13, while the hens in the other 2 matings are doubtless RR's. The former give 17 per cent syndactyl offspring, the latter none at all (in 56 individuals).

Matings 16 and 17 (pen 703) are between pure-bred Dark Brahmas that are probably DR's. About 22 per cent of their offspring are syndactyl—a rather higher proportion than we have found before. Matings 18 to 19 are between progeny of pen 627. In mating 20 the normals were not recorded. The cock in this pen, No. 871, is probably heterozygous, as are also the first two hens, so that nearly 30 per cent of their progeny are syndactyl. From the other 3 hens no syndactyl offspring were obtained. Evidently the two sets of hens have a very different gametic constitution. The existence of two sorts of families is one of the strong arguments for the segregation of this character.

We next come to the pens (matings Nos. 24 to 42) which were especially mated to study the inheritance of syndactylism. I had now, for the first time, two parents with syndactylic feet. On account of imperfection of dominance decision as to gametic composition of any parent must largely rest on the make-up of the progeny. Table 24 gives the most reasonable classification of the parentages.

Summarizing the foregoing, and comparing the totals with Mendelian expectation, we get the result shown in table 25.

A comparison of realization and expectation in table 25 shows that the proportion of syndactyls is always less than expectation, not only for dominants and heterozygotes together, but even for pure dominants alone. The proportion of syndactyls obtained diminishes, to be sure, in accordance with expectation (on the assumption that they are pure dominants), but the numbers lag behind, in the higher proportions 40 to 25 per cent. So we reach the conclusion that, as in polydactylism, so in syndactylism dominance is very imperfect. But there is this difference, that in syndactylism dominance is so imperfect that the dominant condition rarely shows itself in heterozygotes and even fails in many pure dominants. The striking fact, the one that assures us the segregation is nevertheless occurring in this case too, is that some families (whose two parents are extracted recessives) throw 100 per cent recessives.

These studies on syndactylism in poultry may be used for a critical examination of the recent work of Lewis and Embleton (1908) on syndactylism in man. The cases described by them follow the types I have just described in poultry. Their fig. 18 corresponds to my types a and a; figs. 10 and 11 to my type . The "crossbones" referred to by the authors correspond to bones of the "curved toe." The facts presented by the authors support the idea that syndactylism is dominant rather than recessive, but they deny the application of Mendelian principles to this case. Actually, the foot deformities described by Lewis and Embleton are inherited much like syndactylism in poultry. No extracted normal (recessive) extremity produces the abnormal condition. Heterozygotes show much variation, from very abnormal to slightly abnormal (possibly perfectly normal?) appendages. Dominance is, indeed, much more potent than in poultry.

The authors' denial of the application of Mendelism to this case seems to be based on an all too superficial consideration of the hereditary behavior of the character and a tendency to "mass" statistics—a procedure that tends to obscure the interpretation of the data of heredity.

As to the inheritance of type, my statistics are not extensive enough to give a final answer, but if all types be grouped into those with straight and those with curved toes, then in crosses of straight-toed syndactyl and normal 33 per cent of the offspring were of the curved type, whereas in crosses of curved-toed syndactyls and normal 45 per cent were of the curved type. These averages depend on 22 and 15 individuals, respectively. They lead us to look for an inheritance of type when more extensive data shall be available.

Syndactylism is a typical sport, that is, a rather large mutation having a teratological aspect. The question arises, Does it prove to be prejudicial to the welfare of the species? The breeder who has only a few individuals of a rare sport feels their loss more than that of normals and the general impression left in his mind is that the sport is less capable of maintaining itself than the normal form. Assembling the data, consisting of about 40 individuals of each kind, it appears that the death-rate is not very different in the two lots; the slight excess of that of the syndactyls is sufficiently accounted for by the circumstance that no normals were reared during the period of greatest mortality (the summer), but were destroyed or given away as soon as hatched. It is probable, therefore, that syndactylism, under the conditions of the poultry-yard, has little life and death significance, but is one of those neutral characters whose existence Darwin clearly recognized.