elements are often distinct from each other, muscle tendons and nerves are bifid, and in many cases the bones of the extra digit are more or less closely united to those of a normal toe. The question next arises as to whether these digital bifurcations are due to external influences or to internal variations of the germ plasm. Ahlfeld has observed that digital duplications may be caused 'in utero' by pressure from the thread-like outgrowths of the amnion. He attempted to make this explanation cover all cases of polydactvlism, but there are several serious objections. In the first place, the extra digits generally occur on both hands or on both feet, often on all four extremities (Fig. 1, A-D). The middle digits, moreover, arc not generally affected, but the duplication is chiefly of the first and fifth. Finally, and most important, the abnormality is strongly inherited and may increase in degree during successive generations. Thus Fackenheim cites the case of a woman born of normal parents. She had the little finger duplicated on each hand. Of two sons, one inherited the mother's extra fingers and the other had besides extra small-toes on both feet. Of eight children, three were normal, three had six toes and two had six fingers on both right and left extremities. In three succeeding generations the abnormality appeared, now on the hands, now on the feet, and in two cases on all four extremities; in two cases seven toes were present on both feet.
Similar observations have been made by Poulton and Torrey in families of cats. It is evident that extra digits produced by the chance pressure of amniotic threads would not be inherited, and that such chance pressure would certainly affect now one digit, now another; whereas, we have seen that the first and fifth digits are chiefly affected. Of twelve cases studied by the writer all were of the latter type.
It will be observed in Fig. 1, A-D, which represent the extremities of one child, that the fifth digit is affected differently in each case. In fact, it has been pointed out that occasionally no extra digit may be produced, that the first or fifth digit may simply be abnormally large. These facts, together with the frequent inheritance of the extra digits, show that we have to do here with variations of the germplasm. The first digit of man has been modified, and the fifth slightly reduced. Variation most often affects organs whose structure has been recently changed, and the variation or duplication of these digits might be naturally expected.
We are warranted, then, in assuming that the abnormal occurrence of six or seven digits on the five-toed extremity is not due to reversion. They are rather duplications of the normal digits, produced either by external influences or, more frequently, by germinal variation.
As the five-toed extremity is the primitive type of mammalian foot it is but natural to conclude that the appendage with less than five toes has lost some of the original number of evolutionary changes. The
