the mesoderm would thus be brought to the surface and the ectoderm buried beneath it. The same effect might be obtained, however, by transplanting the mesoderm alone. While the difficulties in this procedure are great, they have now been in a large measure overcome. The positive experiments are too few in number to warrant any very definite statement, but they do indicate that it is immaterial which surface of the mesodermal disc faces laterally.

A much greater variety of experiments could be had by experimenting with positions intermediate between the upright and inverted positions, i.e., with limb buds turned, say, 90° instead of 180°. Such experiments may yield very interesting results, but as yet there has not been sufficient time to carry them out, nor has the effect of implanting the limb exactly in the midline been studied.

The experiments with superposed buds were made in the same way as the above, except that the mesoderm of the host was not excised. In the case of half buds, more combinations are possible, as described in the section dealing with this group. Both here and in the superposition experiments all possible positions with regard to the placement of the graft within the limitations stated above were experimented with. Relations of harmony and disharmony proved to be the same here as in the case of whole buds.

The total number of cases of which records have been kept is 462. The analysis is based, however, upon the 271 individuals which yielded positive results. The identity of the individual cases has been maintained by rearing each in a container by itself and keeping a separate history of each. These histories consist in notes and in sketches made from time to time directly from the living specimens, mostly with the aid of the camera lucida.

In dealing with so large a mass of material it has of course been necessary to select typical cases for presentation, and in order not to interrupt the continuity of the general account, the individual histories, as far as given, have been gathered together in an appendix. The main body of the paper has been divided in

accordance with the outline presented above. The larger groups of experiments have been considered apart from each other, and each subgroup is treated in a special section. The peculiar features of each of the larger groups have been considered at the beginning, and the results of the experiments summarized separately at the end of each main section. The more general questions are treated in the final chapter.

It has been thought best to provide numerous illustrations in order to avoid lengthy descriptions. Since it was not possible to keep a complete pictorial history of each case, those were selected for drawing that promised typical or otherwise interesting results. Unfortunately, however, it was not always possible to predict what the outcome of an experiment would be, so that some important cases were not drawn in early stages, while others of less interest were."

GENERAL FEATURES OF THE DEVELOPMENT OF THE

TRANSPLANTED BUDS

The development of the transplanted limb buds must now be considered in comparison with normal development. When the normal limb bud appears it is a round prominence just below the pronephros. It soon becomes more sharply marked off from the background and begins to 'point' dorsoposteriorly. The radial border of the fore arm and hand is at first ventrolateral, then ventral, and the first digits to arise are the first and second. The third and fourth digits appear later on the dorsal border of the hand, so that there is never any difficulty in distinguishing the ulnar from the radial border unless the third and fourth digits are entirely suppressed. The palmar surface of the hand faces at first ventromedially and later medially.

The transplanted limbs, both heterotopic and orthotopic, give evidence of their orientation early in development, inasmuch

7 Almost all of the preliminary sketches and many of the finished drawings were made by Miss Lisbeth Krause. The former, which were pencil sketches, had to be redrawn for reproduction. For this part of the work and also for a number of the original drawings I am indebted to Mr. A. Hemberger and Mr. H. D. Rhynedance.

* Harrison, '18, p. 419.

as their direction of 'pointing' is determined principally by the bud itself. In two of the combinations (homopleural dorsoventral and heteropleural dorsodorsal), they point anteriorly or dorsoanteriorly; in the other two (homopleural dorsodorsal and heteropleural dorsoventral) posteriorly or dorsoposteriorly like the normal. The subsequent development in the latter case is normal, but in the former there is a tendency for the limb to stick out more sharply to the side or to rotate more or less from the position in which it would be found were the position determined entirely by the orientation of the bud itself. Nevertheless, the palm tends to face ventromedially, or else the limb is so rotated that it faces more ventrally or anteriorly. In order to determine whether the limb is right or left, it is necessary to be able to distinguish between the palm and the back of the hand, which is not always so simple as it might seem. It can usually be done, however, by noting the digits, which are frequently' slightly flexed. When there is uncertainty, it is necessary to resort to sections, in which case there is no difficulty in distinguishing between the two faces, because of the much greater thickness of the soft parts on the flexor surface of the skeleton.

The duplicities that arise are of all grades and kinds, and occur in very different proportions in the several experiments. Sometimes they make their appearance very early, sometimes late in development. In the orthotopic grafts reduplication is far more common when the developing limb and the substratum are of opposite sides. In such cases the doubling member nearly always appears as a bud posterior to the main limb, growing there into a limb of proper asymmetry. The extent of reduplication may include the whole limb from the shoulder down, or only certain of the digits. The duplicate limb is as if it were mirrored from the original in a plane which is perpendicular to the plane of the proximodistal axes of the two limbs and which cuts the axes of the two limbs at their junction, at an angle which varies from almost 0° to 90°. In the former case the two members are almost parallel, in the latter they diverge in the opposite direction at almost 180°, the mirror plane bisecting the angle between them

Bateson, Materials for the Study of Variation, p. 479.

(fig. 3). In the present paper the relation of the mirror plane to the long axis of the limb has not been taken into account for purposes of description, the relation only to the dorsopalmar and the radioulnar axes being stated; i.e., the degree of divergence of the two members is not taken into account. Thus, when the mirror plane is parallel to the radioulnar axis, the limb is said to

Fig. 3 Diagram showing mode of reduplication. PR, primary limb; P.DU, posterior reduplicating member; A.DU, anterior reduplicating member; MP1(R), primary (radial) mirror plane; MP.(U), secondary (ulnar) mirror plane; 1 to 4, first to fourth digits, respectively. S, location of section shown in figure 4B. Dotted lines show the outlines of limbs as they would have been had there been no coalescence.

be mirrored in a palmar or a dorsal plane, according as the palms or the backs of the hand face one another; when the plane is parallel to the dorsopalmar axis, the mirroring is in a radial or an ulnar plane, according as the radial or ulnar borders of the limb face one another (fig. 4, A). Intermediate planes are described as radiodorsal, ulnopalmar, etc. (fig. 4, B). No attempt has been made for the present to measure accurately the angles of mirroring. It has been found, in agreement with Bateson, that

when there is a double reduplication, then the two mirror planes intersect at the bifurcation in a line perpendicular to the proximodistal axes; i.e., so that with reference to the radioulnar and dorsopalmar axes the planes of reflection face one another (fig. 4). Considerable deviation from this rule has, however, been noted in certain cases, and the amphibians do not seem to follow it with the same regularity as the arthropods, according to Bateson.10

Fig. 4 Diagram of reduplication, sectional view. In A the mirror planes are radial (MP1) and ulnar (MP2), and a certain amount of coalescence between the primary and the anterior reduplicating members is shown, as in figure 3. In B the mirror planes are radiodorsal (MP1) and ulno pulnar (MP2). D, dorsal; PAL, palmar; R, radial; U, ulnar.

EXPERIMENTAL

A. Limb buds implanted in abnormal location-heterotopic transplantations

In nearly all of the experiments in this group the limb bud was implanted on the flank of the embryo at the ventral border of the myotomes between the region of the fore and hind limbs. In a few cases it was placed on the side of the head between the eye and the ear, but the grafts were absorbed in all of these except

10 Op. cit., p. 552.