embryology he can understand much of what may be expected in this class of variations of structure.
We will turn now to the fifth and last of our headings, that of morpho-physics. It is only of recent years that methods of experimentation, as distinguished from methods of observation only, have been applied to embryological problems. Naturally under the circumstances many crude experiments have been undertaken, many of doubtful validity; but there have also been many others, soundly planned, rightly executed and fruitful in results. Already the new conclusions constitute an increment both large and precious to our stock of embryological knowledge. One important class of these experiments has been based upon the discovery of the possibility of grafting parts of amphibian embryos on to one another; or to get two large pieces of two distinct embryos, or even two halves of two embryos, to grow together. The grafting experiments which have already been made are very numerous. Let me present one or two examples of the sort of results that these experiments yield. If the halves of two species of frog in a very early stage are grafted together, they will unite perfectly, but it is found that the epidermis of the species which forms the anterior half of the graft will spread to a certain extent over the posterior half, thus showing that the skin can actually crawl over the underlying tissues. It is probable, indeed, that the migration of epithelial cells along the surfaces upon which they rest is a very general phenomenon, and plays a very important part in the animal economy. In another series of experiments the embryonic optic vesicle has been removed and grafted on to a new part of the larva. Where the optic vesicle comes in contact with the epidermis it causes the epidermis to form a typical lens for the eye. Thus it is proved that the formation of the lens is not a specific function of that part of the epidermis from which it is normally produced, but is a potential function of the entire embryonic epidermis which may be called forth into activity by contact with the growing optic vesicle. I believe that we have in this an illustration of one of the fundamental principles of the establishment of structure and that much depends upon the interaction and mutual stimulation of parts.
Another class of experiments has been conducted by those who have been somewhat jocosely named the 'egg shakers.' An egg during an early stage of segmentation is divided artificially into its natural segments, or into groups of such segments, as the case may be. In many cases this division can be accomplished by shaking the eggs somewhat violently so as to break the segmentation spheres apart; hence the name above quoted. Now it has been demonstrated that in some cases fragments of a single egg will develop into an embryo perfect apparently in structure, though only of say half the normal size,
