disappear long before birth. These indications, and similar ones, must have some meaning.
Now whatever qualities an animal exhibits after birth are attributed to heredity. May it not be that all the intermediate stages are also inheritances, and, therefore, represent phases in ancestral history? If they be, indeed, clues to ancestral conditions, may we not, by patching together our observations, be able to interpret the record, just as the history of ancient peoples has been made out from fragments in the shape of coins, vases, implements, hieroglyphic inscriptions, etc.?
The results of reflection in this direction led to the foundation of the recapitulation theory, according to which animals are supposed, in their individual development, to recapitulate to a considerable degree phases of their ancestral history. This is one of the widest generalizations of embryology. It was suggested in the writings of Von Baer and Louis Agassi z, but received its first clear and complete expression in 1863, in the work of Fritz Müller.
Although the course of events in development is a record, it is, at best, only a fragmentary and imperfect one. Many stages have been dropped out, others are unduly prolonged or abbreviated, or appear out of chronological order, and, besides this, some of the structures have arisen from adaptation of a particular organism to its conditions of development, and are, therefore, not ancestral at all, but, as it were, recent additions to the text. The interpretation becomes a difficult task which requires much balance of judgment and profound analysis.
The recapitulation theory was a dominant note in all Balfour's speculations, and in that of his contemporary and fellow-student, Marshall. It has received its most sweeping application in the works of Ernst Haeckel.
Widely spread through the recent literature is to be noted a reaction against the too wide and unreserved application of this doctrine. This is to be naturally expected, since it is the common tendency in all fields of scholarship, to demand a more critical estimate in research, and to undergo a reaction from the earlier crude an I sweeping conclusions.
Improvement in Methods.—Another feature of the work in Balfour's period was increasing attention to methods of preparing material for study. The great problem is to bring tissues under observation, with the normal relations as little disturbed as possible,* so that the prepared material will represent the conditions existing in life and no others. "Many of the most important elements of cellstructure are invisible in life, and can only be brought to view by means of suitable fixation, staining and clearing." One great danger is that pseudo-structures will be artificially formed by the action of reagents. On this account great attention has been given to every
