"In accordance with this law [of differentiation] we encounter no instances, _either in normal or pathological development_, of the transformation of a cell of one kind of tissue into a cell of another kind of tissue; and further we encounter no instances of a differentiated cell being transformed back into an undifferentiated cell of the embryonic type with varied potentialities."[12] Again, we have the following from one of the foremost pathologists, as to the strict and rather narrow limits of even pathologic change: "Epithelium and gland cells ... never become converted into bone or cartilage, or vice versa; while, again, it may be laid down that among epiblastic and hypoblastic tissues, on the one hand, and mesoblastic tissues on the other, there is no new development or _metaplasia_ of the most highly specialized tissues from less specialized tissues; a simple epithelium cannot in the vertebrate give rise to more complex glandular tissue, or to nerve cells; in regeneration of epithelium there is no new formation of hair roots or cutaneous glands. The cells of white fibrous connective tissue have not been seen to form striated or even non-striated muscle."[13] [Footnote 12: _Science_, March 29, 1901, p. 490.] [Footnote 13: J.G. Adami, "Principles of Pathology," pp. 641-642.] As implied by these quotations, a constant and progressive differentiation of cells prevails in the developing embryo; and when complete, certain groups of cells act as specialists in doing only certain kinds of work for the body. These cells maintain their specific characters in a very remarkable degree under normal conditions. Under various abnormal conditions, however, these cells may become modified as to functions, so that cells or tissues of one type may assume more or less completely the characters of another type. "But," as a very high authority declares, "the limitations in this change in type are strictly drawn, so that one type can assume only the characters of another which is closely related to it. This change of one form of closely related tissue into another is called _metaplasia_.... "When differentiation has advanced so that such distinct types of tissue have been formed as connective tissue, epithelium, muscle, nerve, _these do not again merge through metaplasia. There is no evidence that mesoblastic tissues can be converted into those of the epiblastic or hypoblastic type, or vice versa_."[14] [Footnote 14: Delafield and Prudden, "Text-Book of Pathology," pp. 62, 63.] This modification of function among the cells which sometimes goes on in the developing embryo, or under pathologic conditions, is very closely analogous to the variation which goes on among species of animals and plants. But, as we shall see later, there is a well marked limit to this variation among species, just as we see there is in the variations among the cells. Practically the same general laws hold good in each case.