CLASSIFICATION OF THE PRIMORDIA 91 



method expounded in § 74 ought to be followed. Leaves between 

 which social equivalence exists ought to be taken in successive 

 states ; and, moreover, the method of the curves of develop- 

 ment ought to be appHed. (See § 49.) We may investigate in a 

 similar way the embryology of the internodes, the scales, the 

 trichoms, etc. 



We might be tempted to say, for instance, that the terminal 

 shoot of the main stem produced in the course of the fourth 

 year represents a further state of development of a tree than 

 the terminal shoot produced in the course of the third year. 

 This would be, in reality, inexact. The successive yearly seg- 

 ments do not coincide with successive stages of development. 

 These segments (shoots) are individuals of a certain order 

 (zl + I ; see § 74), which are produced successively by means of 

 buds ; we may call them hud-generations. The successive bud- 

 generations are superposed, as if they were grafted on each 

 other ; this results in social relations. Therefore the differ- 

 ences between the successive yearly (terminal) shoots of a main 

 stem ^ are not embryological but social differences. (See 

 gradation. Part VIIL) 



Cuttings : We may take, for instance, the terminal shoot 

 of a main stem as a cutting and cultivate it separately ; the 

 social relations with the main stem being suppressed, the next 

 bud-generation produced by the cutting is more or less different 

 from the bud-generation which would have been produced if the 

 previous relations had been preserved.^ In certain cases the 

 produced modification is very distinct. 



Grafts : In a similar way a terminal (or any other) shoot 

 ^ + 1 of a specimen x may be grafted upon another specimen 

 x^ ; new social relations are brought about and in the next bud- 

 generation produced by ^ + i the properties may be modified. 

 If the material is strictly monotypic — in other words, if there is 

 no specific difference between the specimens x and x^ — the 

 observed modifications are a consequence of the social changes. 

 But if X and x^ belong to different species ^ the observed changes 

 depend not only on social causes, but also on a specific influ- 



1 Such differences always exist. In many species they are very apparent ; 

 for instance, with reference to the properties of the leaves {Cecropia, Morus 

 nigra, Symphoricarpus racemosa, many Leguminosce with compound leaves, 

 etc.). 



2 Suppose the cutting is the terminal shoot of a main stem developed in the 

 course of the tenth year. It may happen that the next shoot (produced by 

 the cutting) has leaves similar to those produced by the tree in its first, second 

 or third year. Starting from the erroneous idea that the tenth year represents 

 a further state of development than the first, second or third years, and from 

 the principle that the successive yearly shoots coincide with successive states 

 of the phylogenetic evolution, we are tempted to say that the cutting^ shows 

 a reversion to an ancestral state. We are deceived, I think, by an illusory 

 appearance. 



" Example : Pyrus cydonia grafted upon Cratagus monogyna. 



