514 ANNUAL REPOUT SMITHSONIAN INSTITUTION, 1908. 



and forms of the aspect of species which fully support all of the con- 

 clusions drawn from the observations on the evening -primroses. An 

 examination of the facts easily brought together allows us to see 

 that certain general principles in the organization of the plant and 

 in its behavior in these breaks or saltations in heredity may be 

 made out. 



The first and most important of these is one which was advanced by 

 De Vries speculatively before he began his experiments in heredity, 

 namely, that the plant is essentially a complex group of indivisible 

 unit characters. These unit characters may not always be expressed, 

 or recognizable in external anatomical characters, since they may 

 be in a latent condition, or totally inactive, or external taxonomic 

 characters may really consist of several elementary qualities, but 

 these are not shown in any intermediate stage, although they may 

 be modified within the limits of fluctuating variability. 



Any plant, supposedly, includes thousands of unit characters, and 

 as they are essentially qualities, or capacities, they do not usually 

 coincide with the characters ordinarily used in taxonomic descrip- 

 tions. As an illustration, the phases of geotropic sensibility of an 

 organ may be considered as a unit character. Thus a branch is either 

 apogeotropic, directing its tip directly upward, or it may be diageo- 

 tropic placing its axis in a horizontal plane, at right angles to the 

 action of gravitation, or it may undergo a mutation and " weep " or 

 direct its tips directly downward. In any case, however, it possesses 

 one of these three forms of reaction. It does not follow, however, 

 that all branches are actually in one of these three positions, for other 

 forces to which it reacts may operate to place the axes in various 

 planes, and the position of the branch may express just such con- 

 currence of elementary characters as alluded to above, or, indeed, the 

 geotropic unit character may be latent, and the organ may respond 

 to other forces exclusively. Similar analyses must be used in the 

 delineation of all unit characters, and it is obvious, without further 

 discussion, that we will never be able to uncover, either theoretically 

 or actually, more than a few of these indivisible units. The forms 

 and activities by which we recognize plants must by no means be 

 taken to be simple in their constitution imless proved to be so, and the 

 modification of a character in hybridization and otherwise must be 

 taken as proof that it is not an elementary feature. 



In tracing the development of species it is found practicable to des- 

 ignate them as retrogressive wdien a distinct unit capacity or unit 

 character is lost or, rather, becomes latent. Thus the loss of color 

 must be of this kind, and also the loss of geotropic reaction, while 

 the power of forming laciniate leaves when shown by mutants from 

 a simple-leaved type would be estimated as a progressive mutation, 

 as the group of characters concerned belong to a more highly organ- 



