No. 504] ORIGIN OF STRUCTURES IN PLANTS 781 



arrested development, or other differentiation, the various 

 types of trichomes now to be found in Juglan's species 

 have arisen. Certain of the trichomes are evidently more 

 closely related to one another than to other types, and 

 thus the trichomes are not all of equal age, hut have been 

 derived from an ancestral form at various times in the 

 history of the plants which bear them. It is very prob- 

 able, for instance, that the 6-celled type is more nearly re- 

 lated to the 8-celled type of trichome, than it is to either 

 of the larger forms, but it would be difficult to say which 

 represents the more ancient type. In development, and 

 probably in origin, the types of trichomes thus behave 

 as if they were separate organisms, or independent units 

 of a complex organism. This is not the same as saying 

 that each type of trichome is a " unit character," al- 

 though certain observations which I have made on the dis- 

 tribution of trichomes in another plant, as well as the re- 

 version of the trichomes in the second generation of 

 Juglans calif ornica X Juglans regia would justify this 

 conclusion. Should it be the experience of other observers 

 also that each type of trichome has its peculiar area of 

 distribution in a plant, the conclusion that each form 

 of trichome represents a separate unit character could not 

 be avoided, and from such structural studies as above re- 

 ported we should be able to trace their very origin as 

 separate portions of the tissue of the plant. 



In addition to the normal types of hairs in walnuts, 

 as given above, there are also other types. Of such, there 

 are certain abnormal forms which are evidently related 

 to the already existing trichomes, of which they are slight 

 modifications, and one aberrant type which is essentially 

 different from these. The origin of the aberrant form 

 was seen also, and was found to be as different from the 

 mode of origin of the normal trichomes as the mature 

 aberrant type is different from the mature normal form. 

 In brief, its departure from the normal takes inception in 

 the orientation of the first cell wall, which is longitudinal 

 in place of being transverse as is usually the case (Fig. 



