780 



THE AMERICAN NATURALIST [Vol. XLII 



reasons the trichomes are favorable for the study of the 

 origin of plant structures, as I recently found while work- 

 ing on certain hybrids, a detailed account of the results 

 of which will be given in another place. 



The walnuts, to which reference is made, bear 4 or 5 

 types of multicellular hairs, besides certain abnormal and 

 one aberrant type. These are composed of 6, or 8 cells, 

 or about 16 or about 32 cells. A close study of the de- 

 velopment of the trichomes, in which mitotic figures were 

 used as indicators of the course of cell division, showed 

 the following to be facts: (1) In the earliest stages of 

 development of all of the normal trichomes, the sequence 

 of the first two, or three cell divisions was the same; 

 (2), the sequence of cell divisions of the 6-celled and 

 the 8-celled trichomes, during the entire development, is 

 consistent; (3), the 8-celled trichome recapitulates faith- 

 fully the sequence of cell divisions of the 6-celled type 

 up to the six-celled stage, and then adds two divisions in 

 an order not departed from. Certain facts indicated that 

 the late cell divisions of the two larger forms of trichomes, 

 namely, those with about 16 and about 32 cells, do not 

 hold to a sequence so closely, but further study of these 

 difficult trichomes might modify this conclusion. These 

 facts indicate that all of the multicellular trichomes may 

 have originated in a common ancestral form and that by 



