10 DEVELOPMENT OF SIUM CICUTAEFOLIUM. 



leaves, and physiological causes may determine which atavistic form 

 shall appear and may in addition give rise to forms or details of form 

 which are in no sense atavistic. 



A point in favor of the interpretation of the tripartite leaf as being at 

 least in a general way atavistic may be found in the fact that ternate 

 division of leaves is notably prevalent among the Umbelliferae ; but it 

 should not be forgotten that a ternate condition is a necessary logical 

 transitional stage between any " simple " leaf and a pinnate leaf, and its 

 significance must rest solely upon this logical relation. To show how- 

 gratuitous is any attempt to draw from the juvenile leaves conclusions 

 regarding the leaf- form in ancestral groups it is only necessary to inves- 

 tigate the conditions found in the few instances in which the ancestry 

 of a species is definitely known. Through the kindness of Dr. D. T. 

 MacDougal, of the New York Botanical Garden, I am enabled to pre- 

 sent photographs of the seedlings and adult rosettes of Onagra laniarck- 

 iana and O. rubrinervis, the former species being the parent of tlie 

 latter. They represent the closest relationship possible between two 

 species, and yet those who are experienced in their culture separate 

 them with unfailing accuracy in the earliest juvenile stages, A com- 

 parison of the seedlings and the adults in Plates IV to VI makes evident 

 the fact that the seedling of Onagra rubrinervis resembles much more 

 closely the adult condition of the same species than either the seedling 

 or the adult condition of its parent species, Onagra laviarckiana. 



Turning now to other regions which have been supposed to present 

 ancestral types, particularly to the two regions most strongly exploited 

 by Cushman (1902, 1903, 1904) in his recent papers in the American 

 Naturalist, we find conditions which are totally different from those we 

 have just considered in the juvenile stages. The first of these regions is 

 that at which a perennial herb resumes growth in the spring after hiber- 

 nation. At the close of the first season from seed the leaflets of Sium 

 cicutaefolimn become more or less dissected into narrow segments, as 

 shown in fig. 2. This modification is very often apparently related to 

 the more aquatic condition of the habitat, and is in line with modifica- 

 tions in many other species whose dissected leaves appear to be corre- 

 lated in some way with aquatic conditions ; but it can be shown that this 

 change in the form of the leaf of Sinm takes place, though less per- 

 fectly, when the plants are not supplied with an unusual quantity of 

 water. Plants taken up as seedlings in June and grown under meso- 

 phytic conditions in the window of a laboratory at the University of 

 Chicago and later in the conservatory of the same showed dissection of 

 the leaves similar to that exhibited by plants that were submerged in 

 imitation of the effects of autumn rains. 



