268 



THE AMERICAN NATURALIST. [Vol. XXXVIII. 



•on former occasions (38, p. 5 i et seq.) In Araucaria, the three 

 species investigated may be similarly differentiated from one 

 another. The same rule is applicable to Tonrya taxifolia which 

 is thereby separable from the other species ; likewise to Cupress- 

 oxylon daivsoni, Tsuga canadensis, Larix americana, and, among 

 the pines, P. lambertiana, P. clansa, P. sabiniana, P. tceda, P. 

 palustris, and P. ciibensis. It is to be observed, however, that 

 the constancy which characterizes this feature in Cordaites and 

 Araucaria, is wanting in the higher Abietineai. In Larix there 

 is such variation that very careful scrutiny is required, while in 

 the genus Pinus, the number of exceptions to the typical charac- 

 ter increases greatly, and is liable to cause some difficulty in 

 the final determinations unless much care is exercised. Pinus 

 t(eda offers a conspicuous illustration of this fact, as may be seen 

 by reference to the analytical key. It is therefore manifest that 

 the value of the bordered pits for taxonomic purposes is most 

 clearly defined in the lower types of the Coniferales, and that 

 their value diminishes steadily, with an advance toward higher 

 forms of organization and development. In all cases where 

 exceptional forms introduce diagnostic difficulties, these may be 

 overcome by the controlling effect of associated characters. 



We are now in a position to examime the data at hand with a 

 view to determining the bearing of the bordered pits upon ques- 

 tions of phylogeny. 



Having reference to the origin of the bordered pit and the 

 various modifications it presents in the course of development, it 

 cannot be doubted that the hexagonal, multiseriate pits of Cor- 

 daites, Araucarioxylon, Araucaria and Agathis place those gen- 

 era m a relatively inferior position, a view which gains a large 

 measure of support from the well known and extensively multi- 

 senate disposition shown in Heteranginm gnevit (49, P- 340. 

 but the facts so far discussed have not as yet thrown any special 

 1 It I ^ the relative positions of the separate genera. 



I species of Cordaites shows that the 



i exammatic 



bordered pits exhibit 



uch wuder range of j 



any other genus covered by the present studies. If then 

 accept the genera] principle with respect to the development of 

 tne bordered pits as already illustrated, it cannot be doubted that 



