27 i THE AMERICAN NATURALIST. [Vol. XXXVIIL 



ical data. It will nevertheless serve a useful purpose at the 

 present moment, to ascertain the general sequence based upon 

 the percentage distribution of the principal variants as follows. 



Comparison of the principal variations in the serial arrange- 

 ment of bordered pits, by percentages. 



With respect to specific differentiations, it has already 

 appeared that the bordered pits may be employed with success 

 in Taxus and Torreya. In Cupressus this rule also applies to 

 C.pisifera, and C. macrocarpa, both of which are distinguished by 

 having their pits in one row or pairs, while the remaining seven 

 species have strictly uniseriate pits. An instructive example is 

 afforded by Cnpressoxylon daivsoni. In this species, which is of 

 early Tertiary age (Lignite Tertiary), the pits are typically 2- 

 seriate, being disposed in a very compact manner similar to that 

 found in existing Sequoias. But in a series of eleven speci- 

 mens, it is clearly seen that two variants are represented — the 

 second being a i -seriate form. These variations are also found, 

 as in the other Coniferales, to be directly related to variations in 

 the size and rate of growth of the tracheid. It cannot be 

 doubted then, that C. daiosoni is a more primitive representative 



