No. 448.] NORTH AMERICAN CONIFERALES. 



than any species now existing, and that it is substantially 

 ancestral form of the genus, so far as we know. 



In Larix the four investigated species may be different^- 

 pretty fully, and this rule applies with particular force tc 

 amcricana, and L. occidentalis, both of which are distinguishec 

 a 2-seriate form. Among the pines, P. hwibcytiaiuu /'■ ^'l'" 

 P. sabiniana, P. tceda, P. palustris, and P. ciibnisis are rer 

 differentiated from the others by the 2-seriate i)its. In all ( 

 cases than those specifically indicated, the bordered pits a 

 an inadequate basis for specific differentiation. 



It is now apparent that segregated, round or oval i)its m 

 row miist be taken as representing the highest type ot tle\ ■ 

 ment in the Coniferales, and any deviation from this nuis' 

 taken to indicate the survival of more primitive condit 

 pointing to derivation from a type hke that of Araucaria or 

 daites. From this point of view, the occurrence of pits in 

 rows in Larix americana, Torreya taxifolia, Sequoia, 7'^?/<i,'v? 

 adensis and various species of Pinus, indicates the svu" 

 of ancestral characters which are partial to the extent ot 7. 

 and complete to the extent of 10.8 f. That such deviations 

 the usual type of structure are either survivals or revere 

 which serve to indicate a common origin, cannot be doti 

 more especially as they do not occur at a fixed point nea" 

 original type, but they arise sporadically in widely separ 

 genera. The tendency of such evidence then, 

 mon ancestry for the various genera 



he Tax 



Coniferae, a view which is greatly i 



crthened bv the tes 



afforded by the spiral tracheids of Lanx ami vicana, seiu o 

 and Pinus tcsda. The provisional conclusions which we 

 reach are, that there were probably four main lines of des 

 from the original stock represented by Cordaites : 



1 Araucaria and Agathis. 



2 Ginkgoales. 



3 Taxaceae. 



4 Coniferae. 



