[pENHALLow] N. AMERICAN TAXACE^ AND CONIFERiE 41 



somewhat widely under ditt'erent conditions of growth and even in differ- 

 ent i^arts of the same tree, and these variations are of such a nature that 

 it would be quite possible for the wood in a branch of Torreya to present 

 much the same aspect as wood taken from a stem of Taxus. With these 

 considerations in mind, it becomes possible to construct a differential key 

 for these two genera, and this, together with a systematic diagnosis of 

 the family and of each genus, will appear in the following pages. 



The Taxaceœ and Coniferœ possess a number of structural features 

 in common. These are to be found first in the transverse section, in the 

 usually regularly radial disposition of the tracheids. In the radial sec- 

 tion the radial walls of the tracheids of both the spring and summer wood 

 are marked by the presence of conspicuous bordered pits. In the Taxa- 

 ceœ these structures are relatively small, and always in one row, generally 

 occupying the full width of the narrow tracheids. In the Coniferœ, on 

 the other hand, they are, with the exception of Juniperus, usually large 

 and oval or round, and not infrequently two- or three-seriate. In both 

 families bordered pits occur on the tangential walls of the summer wood, 

 and in a very few cases on the tangential walls of the spring wood of 

 certain Coniferœ. 



Apart from the details already considered as differentiating these 

 two families, there are few anatomical features which belong distinctively 

 to the Coniferœ, and they are to be regarded as of subordinate value. 

 Thus in transverse section the tracheids, except in Juniperus, are, as a 

 rule, much larger, and there is often a more marked and abrupt contrast 

 between the spring and summer woods. In the radial section the Con- 

 iferœ commonly show Sanio's bands, which are wholly wanting in the 

 Taxaceœ, so far as it is possible to determine from our present investi- 

 gations. 



Anatomical considerations show that the sequence of genera and 

 also the limitations of those groups, as defined on the basis of general 

 morphology, require some readjustment. It will, therefore, be desirable 

 to consider somewhat in detail the various points of affinit}'' which justify 

 the arrangement embodied in the present treatise. 



The four genera Libocedrns, Cupressits, Thuya and Juniperus fall 

 into a natural grouj), of which the common characteristics are the pre- 

 sence of more or less numerous resin-cells, the chiefly simple pits on the 

 lateral walls of the ray-cells, the thin or sparingly pitted terminal walls 

 of the ray-cells, and the absence of resin-joassages. A more critical 

 examination of the distribution of the resin-cells shows that Libocedrus 

 and Juniperus approach one another somewhat closely in the fact that 

 these elements are disposed in tangential bands, while in both Thuya and 

 Cupressus they are commonly scattering and often appear only in somewhat 

 distant growth-rings. The affinity between the first two genera is also 

 greatly strengthened by the great similarity of the terminal walls of the 



