No. 448.] 



NORTH AMERICAN CONIFERALES. 



in the later growth and are wholly wanting in the two or three, 

 last formed tracheids. In T. brevifolia they become very imper- 

 fect in the outer summer wood and tend to disappear completely, 

 only vestiges remaining in the last formed tracheids. In T. chs- 

 pidata the spirals are generally absent from the summer wood, 

 or when present, they are merely vestigial. The angle is some- 

 what greater —about 7 deg. — than in Torreya, and this fad is 

 apparent with respect to certain species without special measure- 

 ment. The four species appear to be paired off in such a wa\ 

 as to represent a mean difference of about 10.9 deg. as between 

 T. canadensis and T.floridana on the one hand, T. hrcvifolia and 

 T. cuspidata on the other. In all cases the angles of the spirals 

 are quite distinct from those of the lines of striation. The fol- 

 lowing details are based upon the average of ten determinations. 



A comparison of these results in detail emphasizes the fact 

 that the distribution of the spirals, as between spring and sum- 

 mer wood, is in direct harmony with the principles already stated, 

 and furthermore, that the angles at which the spirals develop do 

 not afford an adequate basis for generic differentiation. It is 

 nevertheless possible to recognize sub-generic groups in such 

 wise that in both genera a general line of division may be estab- 

 lished at 70 degrees. In the case of Torreya californica, the 

 very low angle of 46.2 degrees may be regarded as a differential 

 character of specific value. 



In the genus Pseudotsuga, spirals are confined to the tra- 

 cheids of the spring wood. This has a partial exception m 

 P. macrocarpa, in which vestigial spirals may be observed in the 

 earlier tracheids of the summer wood. In this species the mean 



