12 ANATOMY OF THE GYMNOSPERMS 



fact that these species are among the most highly specialized 

 representatives of the entire phylum, and that they therefore 

 stand as representatives of the highest order of development. 

 The sources of error having been determined by such tests, cor- 

 rections were applied to the key in such a way as to eliminate 

 them, while the original diagnoses were further modified to meet 

 the special requirements. It is manifestly impossible to construct 

 a key capable of providing for all exceptional cases. These can 

 only be met by the experience of the observer or by final refer- 

 ence to and critical comparison with type specimens. But the 

 experience so far gained justifies the belief that as now presented 

 the key affords sufficient data for the recognition of species under 

 all ordinary circumstances, and there is no reason for hesitation 

 in stating that it is fully as efficient in this respect as the keys 

 usually employed for the determination of species on the basis 

 of external morphology. 



In the employment of this classification the novice will en- 

 counter certain practical difficulties the nature of some of which 

 it may be well to indicate. In the genus Picea the differentiation 

 of species is attended with more than the usual difficulty, and 

 the same fact appears once more in the second division of the 

 genus Finns. This appears to be the result of a general advance 

 toward a higher type of development, in consequence of which 

 there is a more uniform distribution of similar characters among 

 the various species. This feature also appears occasionally in 

 other genera, especially in Juniperus, where it is not altogether 

 easy to differentiate J. nana from J. communis, of which it has 

 commonly been regarded as a varietal form. But J. rigida shows 

 precisely the same relations to both of these, and I therefore 

 prefer to retain the specific status of all three, though somewhat 

 provisionally. In Pseudotsuga there is as much anatomical differ- 

 ence between P. Douglasii and P. macrocarpa as there is between 

 any well-known and well-recognized species. There is therefore 

 no reason for assigning the latter to a varietal position, and it 

 should be given the status of a species, as correctly suggested 

 by Sargent. In the genus Pinus, P. Murrayana cannot be regarded 



