340 THE AMERICAN NATURALIST. [Vol. XXXVIII. 



finally reduced to 1-2 in the summer wood of P. balfoiiriana, 

 point with much definiteness to these particular species, while 

 among the hard pines the occurrence of large, oval or squarish 

 pits, I or rarely 2 per tracheid, segregates a group of four 

 species. Detailed as these features are, they are not accidental, 

 but of such constancy as to admit of no hesitation in accepting 

 the conclusions to which they point. 



The length of the ray cell is subject to considerable variation, 

 not only within the limits of an individual, but as between one 

 species and another. Our studies, however, do not permit the 

 formulation of a law applicable to specific differentiations, even 

 if such a law does exist, which present evidence leads us to 

 doubt ; but details of length, in terms of spring tracheids, have 

 been incorporated in all the diagnoses, since they are often very 

 suggestive and thus may assist in the ultimate recognition of the 



The form of the cell is of more evident value, although too 

 much §tress must not be laid upon it. The cell is either 

 straight, as in Juniperus, Libocedrus or Picea (Fig. 21), or it 

 becomes fusiform through contraction of the extremities as in 

 Cupressus, Sequoia, Taxodium, etc. (Fig. 19). As a well defined 

 differential character its value is only one degree higher than 

 the length of the cell, and it has been introduced into the diag- 

 noses for the same reason as a controlling factor of secondary 



In the higher Coniferae the medullary ray is distinguished by 

 the presence of an element which differs materially in its struc- 

 ture from the associated parenchyma cells. These elements 

 have been designated by De Bary (9, pp. 491-492), as "ray 

 tracheids." Their structure is so peculiar, and they present 

 such important relations to classification and development, as to 

 necessitate a somewhat detailed account of them, to some extent 

 in recapitulation of well known observations by De Bary (9, P- 

 491) Hartig (19, p. 13) and Goppert (17). 



As stated by De Bary, the ray tracheid resembles the paren- 

 chyma cells, from which they differ, however, in the presence of 

 bordered pits on all their walls. Furthermore, such pits not 

 only differ materially in form and size from the bordered pits of 



