202 



THE AMERICAN NATURALIST [Vol. XL 



narrow border left free, where in their upper portions they over- 

 lap their neighbors. The anatomical features cannot be made 

 out in this section as the magnification is not sufficient. 



Plate 4, Fig. 1, shows a transverse section through the woody 

 cylinder of the younger branch shown on Plate 3, Fig. 5. By 

 using a lens it may be seen that at this stage the cylinder (consists 

 of clearly separated bundles. On the right a single leaf-trace is 

 passing off. 



Plate 4, Fig. 2, ref)roduces a cross section of the basal portion 

 of a leaf from the margin of the same young branch. It may be 

 seen that there are several fibrovascular bundles [)resent. These 

 have originated from the single trace previously described. Of 

 the Cupressineje, Sequoiineee, and Araucarinejc, the only Conifers 

 with which Brachyphyllum has ever been plac^ed, the latter group 

 alone have the branched leaf-trace. 



Plate 4, Fig. 3, shows a similarly branching leaf-trace from one 

 of the l)road leaves which clothe the uj)per and lower surfaces of 

 the stem in the genus Ikachyphyllum. 



Plate 4. Fig. 4, shows a portion of the latter inider a higher 



Plate 1, I-i-. 1. |),M't .,) a hranch of Hrach\ pliyUum in trans- 



Although the iiuitcriai was j)assed through a nund)er of solvents 

 of resin, used in a hot conchtion, there is still some matter present 

 in the lumirui of the resin-passages. This appears to be of a 

 mucilaginous nature and is coniparaUe to that found among living 

 Cunijrr. nnl/i n, Ihr g, >n ra \rauraria and Damrnara. 



Plate 4, Fig. r., shows a transverse section of a diseased branch 

 of Hra<-h>pli\lhnn, in uhid. there is one ],articularlv large resin- 

 ca^M^ The fo^.i! nuK ilai^e. a. ue eouMdcr it to he. i. present 

 aUo ni ihi^ inM.niM- \\<- ha\e, ni fad. f<.Mnd ihi^ ^nl..tan.r to 



and those which had become verv rotlen in the pi-oeess of fossili- 

 zation. The mucilaginous confrnis of l/n- rr.siti-cdtKils ofjord an- 

 other reason for associating lirarh i/jdnf// in>i inf}, Ihr IniNninnnr 

 rather than with the Cupressinerp or the Srqnonuctv. 



The most important argument, liowever, in favor of the .A.rau- 



