
          to peruse. There are some improvements on the
 single files as published in the Philad [Philadelphia] Acad [Academy] Proceedings
 some deceptive external, or "type", affinities having
 been corrected in the Phila [Philadelphia] Proc [Proceedings] I have thought
 that Loasaceae and Passifloreae might connect, but
 the genera being tropical almost entirely I could not
 make sure, and found cause to disconnect them.


 I have paid most attention to Dicotyledons, and
 have arrived at the following results : 1) There is a 
 positive single file [crossed out: arran] connection, a beginning and
 and end. 2) If subdivided into the 5 chief heads which
 you find in the St. Louis Proc. [Proceedings], each of these files
 presents two points, where a striking identity of external,
 obtains; so as to be realizable by the figure of a "traversing"
 of files, assuming a common character of the
 points of intersections. 3) These intersections, producing
 three parts to each file, divide the same say as to pies
 sent as many distinct habitus, terminated at either
 [added: end] by a epigynous form belonging to that habitus, as e.g.
 Ribes and Rhipsalis at the extremities of the Crassulaceae
 saxifragous [?]. 4) All the [added: respective] tripartitious of these
 bars seem to round off into themselves, besides connecting

        