132 PROCEEDINGS OF THE AMERICAN ACADEMY 



the former groups on account of their indehiscent 1 -seeded fruit, 

 which is called an achene. But the carpels do not become such 

 achenes as characterize the tribes to which he refers these genera, 

 where the ovules are solitary, for the ovules are here always two 

 and pendulous, while the mature fruit here, though indeed tardily or 

 perhaps never dehiscent, still is found to open more or less readily by 

 the ventral suture on dissection. F'ocke (in Engler & Prantl) recog- 

 nizes this objection, and accordingly forms two subtribes, HolodiscecB 

 and Ulmariecs, for the two genera, quite unnecessarily, for they fall 

 very naturally among the Spirceece and Gilleniece. In fact Holodis- 

 cus * is a true Spircea, except in the thinner and less dehiscent carpels 

 and fewer ovules, and may well be kept as simply a section of that 

 genus. The Mexican S. parvifolia appears, from the descriptions 

 that are given of it, to belong rather to this section than to the section 

 Petropliytum, where it is placed by Maximowicz. Filipendula is much 

 more distinct in its stipules and divided leaves, cymosely paniculate 

 inflorescence and capitate stigmas, and is more nearly related to Gil- 

 lenia. Too great stress should not here be laid upon the position of 

 the carpels (in Filipendula alternate with the sepals when of the same 

 number), as their number may be considered as normally ten in all 

 cases, with irregular or regularly alternate suppression. Focke in- 

 cludes Chamcehatiaria in Sorbaria, but why he does not with as good 

 reason include Spirceanthus is not evident. The three genera are 

 closely allied, very much in the same way as the three sections of 

 Mriogynia. But Chamcehatiaria is sufficiently well marked to be re- 

 tained as the American representative of the group. Finally, the 

 separation of Anmcus upon its herbaceous habit, compound leaves, 

 dioecious racemose-paniculate inflorescence, etc., leaves Spircea itself 

 still a large genus, but well defined, and with a good degree of 

 uniformity in its characters. , 



Eremiastrum Orcuttii. Pappus consisting of five white ob- 

 long-ovate laciniate paleaj and as many inner alternate bristles twice 

 as long : in every other respect — habit, foliage, pubescence, invo- 

 lucre, etc. — the nearly exact counterpart of E. bellioides. — Collected 

 in the southeastern part of the Colorado desert, in San Diego County, 

 California, by C. R. Orcutt, April, 1889. By the character of the 

 pappus this plant is a Chcetopappa. From its close resemblance to 



* The ovules in the species of this section seem to be very often abortive. 

 Upon the man}' apparently niatiire specimens of the forms of 5. discolor in the 

 Gray Herbarium I have succeeded in finding but a single seed. 



