4 MINNESOTA STUDIES IN PLANT SCIENCE 



genera, but his arrangement of the genus Micranthes is less fortunate since 

 it is made to include both epigynous and hypogynous species. In a later 

 treatment of these species Small (28) substitutes Necker's earlier generic 

 name Hydatica for the Spatularia of Haworth. In other respects the 

 arrangement of the species described is essentially the same as that in North 

 American Flora (I.e.). 



Discussion of the Proposed Classification 



The classification here proposed is founded on the characters of the 

 hairs, the receptacle, the carpels, and the gland.' All structures concerned 

 are shown in Plates I, II, and III. While the present work materially 

 increases the number of sections, the characters on which these are based 

 are believed to be of sufficient taxonomic importance to justify the separa- 

 tions. 



The hairs were first made use of in the classification by Godron (12), 

 and while Engler (5, p. 75) based his primary distinctions on seed charac- 

 ters he offers a substitute classification, based on the characters of the hairs, 

 for the larger categories. 



Two types of hairs are to be found in the groups to be described, namely, 

 the uniseriate type ("ein-reihig" or "articulati" of Engler) and the multi- 

 seriate ("mehrreihig" of Engler). The former type is composed of a 

 single row, the latter of several rows of cells. The sections to be described 

 thus fall into two divisions, the uniseriate-hair type represented by the sec- 

 tions Arabisa and Heterisia* the multiseriate-hair type represented by the 

 remaining sections. 



In some remarks on the use of hairs in classification Engler (5, p. 75) 

 makes the following statement : "Da die Haare soweit meine Beobachtungen 

 reichen, bei den Arten -der einzelnen Sectionen stets gleichartig sind, so ist 

 eine darauf gegriindete Uebersicht jedenfals natiirlich." However, neither 

 in his Monographie nor later in Die Natiirlichen Pflantsenfamilien did he 

 adhere consistently to this statement inasmuch as he failed to recognize 

 the characters of the hairs in those species which, in the scheme of classi- 

 fication here proposed, are included in the sections Arabisa Tausch, and 

 Heterisia Rafinesque (the former being synonymous v/ith the genus Spat- 

 ularia of Haworth). Engler furthermore broke up the genus Rohertsonia 

 Haworth, setting apart in a new section, for which he retained the name, 

 a group of species with multiseriate hairs, but at the same time incon- 

 sistently failing to separate the species of Haworth's genus Spatularia 

 {Arabisa Tausch) on the same grounds. 



• In this paper the disc or nectary is spoken of as the "gland." 

 «H«TKBisiA (Raf.) sect, nov., Heterisia Raf. Fl. Tell. 2:69. 1836; as genus. 



