and G. robusta and G. cuneifolia in the one in which all or some of the outer akenes are 1-2 dentate or auriculate-bordered at the summit; the "akenes (at least outer ones)," in G. robusta, are stated to be "obliquely auriculate or broadly unidentate at summit," and the "outer akenes," in G. squarrosa, "commonly (but not always) corky-thickened and with broad truncate summit, those toward the centre narrower and thinnerwalled and with smaller areola." The accuracy of these observations appears to be fully borne out by the material which I have examined. The akenes of the disk in G. camporum and G. cuneifolia are compressed, and in G. camporum they are usually biauriculate (see Fig. 2), or, more rarely, unidentate at the summit, as indicated by Professor Greene in his "Manual of the Botany of the Region of San Francisco Bay" (p. 171); those of G. squarrosa are four-angled, and without the distinct auriculate appendages of G. camporum (see Figs. 3 and 3a which were drawn by me from one of Drummond's specimens). The conical urceolate shape of the heads of G. squarrosa, which is stated in the U. S. P. to differentiate that species from G. robusta (in which the heads are described as depressedurceolate), does not seem to be a character of much value, judging by the specimens which I have examined; the photographs of undoubtedly authentic G. squarrosa reproduced on Plate II. will show that the heads of this species may be very depressed indeed. The evidence afforded by the vegetative characters of such polymorphous species as G. squarrosa and most of the western species is also of uncertain diagnostic value, but the upper leaves of G. camporum are usually broader relatively to their length than those of G. squarrosa (compare Figs. 4 and 4A, which represent average shapes in the former, with Figs. 5 and 5A, which represent the predominating shapes in G. squarrosa.) The foliage of G. camporum is also usually paler than in G. squarrosa, and the leaves are more rigid in the former, as stated under G. robusta in the "Synoptical Flora," I. c., p.

119.

A possible reason of confusion in connection with the drug of commerce may have arisen from the fact that the terms "Grindelia squarrosa" and "Grindelia robusta" have been loosely applied to the drug by dealers.

I was, for instance, informed by a San Francisco dealer that when eastern drug houses required "Grindelia squarrosa" and "Grindelia robusta " to be separately provided, the plant growing in the marshes (G. cuneifolia, and its variety paludosa) was supplied for "Grindelia robusta," while G. camporum, the plant of the dry hills and plains, was supplied for "Grindelia squarrosa." These designations are, to some extent, justified by the facts, inasmuch as G. cuneifolia, var. paludosa, Jepson, is the largest of all the Grindelias, attaining a height, according to Jepson, of five feet in the Suisun marshes, and hence deserving of the name “robustissima," while G. camporum is probably as squariose as any of the Grindelias.

It is interesting to record that Nuttall's type-specimen of G. cuneifolia