180 BOTANICAL GAZETTE [SEPTEMBER 
making the anatomical diagnosis a little more complete. Moreover, 
I believe that the internal characters observed in representatives of 
the three sections mentioned above will strengthen the former view 
of considering Sisyrinchium as one genus, and a very natural 
one. 
The material examined consists of S. anceps Cavan., which is 
very frequent in sandy soil among rocks on the Potomac shore (D. C.); 
S. angustifolium Mill., from sandy hills near Brookland (D. Ci 
S. montanum Greene, from gravelly soil along creeks on Long’s Peak 
(Colorado), at an elevation of about gooco feet; S. xerophyllum 
Greene, which I found quite abundant in low pine barrens neat 
Eustis (Florida); S. grandiflorum Dougl., from dry hillsides in 
Oregon; and S. californicum Ait. {. from swamps at Bodega (Cal.). 
The last two species were kindly sent to me by Miss ALICE EasTwooP 
and Mr. THomas Howett. To these may be added the tuberous- 
rooted S. alatum Hook. from lava fields above Cuernavaca, Morelos, 
Mexico, at an elevation of 8500 feet, collected by Mr. C. G. 
PRINGLE. 
As stated above, the general habit of these species is very much 
the same, especially the inflorescence. This represents in = 
species of Sisyrinchium a rhipidium,3 and this type of monochasium 
is frequent among the Irideae. Whether the inflorescence is single 
and terminal or accompanied by a lateral, the flowers are always 
surrounded by green, leaflike bracts (L, L', and L’, fig. 2); @™™ 
branaceous foreleaf occurs at the base of the lateral inflorescence 
(P, jig. 2) and at the base of the axillary flowers (P?—P?, fig. 6). The 
successive development of the flowers may be seen in the diagt@™ 
(fig. 6), where five flowers (z~5) are situated in the same plane, a” 
where z is the terminal, the earliest developed. The various stag 
of these flowers are shown in figs. 3-5, but the two bracts (Lt, L’) 
are only indicated. The stamens, especially the anthers (St, figs: 
4, 5), are much farther advanced than the outer (P*) Z 
perianth leaves (P?); I have found no deviation from this type i? 
inflorescence in any of the three sections, but the relative length ° 
the green bracts (L-L?, fig. 2) varies very much, and is conside ae 
a specific character, with no good reason, however. Bracts 
3 Compare Etcuer, Bliithendiagramme 39. fig. 204. 
