1908] HOLM—SISYRINCHIUM 181 
foreleaves, as we know, are very frequently subject to variation, 
sometimes in accordance with the surroundings 
The rhizome is usually very short and cespitose, but in some 
species, S. californicum for instance, the rhizome is horizontally 
creeping, with the internodes quite distinct. The leaves are two- 
ranked, equitant, and the aerial stems strongly compressed, except 
in S. grandiflorum, in which leaves and stems are almost cylindric. 
It would be interesting to know whether the species of these three 
sections germinate in the same way, but so far this point does not 
seem to have been studied. For this reason I have followed S. 
angustifolium, which is so very frequent in the vicinity of Brookland, 
from seedling to mature plant, and the structure of the seedling is as 
follows. The cotyledon (Cot, fig. 1) is epigeic and consists of a sheath- 
ing base and a long filiform blade, the apex of which remains inclosed 
within the seed for some time, as shown in the accompanying drawing. 
Three leaves (L*-L3) are developed from the plumule during the 
first season, and the structure of these leaves agrees with that of the 
later ones, being ensiform and green. The primary root (KR) grows 
and remains active during the first season, and becomes ramified ; 
secondary roots (r) develop also, and these proceed from the base of 
the cotyledon. It requires several years for the plant to reach the 
flowering stage, and until then the small rhizome remains as a mono- 
podium similar to Iris. This method of germination represents 
Kirss’ type 5,4 and corresponds, to some extent, with that Laven 
frequent among the dicotyledons where the seed leaves are epigelc. 
It is characteristic of several Liliaceae (Allium, Bowiea, Asphodelus, 
etc.) and of Agave, but not of Iris; in J. Pseudacorus, for instance, the 
apex of the cotyledon remains inclosed within the seed, and does not 
ome free as in Sisyrinchium. So far as the external structure of 
the various members of the three sections mentioned above is con- 
cerned, there seems no plausible reason for dividing the genus, and 
We shall see that the internal structure also does not warrant any such 
Segregation. Whatever distinctive characters have been observed 
are merely sectional, and it is hardly necessary to call atten- 
Non to the fact that much more pronounced deviations in struc- 
Taps ttt8e 2Ur Morphologie und Biologie der Keimung. Untersuch. Bot- Inst- 
12572. 1881-1885. 
