212 MR. J. G. BAKER ON SCILLEJ AND CHLOROGALEEX. 
segments are distinctly incurved at the point, or even along the 
sides. In colour and nervation there is a wide range in the two 
tribes, and within the bounds of each of three larger genera, Sella, 
Ornithogalum, and Urginea. For instance, in Ornithogalum we 
may trace at least three distinct types :—the first, represented by 
arabicum, in which the colouring is uniform and the veining alto- 
gether fan-like and inconspicuous; the second, represented by 
umbellatum, in which there is a broad, laxly many-nerved dash 
of green down the back, which leaves only a narrow rim of white 
down each edge; and the third type, represented by narbonense, 
in which there is only a distinct narrow band of green down the 
keel on the back, which contains three or four closely crowded 
parallel more or less decidedly anastomosing nerves. In Silla 
there is always a narrow single-nerved keel; but there is great 
difference in colour and in the manner in which the divisions 
spread in the fully expanded flower. In this latter point our 
three English species, verna and autwmnalis on one side, and our 
common wild Hyacinth on the other, represent the extremes. 
In Urginea there are two well-marked groups—one characterized 
by narrow-keeled one-nerved segments, like Scilla, and the other by 
broad-keeled many-nerved segments, like Ornithogalum umbellatum 
or Albuca. 
Stamens.—The position of the stamens furnishes great help in 
characterizing the genera of Scillew ; but in the great majority of 
the genera I have called perigynous they are placed very near the 
base of the divisions. Only in the Agraphis section of Scilla are 
they decidedly biserial and unequal. Of the other perigynous 
genera or groups, they are placed a little above the base in Euco- 
mis,W hiteheadia, and Scilla, section Ledebouria. In Ornithogalum 
there are sometimes curious differences between the alternate fila- 
ments. Sometimes the alternate ones, and occasionally all the 
six, are dilated aud petaloid at the base or even throughout, with 
three cusps, the central cusp bearing the anther, as in the section 
Porrum of Allium. Usually the filaments are only alittle shorter 
than the perianth, the only conspicuous exception to this being 
in Drimiopsis. Often the stamens are slightly declinate. The 
six filaments are always fully developed; but in one section of 
Albuca the alternate anthers are commonly abortive. I have not 
observed any noteworthy variation in the dehiscence of the anthers 
or in the way they are fixed upon the filaments. 
Ovary and Capsule.—In only one instance in the two tribes, 
