10 
fect : but the common case is that this lower peduncle carries only one or two male flowers, and the pedicels disarticulate off before the 
upper peduncle blooms. In some plants, in some of the spathes there are pedicels on the lower peduncle which drop early, in other ae 
“spathes the pedicels are aborted altogether. We next come to species as O. suffruticosa, in which the lower pedicels are always wanting, 
and in this the rudimentary lower peduncle is often itself nearly aborted. From these observations I have been led to attach small import- 
ance to the distinction between one peduncle to the spathe and two peduncles to the spathe hitherto much relied on in the specific cha- 
racters. 
The form of the stamens in Commelyna is shewn in Tab. XII. figure 3: and also in Tab. IX. (by Roxburgh’s artist). The cru- 
ciform top to the barren filaments is yellow and without any trace of pollen... In most of the species. the central deformed perfect stamen 
has yellow anthers, and the two lateral perfect stamens have blue anthers. Though the anthers are thus unlike in form and color, the pol- 
len therefrom is exactly alike, ellipsoidal yellow, nearly smooth. In C. suffruticosa, the perfect anthers are dissimilar in form, but all of 
them yellow in color. In Commelyna, in all the species, the anthers are white till a very late period of their development, and only gain 
their different colors just before expansion. Another specific character has been taken from the degree of inequality in the size of the pe- 
tals. In my paper on Commelynacew above referred to, I divided out the species mainly by characters taken from the capsule and seeds. 
Dr. Hasskarl has been led to sub-divide the genus on the same principles. It has been objected to me, by the highest authority, that in 
the Commelyna communis bundle of the Kew Herbarium, the capsules are as often 4-and 3-seeded as 5-seeded. This would be the case if 
Commelyna clavata and Commelyna sikkimensis are admixed, as they assuredly would be if the bundle has only been hand-and-eye sorted 
and never worked. It is possible, indeed, that hereafter Commelyna sikkimensis and C. clavata may finally be reduced to varieties of C. 
communis. I can only reply that I have examined now a very large number of specimens of Commelyna, and that I feel confident no 
better characters for limiting the species than the capsules and seeds will soon be found. 
The capsule of Commelyna communis contains three cells : the two larger cells contain each two superimposed seeds and 
dehisce : the third cell is placed lower down on the axis of the capsule and contains one larger and differently shaped seed, and is indehi- 
scent or very rarely dehisces very late. Vid Tab. I. figures 1, 2, 3, 4, 5, 6, 7, 8. Commelyna salicifolia, Tab. II. and Commelyna bengal- 
ensis, Tab. IV. exhibit the same arrangement of capsule and seeds. In Tab IV. figure 4 (copied from Wight) the capsule is seen dehiscing, 
the four seeds are shewn ready to drop in one valve, and the other valve has one seed in the indehiscent cell. The two superimposed seeds 
in one of the larger cells fit close together, and when so fitted resemble much the seed in the solitary cell and are a little larger than it. 
The capsule of Commelyna sikkimensis Tab. VI. is the same in structure and dehiscence : but each of the larger dehiscing 
cells contains one seed only. The capsule is here three seeded, the seeds being similar in form, but that in the closed cell being the 
smallest. 
The capsule of Commelyna appendiculata Tab. XTIT. and the seeds are generally arranged as in ©. sikkimensis, but the 
third unequal smaller cell is reduced : and on the same plant are often seen cases where it is entirely obsolete, and the capsule comes out 
regular 2-celled, as in Tab. XTIT. figures 8, 9, 10. 
The capsule of Commelyna obliqua, figure 9, is seen regularly 3-celled, 3-loculicidally dehiscent with one seed in each cell ; 
all the seeds being similar : or on the same plant it may be found regularly 2-celled, exactly as in the reduced form of capsule of Commelyna 
appendiculata. Finally, Commelyna suffruticosa always has the capsule equally 2-celled loculicidally 2-valved with one seed in each 
cell. 7 
Miquel relates of Commelyna bengalensis (2) as found in Java that the capsule contains three cells and two seeds in 
each cell. Tf this plant is rightly described, it will be a new genus for Dr. Hasskarl, and certainly cannot be Commelyna bengalensis, Roxb. 
or any species of Commelyna known to me. 
The seeds of Commelyna are attached by a linear hilum ; Vid Tab. I. figures 7, 8. When the ovule ir. the cell is solitary, the 
growing point is at the middle of the seed. Vid Tab. I. figure 6. But when there are two ovules in a cell, each seed has its growing point 
near the end. Vid Tab. I. figure 5. a 
Species infra descripte. 
Sect I. a. Hucommelyna. 
1. Commelyna communis. 
2 salicifolia. 
3. CO. Hasskarlii. 
‘A ae 4, 0. bengalensis 
b. Dissecocarpus. 
5. 0. clayata. 
Sect II. 
6. C. rajmahlensis, 
Wo © sikkimensis. 
BG Kurzii. 
©, obliqua. 
; 10. C. suffruticosa, ; 
(0) 
ii, (eh appendiculata, 
