22 DR. J. D. HOOKER ON WELWITSCHIA. 
lateral lobes, and gradually develope into the two inner leaflets. These are succeeded by 
two other rings, formed at the base of the nucleus—the outer being the staminal tube, 
the inner the coat of the ovule: I cannot satisfy myself which of these two is developed 
first; but both the nucleus of the ovule and its coat, which is open and truncated, project 
far beyond the staminal ring for a considerable time; and it is some time longer before 
either are wholly included in the perianth. 
The fully formed flowers (Plate VI. figs. 5, 6, & 7) are sessile and much compressed, 
broadly obovoid-spathulate, shorter than the scale (ths of an inch), plane towards the 
axis of the cone (on the inner side), convex on the outer face. 
Perianth.—The two outer perianth-leaves are strictly lateral, very narrow, spathulate, 
faleate, acute, and keeled or almost winged at the back; their bases are connate or very 
closely approximated on the inner side of the flower (Plate VI. fig. 5), and more widely 
separated on the side opposite to the scale (fig. 6). The two inner segments of the 
perianth (which are not present in the female flower) are antero-posterior, confluent into 
a compressed tube below, spreading into two orbicular concave lobes above, of which the 
inner overlaps the other by its margins all round (Plate VI. figs. 7 & 8). 
These four leaflets of the perianth are very thin and transparent, formed of two layers 
of epidermis (figs. 8 & 9), enclosing slender thin-walled spicular cells, which are so placed 
as to present the appearance of radiating nerves; these cells have uniform or very 
obscurely dotted walls and a conspicuous cavity, and they are quite colourless and 
transparent. There is no vascular tissue in the perianth of the specimens which I have 
examined, though the two vascular bundles running up to the ovule are evident within 
the solid base of the flower (fig. 14). 
A comparison of the perianth of the hermaphrodite flower with that of the female will 
be found appended to the description of the latter. In some respects it resembles that 
of the male flower of Casuarina more than any other plant I know; but the comparison 
cannot be carried far. The two outer leaflets of the perianth may perhaps more properly 
be considered bracts; but there is no evidence of consequence in favour of this view. I 
Stamens.—The six stamens (Plate VI. figs. 7, 10, 11, 12) are united halfway up intoa — 
fleshy tube, which is adnate at the base to the inner leaflets of the perianth ; the filaments 
above this are fleshy, terete, and irregularly inflexed in sestivation, reflexed and exserted 
after anthesis. The anthers are capitate, obscurely 3-lobed, 3-celled, and dehisce at 
the apex by a tricrural slit (fig. 12) ; the three posterior are much smaller than the others. 
The pollen (fig. 13) is very minute, ellipsoidal, transparent, and yellow; each pollen-grain 
is sooth to z5sth of an inch long, and consists of a delicate hyaline extine, which is (when 
preserved in alcohol) longitudinally wrinkled, and encloses grumous contents. I have not 
been able to determine the nature of the contents of the pollen-grain, nor whether cell- 
division takes place before the tube is protruded. I have never found pollen-grains on 
any part of the ovule of the hermaphrodite flower. 
In being hexandrous, this staminal whorl presents the only marked exception to the 
binary arrangement that prevails throughout this plant; the only other apparent excep- 
tion is in the rachis of the cone, which is traversed by four triplets of vascular bundles, 
of which triplets the two outer bundles alone seem to supply the scale and flower. 1 
