1882 .] 
417 
[Trelease. 
visiting the flowers for nectar, such bees, supporting themselves by 
the spreading lobes of the perianth, would necessarily bring a given 
part of their backs in contact with the end of the style. If the 
flower be newly expanded, a good part of the pollen will be thus 
brushed from the stigmatic disk, to remain entangled among the 
hairs which clothe the insect. An older flower, visited in the same 
way, will have its stigma brushed by the same part of the insects 
and so pollinated. In case insect visits do not occur, pollen may 
remain in a fresh condition upon the disk and covering the stigma 
when this organ matures, in which case self-fertilization may 
occur. This does happen in the greenhouse, but in a compara- 
tively small number of flowers. 
Grevillea Thelemanniana (Australia.) — The flowers of this 
species, cultivated under the synonym of G. Preissii, are loosely 
arranged in pendant racemose cymes 6-8 mm. long. Their color 
is on the outside a very pretty shade of red for the most part, 
the tips of the sepals being of a contrasting greenish-yellow. 
The chocolate tint of the inside, when exposed by the reflexion 
of the sepals in the expanded flower, serves as an excellent nectar 
mark. Before anthesis the style has become much longer than 
the perianth, and, its tip being held by the latter, as in Hakea, it 
protudes in a strong arch from the upper side of the flower, figs 
8 and 10. Here the elastic tension which is to aid in separating 
the sepals is much greater than in the species last considered, but 
the process is essentially the same, the tapering part of the style 
beneath the stigmatic disk ultimately forcing the two upper 
sepals apart, when, the dome being broken, the others readily sepa- 
rate for a short distance, and all become recurved, fig. 9. In 
this connection it is interesting to recall the presence of an erec- 
tile spur just beneath the end of the style in some species, which 
probably greatly facilitates the separation of the sepals at the 
proper time, as pointed out by Mr. Bentham (1, 61 ; pi. 1, fig. 5). 
As in Hakea, there is pronounced protandry, the pollen being 
dej)osited on the stigmatic disk before the flower opens, figs. 11, 
12. Here too, provision is made for the attraction of pollinating 
agents by the secretion of nectar through a gland at the base of 
the ovarian stalk, fig. 10, n, the secretion collecting in the 
gibbous sac formed by the united bases of the sepals. The 
entrance to this reservoir is a narrow opening between its con 
PROCEEDINGS B. S. N. H. VOL. XXI. 27 OCTOBER, 1882 . 
