POPULAR SCIENCE REVIEW. 
i64 
fertile anther cells is as great as in any, and more than in 
most. 
It will also be noticed that the fertile anther cells are guarded, 
by the closed hood in which they are concealed, from the wind 
and rain, so that their precious pollen is not vainly dispersed, 
but is carefully hoarded until it can be discharged with effect, 
fn none of the many blossoms which I examined did I find the 
pollen loose in the hood of the corolla; and in no instance in which 
I was able to watch accurately the egress of a bee from a freshly 
' opened flower did I fail to see the pollen adhering to its back 
and flanks. When the bee, smeared with pollen, leaves the 
blossom, the elasticity of the joint at the end of the filament 
• auses the upper anther cells to retire again into their hiding- 
place, 'where they lie protected till a second bee visits the flower. 
This is, at any rate, the rule, but not unfrequently the retreat 
vloes not occur. The upper anther cells catch in the narrow 
opening of the hood, and remain outside exposed. It would 
rhus appear than when the pollen has been shed in part. Nature 
is less careful in her provision for the protection of organs which 
•dre now comparatively useless. 
Another point to be noticed in the stamens is the straddling 
position of the filaments. This is clearly explained by the 
necessity for a free entrance for the bee, whose passage would 
otherwise be obstructed. The shortness and firmness of the 
rilaments is such as to give a fixed point for the motion of the 
ronnective with its terminal cells. In another salvia, a variety 
of sylvestHs, the immobility of the filament, and consequently 
of the point on which the connective moves, is still further 
secured by a slight adhesion of the far end of the filament to 
rhe edges of the upper lip of the corolla. 
The position of the joint 'which unites the connective to the 
filament also deserves a passing notice. The joint is lateral, the 
filament being united to the external surface of the connective, 
;ind not touching it behind, so as in no wa}^ to interfere with 
rhe swing of the latter backwards and forwards (fig. 5). 
A more perfect arrangement than that which I have now 
• lescribed, for the transference of the pollen to the body of the 
bee, cannot, I think, be imagined. Every part of the stamens, 
'*ven to the minutest details of structure, is so contrived as to 
••nsure this result. 
From the stamens I pass on to the consideration of the style. 
'I'his, as in all labiates, is gynobasic, and from the ovary it runs 
at the very back of the corolla, being in contact with the pos- 
terior surface of the hood, and reaching to its very extremity ; 
that is, to a point considerably beyond that where the upper 
anther cells lie concealed. As the style follows the curve of the 
flood, and this is not only closed behind but also slightly in 
