266 The Fertilization of Salvia splendens by Birds. (April, 
order to which it belongs, and, as is generally the case with labiate 
flowers, the lower lip is split into three lobes, a median and 
two lateral, which in this case are of nearly equal size. Here, 
however, the lower lip—usually well developed and affording a 
convenient landing place for insects—is small and of little or no 
use for this purpose. 
The style is exserted to a considerable distance, and the in- 
‘cluded portion ts held quite firmly in a longitudinal fold of the 
upper part of the corolla tube. The forked stigma (s/) is thus 
maintained in the median plane of the flower. 
The stamens are two in number, and of the general form found 
in this genus. Their filaments are adherent to the corolla to 
within a short distance of its mouth, where they become free, and 
run obliquely upward and forward, terminating side by side, close 
beneath the base of the upper lip. The connective which in 
many flowers forms an inconspicuous band between the anther . 
cells, is here prolonged, in each stamen, into a slender longitudi- 
nally-placed rod nearly an inch in length. Each connective (c) is 
attached at its middle by a hinge joint to the end of its filament, 
thus forming an oblique lever with equal arms, which lies with 
its anterior end @ in contact with, or barely protruding from the 
tip of the arched upper lip of the corolla, while its posterior end a’ 
nearly reaches the floor of the tube. If this were constructed as 
the stamens of related plants are, it should bear an anther cell at 
each end; but in reality only a single fertile cell—the anterior— 
is developed, the posterior cell being abortive. Moreover, the 
connectives of the two stamens are coherent for a short distance 
‘back of their insertion, so that the two form, in reality, a single 
- forked lever. 
When a flower first opens, the stigma is immature, its lobes 
being closely appressed, as shown in the upper part of Fig. 2, 
but the anthers are already dehiscent. In other words the species 
is proterandrous. Later, when some or all of the pollen has been 
removed, the stigmatic lobes expand, as shown in the lower part 
of Fig. 2, and expose the now receptive surfaces, and the 
style curves down into the position shown by the dotted line 
of Fig. 1. From my observations, I should say that the life of a 
given flower may be divided into three periods; in the first, the 
anthers only being mature, it is staminate in function; in the 
second, some pollen remaining in the anthers, while the stigmas 
