168 MONSTROSITIES IN OPHRYS INSECTIFERA. 
Fig. 1 and 2 represent the anther and stigmatic chamber of 
two flowers taken from the same spike, considerably magnified. They 
show the rostellate process present in this plant, but converted (Fig. 1) 
into a miniature anther, in which the pollen masses, though devoid 
of caudicles, are composed of healthy-looking granules connected by 
elastic threads in the usual manner. 
Babington has (Man. of Brit. Bot. p. 305) divided Orchis, Gym- 
nadenia, and Aceras, from Habenaria, Ophrys, and Herminium, by the 
presence of the rostellate process in the former, and its absence in the 
latter section. However, this process may be occasionally found as in 
the present instance, in monstrous specimens of the genus Ophrys, 
though in the normal form it is absent. 
At Fig. 3 I have given an instance of the production of two 
perfect labellums in a flower, the remaining divisions of which are of 
the usual form and number. 
At Fig. 4 the anther and stigmatie cavity of a flower is repre- 
sented, exhibiting a curious case in which a third rostellum is pro- 
uced. This feature has a peculiar interest, as the point at which the 
rostellum is placed corresponds exactly with the situation of one of the 
two lateral glandular processes which are so generally present in Orchids, 
-and which represent, in a rudimentary condition, the anthers found in 
Cypripedium. 
[To these observations and illustrations we have added the follow- 
ing from Mr. W. G. Smith.—E»p.] 
Fig. 5 shows an abnormal growth of the same species (O. aranifera). 
lst. Two of the sepals are confluent. 2nd. There is great interest 
attached to the left-hand petal, as it is confluent with, and forms part 
of, a second imperfectly-developed column, and bears an anther-cell 
and pollen-mass (better seen in the side view, fig. 5a). The true 
column only bears one anther-cell and pollen-mass, the other being 
abortive, but leaving a trace of its presence and position. 
