13 
stamens but as a kind of stipules (,,Stipulargebilde”) such as 
in Ornithogalum and Allium. This conception which seems quite 
acceptable regarding the motives cited by Eichler, in sustained 
by the fact that. they are never affected by the thorough alte- 
rations which the five stamens undergo. As the latter, see infra, 
can become perfect pistils, the barren stamens might at the 
same time be induced to change their character and fall back 
to their supposed rea] nature. But no trace of such a meta- 
morphosis could be found in the cases I had the opportunity 
of examining. 
The commonest condition met with in the Batavia flowers 
is represented in fig. 17 d-e. Instead of 5 stamens (or in hex- 
amerous flowers 6) we find 5 (6) pistils, differing from the central 
pistil only in size and in the length of the stalk on which the 
ovary stands. Besides, the above mentioned ,stipules” are 
broadened and multipartite. In this way they take the shape 
of a continuous envelope which as a kind of second perianth 
encloses the pistils. That the latter actually occupy the places 
of the stamens appears from their alternation with the stipules, 
as indicated in the figures where we have left out one pistil 
between those that have been drawn. Still more convincing is 
the appearance of intermediate forms between anther and pistil. 
In fig. 17 f-h I have drawn three of them: in / we find closely 
cohering with a theca on a normal filament a pistil differing 
only in size from a normal one. One should be inclined to think 
that this pistil has been developed at the cost of one of the 
thecae or, to express it differently, takes the place of one of 
the thecae. But as the anthers of Telanthera are unilocular like 
in some other Amarantaceae, this supposition must at once 
be rejected. Rather a bifurcation is to be thought of, also on 
account of the fact that the filament now and then bears two 
pistils — either complete or defective — besides a distinct theca. 
These facts are well illustrated in g, where at the top of the 
filament stands not only a normally shaped pistil, but also the 
upper portion (stigma) of another and moreover a somewhat 
rudimentary theca, and in 4 where we see even two pistils, 
