family, because in some of its members the fusion be- 
tween stamens and style is only partial and the length of 
the adnation is relatively short. Indeed, in some species 
of Neuwiedia and A postasia, the filaments of the anthers 
are partially recognizable, but this character is not even 
constant within a given genus. In A postasia papuana 
(Plate XII), for instance, the filaments are completely 
obliterated or reduced to a mere connective tissue. In 
this respect Cranichis crumenifera (Plate XII) or any 
species of the genus Spiranthes or EHrythrodes of the 
Neottioideae might likewise be removed from the Orchid 
family, because structurally the column is quite homo- 
logous in these taxa. Should we express the differences 
between families by a degree of adnation of these organs 
in millimeters ? 
The same incomplete fusion is also observable in Va- 
nilla anomala (Plate X11) (Neottioideae) where the struc- 
ture of the column is comparable to that of Neuwiedia, 
although only one anther is fertile, while the other two 
are expressed by the traces of the filaments as staminodes. 
Vanilla anomala exhibits a further important feature, 
viz., the versatile anther. This character, along with 
others, has also been marshalled to support the separation 
of the A postasioideae into a distinct family. The occur- 
rence of a versatile anther is not limited to this single 
species in the Neottioideae; there are a number of other 
genera and species which possess the same characteristic : 
e.g. Cephalanthera, Psilochilus, Limodorum, Gatleola, 
Didymopleaiclla, ete. In other instances, the anther is 
basifixed and attached rigidly to the column. Vanilla 
Griffith var. formosana (Plate XII) clearly illustrates 
this situation, but a similar method of attachment is also 
present in several species of A postasioideae. 
In viewing the other columnar structures as depicted 
on Plate XII, our attention is focused on another point of 
[72 ] 
