equal separate pollinia; and 4, two undivided pollinia. 
Most of the genera in this scheme fall within the first 
two categories, each forming a distinct line parallel with 
the other. At the end of each line are those genera which 
are characterized by either two entire pollinia or four 
equal pollinia. It is indeed quite revolutionary to derive, 
as Holttum did, the monopodial orchids from such sym- 
podial ancestors which are also characterized by two 
notched pollinia and then terminating that line by four 
equal pollinia, when it has been repeatedly demonstrated 
that the two-pollinia stage is a derived condition through 
reduction from the less specialized 8 pollinia per anther. 
The emphasis on four pollinia in two pairs or on two pol- 
linia, each consisting of a cleft pair, 1s more semantics 
than evolutionary reality. This has contributed on several 
occasions to false claims in relationships. 
The four distinct categories mentioned by Holttum 
apply only to the Asiatic monopodial orchids where the 
pollinia are attached toa single, common stipe. Although 
he refers Angraecum and its related genera to the group 
characterized by two notched pollinia (and considers 
them as a primitive group because of their simple, un- 
adorned lip), the angraecoid orchids clearly show ad- 
vanced specialization; in most cases each of the pollinia 
are attached to separate stipes which share a common 
viscidium or each are attached to separate viscidia. In 
highly evolved cases the stipe may be completely elimi- 
nated through further reduction. The phenomenon of 
reduction is a dominating force in the whole orchid family 
and it can manifest itself independently or in conjunction 
with other evolutionary trends, causing only one or 
simultaneously several expressions in the phenotype. 
Probably cognizant of these possibilities, Summerhayes 
has seen it advisable to separate the angraecoid orchids 
into a distinct line based on characters which are quite 
[ 1538 ] 
