and the numbers, connecting each of these groups, as 
representing independent trends which originate from 
and link the groups at different points. 
The sundry orchid systems proposed in the past were 
based essentially on progressive differentiation in a linear 
sequence, which assumes that the family is monophyletic 
in origin. Indeed, one’s first impression is of a linear se- 
quence from A postasioideae to Kerosphaeroideae, espe- 
cially when the groups are evaluated individually. When 
we attempt to assign a definite position and sequence to 
each of these groups, however, after studying their alli- 
ances, we find the linear arrangement to be rather absurd 
and unrealistic, since the groups are constantly in juxta- 
position with each other. The pattern expressed by these 
interrelationships is reticulate and indicate that the fami- 
ly, as a unit, is polyphyletic in origin. 
The general belief that, during the course of evolution, 
Neuwiedia (A postasioideae) gave rise to both Cypriped- 
toideae and the monandrous orchids, is hardly tenable. 
The so-called ‘‘clear cut’’ differentiation between Dian- 
drae (A postasioideae and Cypripedioideae together) and 
Monandrae is obliterated, as was mentioned earlier, by 
the presence of two anthers in the outer staminal circle 
of the Satyrium-complex in Ophrydoideae. In addition 
to this criterion, there are a number of characteristics 
common to both Diandrae and Monandrae which have 
been discussed above as well as documented on Plate 
XIII. L look upon each group as having an equal stand- 
ing with respect to its adjacent group, and, for the pur- 
pose of classification, | am recognizing each as a distinct 
subfamily.' Perhaps some phylogeneticists may object 
to such a conclusion, but we have to bear in mind that 
each group is rather well circumscribed in spite of the 
close interrelationship. 
' For the descriptions of these subfamilies see Appendix. 
[ 78 | 
