The Pentacyclidce. 
151 
considerations. The nature of the anthers in Ericales must be re¬ 
garded as an item of special importance only, so far as our present 
knowledge goes, in the absence of any general biological explanation 
therefore; it must thus be reduced to the same level with habit in the 
discussion of general affinities with other groups, although it may 
have considerable practical value in the diagnosis of the group itself. 
The anther-structure, like the habit, represents the special progress 
made by the group over and above the progress made on the broad 
general principle of economy, and does not reflect a general tendency 
external to the Ericales ; it may, therefore, be accorded little or 
no weight in the search for external affinities for the group. 
Such considerations as these seem to afford no little justifi¬ 
cation for urging that the Vaccinioideae have an origin separate from 
that of the rest of the Ericales, and that consequently they should 
be separated from them in a natural system of classification. This 
origin, we have already hinted, may reasonably be sought in the 
ancestral stock represented by the Rosales; and we must suppose 
that the vaccinioidean line emerged from this stock before the 
appearance of the more advanced members of the Rosales, its 
determining principle being the fusion of petals to form a corolla- 
tube. 
The position adopted in the foregoing may conveniently be 
summarized in the following diagram:— 
Atvxcesbiafi. 
The question arises, to what place are we to relegate the 
Vaccinioideae in a natural system ? This question, however, cannot 
