362 BOTANICAL GAZETTE [may 
although accurately distinguishing these two groups, curiously 
enough failed to see any relationship between them from an evolu- 
tionary standpoint. Nor has any other writer since this classifica- 
tion has been developed pointed out the very evident situation 
which exists. That the prosoplasmas in all cases are derived from 
the kataplasmas is self evident. This fundamental conception, 
together with an interpretation of it to be given later, forms the 
central thesis of this paper. On it has been built the diagram- 
matic presentation of the phylogenetic lines of the higher cecido- 
zoon groups. Following the explanation of these which follows, a 
more detailed discussion of the general theoretical situation will 
be Siiiiiaia 
The phylogenetic lines 
Attention should be called to the polyphyletic origin of Ba: 
the cecidia-making habit having appeared independently in many 
diverse groups. Within certain phyla (ex. Nematoda) it appeared 
among a few closely related genera, while in certain families (ex. 
Itonididae) it probably had an independent origin. in many genera. 
The “‘phylogenetic trees” of the various groups (pls. XXI, XXII) are 
arranged horizontally with some regard to the evolutionary position 
of the cecidozoon groups. Vertically, the figures are arranged on 
the basis of their classification into kataplasmas or prosoplasmas. 
The figures of the galls, while drawn more or less diagrammatically, 
are based on actual species. They have been chosen to represent 
fundamental types only. Practically all of them show the galls as 
seen in section, this being necessary to bring out such salient 
characters as position of cecidozoon, orientation with regard to 
plant parts, and the differentiation of sclerenchyma in certain 
prosoplasmas, this latter being indicated by a dark layer within 
the body of the gall. The lines or “‘trees’’ are based on cecidozoon 
groups of varying degrees of position in the systematic scale. 
None below families have been attempted, although an ideal study 
would present the genus as the unit. At the present time, how- 
ever, this is not possible through lack of anatomical data. In the 
discussion of each group some mention will be made relative to 
the number of genera involved. 
