3°4 
H. F. Wernham. 
of an uniovulate and unilocular ovary. In the light of this last 
consideration, the significance of economy in production of mega¬ 
sporangia, already noted in the present chapter, becomes increased, 
for the tendency to economy is seen to ally itself with the all- 
important function of seed-dispersal. The Compositae afford the 
highest expression of this idea; the bicarpellary ovary here is 
unseptate, contains but a single ovule, and matures into a small 
one-seeded fruit dispersed usually by the agency of a pappus. 
The schizocarp represents an alternative method of securing 
a fruit-body for each individual seed, and in so far as this latter 
condition characterizes an extensive section—about 30%—of 
Tubiflorae, we must regard it, in this group, as representing a general 
biological tendency comparable with zygomorphy in the relatively 
isolated flower; the Diovulatae (Lamiales), we shall find, reflect the 
achievement of these two tendencies in combination. 
In view of these considerations we may conveniently name this 
secondary septation of the ovary “ the tendency to schizocarpy.” 
We have already recognized it in a high degree of realization in 
Boraginaceae, and the Solanaceae represent an early stage in its 
progress; we shall meet with it again as a critical character of the 
Diovulatae. The South American Nolanaceae, comprising some 30 
species, all herbaceous, afford a curious illustration of one way in 
which this tendency has achieved its aim. In these plants the 
structure of the ovary is similar to that of Boraginaceae, but 
septation in all directions has resulted in the production of an 
indefinite number of uniovulate chambers, more or less free, some 
superposed above others, and surrounding the base of a gynobasic 
style. Originally classed by de Jussieu with Boraginaceae in view 
of the last-named character, the Nolanaceae were included in 
Solanaceae by Duval and Baillon, and in Convolvulaceae by Bentham 
and Hooker. The character of the pistil, however, seems to justify 
their rank as a separate natural order, in accordance with Engler’s 
system where they are associated with Solanaceae ; but their regular 
flower and isostemonous androecium urge their inclusion in the 
Transitional Group, and their affinities appear to be rather with 
Boraginaceae, with a proximate convolvulaceous ancestry. 
The question of secondary septation of the ovary must be 
distinguished carefully from that of its primary septation ; the latter 
is one of extreme difficulty, and it is desirable in these papers to 
avoid if possible the necessity for more than the mere statement of 
it. In so far as the angiospermous ovary is to be regarded as 
composed of sporophylls or carpellary leaves, whether by the 
