Tubiflorce. 301 
to the bicarpellary condition of the ovary among the ContortEe and 
Tubifloras taken as a whole—the ovary in Polemoniaceae comprising, 
almost without exception, three carpels. For the rest, the corolla 
is invariably contorted in aestivation ; the epipetalous stamens are 
equal in number to, and placed alternately with, the corolla-segments; 
the ovule-number is 1— CO per carpel. The species are mostly 
herbaceous, rarely fruticose ( Cantua ). 
It seems somewhat incongruous to include a very natural order, 
comprising 200 species characterized by a tricarpellary gynaecium, 
in a group of over 16,500 species every member of which has a 
bicarpellary gynaecium— i.e., the so-called Bicarpellatae. Some 
attempt has been made in these chapters to indicate the significance 
of the bicarpellary arrangement; but apart from any principles 
formulated here, the existence of some such significance is apparent 
from the mere fact that an ovary of two carpels is constant for an 
enormous number of species which are more or less closely related 
on other grounds. 
The strongest claim for the inclusion of the Polemoniaceae 
among the descendants of the apocynal stock would seem to be the 
isomerous and alternating epipetalous androecium; the primary 
economy principle has here reached its highest expression. But in 
the gynaecium, the working of the economy tendency has ceased at 
the stage which we have already noted as so common among the 
Archichlamydeae (supra, p. 112); and in the essential characters 
the flower in Polemoniaceae closely resembles many of the Ericales, 
such as Epacridaceae, Diapensiaceae, in so far the androecium is 
isomerous and alternating, and the gynaecium multilocular. 
It would seem in all the circumstances that we must seek the 
ancestry of the Polemoniaceae further back than in the apocynal 
stock, which is essentially bicarpellary. They may, indeed, represent 
a link between this stock and the Archichlamydeae; and in this case 
they should find a place among Heteromerae—as representative of 
a sympetalous group which adopted sympetaly before its primary 
economy tendencies were fully worked out. However, the discussion 
of the affinities of this order in any further detail is scarcely within 
our field; suffice it to say for the present that we shall regard the 
Polemoniaceae as the descendants of a stock separate from the 
apocynal stock and not so advanced ; this has been shewn on the 
diagram p. 300, where the similarity between this order and the 
Hydrophyllaceae is indicated as the result of parallel development. 
It is conceivable that the Polemoniaceae were derived from the 
geranial stock before carpel-economy had reached the bicarpellary 
