14 NATURAL HISTORY OF PLANTS. 
which we have been able to study have only one carpel at all ages, 
while in the other types, whose fruit is unicarpellary when 
adult, there was a larger number of carpels at some earlier period. 
4, The state of the interior surface of the pericarp: this is covered 
with peculiar hairs in Cuestis, but remains glabrous in the neigh- 
bouring genera Cnestidium and Teniochlena. As regards the per- 
sistence or precocious fall of the calyx, the degree of closeness with 
which it embraces the base of the fruit, the presence or absence of 
an aril—in our eyes these characters are not even of generic value, 
inconstant as they are’ in certain genera which our predecessors 
have regarded as perfectly homogeneous. Thus several authors have 
held Rourea generically distinct from Byrsocarpus and Bernardinia, 
in that its calyx persists, closely applied to the base of the fruit, 
while in the other two it diverges from it, even falling off after 
anthesis in Bernardinia. But we have shown’ that “in the series of 
species from Madagascar we find every transition from the Senegal 
species of Byrsocarpus with spreading sepals, and those of the mimo- 
soid Rowreas from tropical Africa in which the appressed calyx is 
most marked...” In fact this is only a question of degree, so that 
“it is impossible to lay down the law, at what point in this series of 
species the calyx ceases to be that of a Byrsacarpus, and becomes 
that of a true Rowrea.”’ The non-persistent calyx of Bernardinia is 
equally insufficient to make it a distinct genus from Rourea, for in 
the genus Connarus itself, species with persistent sepals, are united 
to others with caducous sepals, without our being able to use these 
differences to found even distinct sections; these two characters can 
then afford no acceptable generic distinctions. This will not apply 
to the accrescence of the calyx, for it is sufficient to separate Rourea 
and Connarus, which genera we have already seen are perfectly dis- 
tinguished by another character. 
Connaracee are distributed? over no wide zone of latitude, but are 
found under almost every degree of longitude in all the warm regions 
of the globe. Not one species it is true has been found in tropical 
‘Australia, and only one in the Islands of the Pacific. But the 
hundred and fifty described species are nearly equally distributed 
over the whole of the warm districts of Asia, Africa, and tropical 
1 Adansonia, vi. 228 (see above, p. 5, note 3). 2 Linpu., Veg. Kingd., 468. 
