T etracyclidce. 225 
There remain for consideration the Oleacese and Salvadoraceae. 
While it is among these two natural orders alone of the cohort that 
we find examples of polypetaly, they display remarkable advance in 
the nearly constant character of having but one or two ovules in 
each loculus of the ovary, which is almost invariably bi-carpellary. 
In the Salvadoraceae, comprising less than a dozen species, the 
androecium is tetramerous— i.e., isomerous with the corolla; in 
Oleaceae, represented by nearly 400 species, the stamens are two in 
number, with rare exceptions. 
This occurrence of oligomery of the androecium throughout an 
extensive group in which the flowers are regular is without parallel 
in the whole series of flowering plants. One can only suggest that 
the case of the oleaceae represents a special extension of the primary 
economy tendency, unassociated with any corresponding adaptation, 
on broad general lines, to insect-visits. 
It seems, therefore, in any case, that we must regard the 
Oleaceae as relatively advanced, and that we must not lay too much 
stress upon the indications of polypetaly which they display. Taken 
in consideration with the other floral characters this polypetaly 
may not impossibly be secondary, as also in the Salvadoraceae, in 
which order a dialypetalous corolla and a hypogynous androecium 
characterize most of the few species. It is, at the same time, 
quite possible to regard this polypetaly as primitive, without viewing 
these two orders as primitive on the whole relatively to the rest of 
the Contortae. Their extreme advancement on the line of economy 
in production of parts, in fact, has led the writer to regard them as so 
far isolated that they may claim the rank of a separate cohort, which 
might suitably be named the Jasminciles. For they differ essentially 
in most of their characters, in addition to those already mentioned, 
from the typical derivatives of the apocynal stock. Contorted 
aestivation of the corolla is of such very general occurrence among 
these latter as to have suggested Engler’scohortal name—Contortae; 
in Oleaceae and Salvadoraceae the aestivation is prevailingly imbricate. 
The relatively frequent occurrence of compound leaves among 
Oleaceae ( Jasminum , Fraxinus ) is another exceptional feature —and 
is, indeed, exceptional for all sympetalous shrubs and trees, which 
usually have simple leaves.’ A further sign of advance in this 
group, and of considerable interest from the general standpoint of 
evolutionary progress, is the appearance of what may be named 
1 Apart from Oleaceae, the Bignoniaceae, and the genus Vitex of 
Verbenaceae, are the only important exceptions to this 
statement, 
