IN CEYLON. 
105 
stamens. In D. crumenata the staminodes of each female 
flower are in number an exact multiple of the corolla seg¬ 
ments and are disposed opposite to and alternating with them, 
whereas the male flower resembles D. qusesita and D. pruriens 
in the indefinite hypogynous character of the stamens. 
There is one species—D. Toposia—in this group which 
shows an unstable staminal whorl in the female flowers, the 
number usually being greatly in excess of the members of 
other whorls, and presents no fixed orientation to them. 
Each female flower of D. Toposia may possess from twelve 
to sixteen staminodes, which form an epipetalous or partly 
hypogynous ring; the stamens of each male flower may be 
from ten to over thirty in number, of unequal length, form¬ 
ing a hypogynous group terminating the receptacle. 
VII. — AFFINITY. 
The phylogeny of the Ebenales has formed the subject of 
much discussion. 
Hiem (p. 62) states that the Ebenacese shows the closest 
affinity to the Olacinese, Styraceae, Anonacese, Ternstrce- 
miaceae, Sapotaceae, and Ilicinese, but also draws attention to 
the points of difference between these natural orders and the 
Ebenaceae. He also states (p. 63) that other orders including 
Tiliacese, Magnoliaceae, and the Euphorbiaceae show affinity 
in a less degree to the Ebenaceae. 
Miers compares them with the Anonacese and suggests 
their grouping among the polypetalous section; Choisy 
points out their proximity to the Ternstrcemiaceae. 
Parmentier, in his monograph on the histology and 
morphology of the Ebenaceae, does not commit himself to 
statements as to the phylogeny of the natural order. 
Hallier, on the other hand, points out that there are many 
features in which the Ebenales resemble many other natural 
orders, and in this respect goes into much more detail than 
Hiern. 
Under the circumstances I cannot do better than point out 
the main facts as supplemented in the present paper, and 
8(1)4 W 
