328 MR. J. MIERS ON THE HIPPOCRATEACEA OF SOUTH AMERICA. 
11. Finally, the signally anomalous arrangement of the floral parts in the tribe Kip- 
pistiee also tends to sever all connexion between the two families. 
Having now enumerated the many strong points which tend to separate these two 
orders, it is only fair to examine, on the other hand, all the characters that can be urged 
in favour of the intimate relationship advocated by many botanists. These may be sum- 
marized in a brief space—their arborescent habit, their evergreen opposite leaves (more 
often alternate in Celastracee), their axillary panieulated inflorescence, their small 
flowers, the presence of five sepals and five petals, with imbricated æstivation, a 3-celled 
ovary, simple style and stigma, and the axial placentation of the ovules. But as these 
circumstances are common to many other natural orders, they cannot serve as evidence 
in favour of any special affinity between the two families, or counterbalance the many 
peculiarities that distinguish the one from the other. The feature which appears to me 
our most important guide in this question of relative affinities, is the place of the disk, 
serving, as it does, to regulate the position of the petals, stamens, and ovary in regard to 
each other. Searching, therefore, among the hypogynous polypetalous orders, with a 
superior ovary of consolidated carpels, with axial placentation and simple style, and 
beginning with those where the stamens are inserted upon the disk—then, through 
another group, where they are placed within the disk, which thus insulates them from 
the petals—and finally where they are inserted more or less completely outside the 
disk, we have thus in linear succession, omitting certain intervals, the following Orders :— 
Aceracee, Amyridacee, Eleocarpee, Tiliacee, Erythroxylaceæ*, Chlenacee, Hippocra- 
leacee, Goupiacee ; Linacee, Humiriaceæ, Chailletiacee, Icacinacee, Aquifoliacee, 
Celastraceæ, Rhamnacee, Staphylinacee, Sapindacee, &c.; and it is therefore in these 
directions that we should seek the affinities of the Hippocrateace and Celastracee. 
It may be useful to make here a few passing observations on the genus Alzatea of the 
‘Flora Peruviana.’ I have already alluded to the Ceylonese genus Kokoona, rightly placed 
in Hippocrateacee by Dr. Thwaites, but which was removed into Celastraceæ on account 
of a supposed affinity with Alzatea. This anomalous genus was established in 1794 by 
Ruiz and Pavon, on a tall tree from the Peruvian Andes, with erect branches, opposite 
or verticillated leaves, and a copious pyramidal inflorescence of many verticellated 
branchlets, ultimately terminating in numerous 1-flowered pedicels. The flowers, of 
middle size, consist of five fleshy petaloid sepals, with valvate æstivation, which are eon- 
tinuous with a broad cupuliform torus, upon which a disk of similar form, extending to 
the base of the sepals, is agglutinated ; there are no petals; but they have five short introrse 
stamens, not situated in the usual manner, opposite the sepals, but alternating with them, 
being inserted outside the disk in each sinus, between the sepals. A superior ovary is 
seated'in the middle of the disk, surmounted by a short style and an obtuse 4-lobed 
stigma; it is 2-celled, with very numerous ovules crowded on each side of the dissepi- 
ment. The fruit is a small fleshy capsule, surrounded by the persistent rotate sepals: it 
is globular, somewhat compressed, bilocular, splitting loculicidally along the margins, 
* In this family I have noticed that the stamens are not actually monadelphous, as usually described, but are 
inserted within the mouth of a tubular hypogynous disk. On this account, and for other reasons, the Erythroaylacee 
have an unsatisfactory position in Linaceæ, where they are placed as a mere tribe by Messrs. Bentham and Hooker. 
