30 



Geunsia. 



2. The petioles of the opposite 2. such a Hne is either absent, 

 leaves are always joined or not conspicuous. 



by a characteristic interpe- 

 tiolar margin of hairs. 



3. The anther-cells open by a 3. The anther-cells open by 

 typical widening of the a long fissure, from the 

 upper part of the parallel apex to the base, 

 fissures; (it is not a mere 



apical hole, as many authors 

 pretend). 



4. The ovary is {3-)5-celled; 4. The ovary is 4-celled; the 

 the cells are 2 seeded. cells are 1 -seeded. 



This, and some other, less easily definable features, deter- 

 mine the general habit, which is typical for each genus, and 

 which enable us, to separate the two genera, even th^n, when, 

 as occurs, we meet with specimina, which are 4-merous and 

 5-merous, and have the same habit. The latter characteristic, 

 taken before as the criterion for separating the two- genera, 

 is not sufficient for this purpose, since we discovered 4-merous 

 specimina, which without any doubt belong to Geunsia, and 

 5-merous ones, which belong to Callicarpa. Many species of 

 Geunsia show a close affinity with certain species of Calli- 

 carpa. This may indicate the phylogenetic relation between 

 them, and may give rise to the supposition that the Calli- 

 carpa-species are developing from the Geunsi'a-species, which 

 may just be in a period of active mutation '). There are a 

 number of facts, which give an indication into this direction. 

 So may f. i. : 



C longifolia vac. a be developed from G. pentandra or G. 

 acuminatissima. 



C. longifolia var. § be developed from G. farinosa. 



C. lanata have the same origin as G. hexandra. 



C. basilanensis be developed from G. cumingiana var. /3! etc. 



The 6 — 7-merous Geunsi'a-species may be another branch 

 of phylogenetic development, arisen from an original 5-me- 

 rous form. 



^) See: ..Systematical Notes". 



