MATERIALS FOR A FLORA OF FORMOSA. . 9 
exception, two parietal placente, which connate themselves in the 
innerside of the capsules. But, it is quite otherwise with the 
present plant, which bears only a placenta rising by itself to mid- 
way from the base of the ovary, ie. a central placenta, as is the 
case with Scrophulariaceous plant. The genus is, therefore, not to 
be referred to Hydrophyllacezx, but should be referred to 
Scrophularinesx. There is a little difference between his state- 
ment and Mr. Maxrno’s drawings, regarding the placenta. So far 
as I can judge from the latter figures, the ovary is perfectly two- 
celled, with axial placente. But, in the figures given in Hooxer’s 
Icones Plantarum, the ovary is nearly, but not perfectly, two- 
celled, as the placentz are free at the apex. Although these 
differences exist, they are, in my opinion, to be regarded as 
results of individual mutability, to which all plants are more or 
less subject. Taking all these things into consideration I think 
the plant should be referred to Scrophularinee, as was stated 
formerly by Bartton and quite recently by Prof. Brann. Finally, 
I may mention one more plant which is also extremely difficult 
as to its reference. It is Tvriplostegia glandulifera Watt. The 
genus was first established by Watiicn on a Himalayan plant 
in his manuscript and also in DC. Mem. VII. t. 4, referring it 
to, Valerianes, as we °see in, “DO yp Prodm iv, p. 642.) 
The genus is, however, transferred to Dipsacee by BENTHAM 
and Hooxer, f. in their Genera Plantarum II. p. 158, with the 
remark ‘Genus hucusque ad Valerianeas quibus inflorescentia 
accedit adscriptum, sed involucellum, semen albuminosum, folia 
et indumentum omnino Dipsacearum et odore Valerianearum 
caret,” and also on p. 1230, with another remark, ‘‘ Stamina per- 
fecta sepe 3 tantum ut in Valeriana affinibusque, et genus 
omnino medium tenere videtur inter Valerianeas et Dipsaceas, 
