60 MB. BENTHAM ON LOGANIACE2E. 



up before arriving at maturity. In both cases the seeds had been 

 already shed. 



I have not seen the Sumatra plant published by Blume under 

 the name of Leptopteris Suniatrana; but neither in his description, 

 nor in his figure of the flower and its analysis can I find anything 

 to distinguish it specifically from the Hong Kong plant. 



Tribe II. EULOGANIEJ3. 



This, the original group upon which the order was constituted, 

 presents also in the strongest degree its peculiarities and diffi- 

 culties. Representing on the one hand the Hedyotidece among 

 Bubiacece, and passing into them by the most gradual steps 

 through Houstonia, it is, on the other hand, as closely connected 

 through Buddleiece with Scrophularinece, and through Geniostoma 

 forms the nearest approach to Apocynece. Towards Bubiacece, a 

 slight adherence of the ovary at its base will perhaps justify the 

 including Houstonia, as proposed by Torrey and Gray, within the 

 boundaries of that family, to the exclusion of Spigelia and Mi- 

 treola ; but, on the opposite extremity, I see no way of establish- 

 ing a distinctive character between Loganiacece and Scrophularinece 

 without bringing over Buddleia and its allies into the domain of 

 the former. The aestivation, upon which I had formerly relied, 

 under the mistaken supposition that it was, as described by 

 Endlicher, contorted in Logania as in Geniostoma, proves of no 

 avail. I had then overlooked the observations of Alph. DeCan- 

 dolle, which I have since verified in a number of species, that it 

 is imbricated with one external lobe in Logania as in Buddleia. 

 The stipules are occasionally reduced in Logania to a slight con- 

 necting line, which always exists in Buddleia, and in some species 

 is expanded into foliaceous appendages, which, although not called 

 by the name of stipules, appear to be of the same nature. The 

 capsules and seeds are essentially the same in both genera, and 

 even in inflorescence and general habit some species of Logania 

 differ very little from Gompliostigma and some of the entire-leaved 

 Buddleias. It is true that most species of the latter genus have 

 dentate leaves, an element hitherto unknown in Loganiacece, and 

 till lately also in the vast family of Bubiacece ; but now, in the 

 instance of Carlemannia, we are obliged to admit dentate leaves as 

 an exceptional character in Hedyotidece, and by analogy we cannot 

 exclude it from Buloganiece. The line of demarcation between 

 Ltoganiacece and Scrophularinece cannot therefore be drawn between 



