504 Prof. Don's Descriptions of Indian Gentianese. 



Although the Gentianece undoubtedly constitute a very natural family, 

 agreeing remarl<:ably in their habit and structure, and also in their sensible 

 properties, they afford very few absolute marks to distinguish them from the 

 other families to which they are related. When taken in an extended sense, 

 the Gentianece may be said to hold an intermediate station between Apocynece 

 and Rubiacece, differing from the former, to which they are more intimately 

 allied, in the larger quantity of albumen, and in the much greater develop- 

 ment of their embryo ; from the latter in their free ovarium, and from both 

 by their persistent corolla, and in the nervation of their leaves. We may 

 compare Craivfurdia with Gelsemium, of which it possesses the twining habit 

 and most of the characters, but the latter is essentially distinguished by its 

 penninerved leaves, deciduous corolla, and concrete carpels, which unite it to 

 Apocynece. Some species of Lisianthus resemble Allamanda in their woody stem 

 and in the structure of their flower, and the twisted aestivation of Apocynece 

 occurs also in Erythrcea and Gentiana contorta. There is an evident affinity 

 between the Rubiaceous genus Oldenlandia and Mitrasacme, which also ac- 

 cords in many respects with Spigelia and Mitreola, but it differs in the imbricate 

 aestivation of its corolla. Seeing, however, the near approach to the valvate 

 form of aestivation in Slevogtia, and that Spigelia and Mitreola agree with 

 Gentianece in habit, I am led to question the propriety of considering them 

 in any other light than as forming a subordinate group of that family. The 

 genus Canscora, by its irregular flowers, and by its resemblance in habit to 

 certain Gratiolex, especially to Torenia, would seem to establish a relation- 

 ship between the Scrophularinea; and this family. 



The essential characters of Gentianece consist in their persistent usually 

 plicate corolla; in the two carpels composing the pericarpium being placed 

 right and left with respect to the axis of the flower ; and lastly, in the nerva- 

 tion of their leaves, which bears a considerable resemblance to that of Mono- 

 cotyledonous plants. These characters only apply to the normal Gentianece, 

 and necessarily exclude the three small groups of Spigeliacecv, Loganiacece, and 

 Potaliaceae, which Dr. von Martius has proposed to separate from them. In all 

 these, however, the relation of the carpels to the axis of the flower is the same 

 as in Gentianece, but they have all a deciduous corolla, and in the last two 

 the leaves are penninerved. Another group, the Menyanthece, consisting of 



