ILLUSTRATIONS OF INDIAN BOTANY. 209 



Tribe IV. Bokuage^e. Ovary consisting of 2 two-celled or separable carpels; Style 

 central baselar, springing from the gynobase. Fruit 2-4-parted, of 2 two-celled or 2-partibl« 

 carpels. Seed exalbuminous,— Shrubs or herbs. This last is divided into 6 subtribes, sometimes 

 differing from each other nearly as much as the tribes do among themselves. These I deem it 

 unnecessary to introduce, there being so few Indian species referable to this tribe. 



Affinities. De CandoUe briefly remarks of these, "an order approaching Labiatcc and 

 VerbenacecB by the structure of its fruit, and SolanacecB and Hydrophyllacece by its floAvers/^ 

 Lindley remarks of Cordiacece, "The plaited cotyledons and dichotomous style first led to the 

 separation of this order from Borrageworts, with which it was formerly associated, chiefly it 

 is to he supposed, on account of the roughness of the leaves. Von Martins remarks that it 

 is in fact much nearer ConvolvulacecB^ from which it differs in its inverted embryo and 

 drupaceous fruit. It seems to me impossible to admit CordiacecB even into the same category 

 as Borrageoe^ the indispensable pecularlties of which are gyrate inflorescence and nucamenta- 

 ceous fruit, neither of which circumstances occurs here." 



With respect to the last two circumstances mentioned, the first, the inflorescence, is nearly 

 met here In the decided tendency towards gyration observable in the accompanying species of 

 Cordia^ and the cotyledons are but very slightly plaited, thus showing so far a strong leaning 

 towards Ekretiaceas in both these points, but differing In the twice dicliotomous style, and 

 ascending not pendulous ovules: but the ovary Is the same in both, a more important character 

 than one derived from the mature fruit. 



In regard to Ekretiacece, Lindley remarks, "a branch of the old BorraginecB distinguished 

 by a terminal style, proceeding from the apex of a perfectly concrete ovary of 4 cells, a baccate 

 fruit, and seeds furnished with thin fleshy albumen. The order is recombined with Borraginecp 

 by Alph. De Candolle, but it seems sufficiently characterized by its concrete carpels, and the 

 presence of a small quantity of albumen. The separate not separable nuts of Borraginece 

 are so peculiar, notwithstanding Cerenthe has them combined in pairs, that a real objection 

 seems to exist to the disregard of so good a mark by the combination with them of these 

 concrete-fruited Ehretiads.'' 



Under his order Ehretiacece Lindley ranges De Candolle's 2nd and 3rd tribes. There is 

 no doubt a considerable difference between them and Borragece^ in the ovary and fruit, and espe- 

 cially in the hypogynous origin of the style of the latter, but in other respects w^e find much 

 affinity and a progressive transition- Between Cordia and Ehretia the greatest difference I find 

 IS in the position of the ovules, ascending in the one, pendulous in the other. The seed, too, dif- 

 fers, being exalbuminous in the former, and more or less copiously albuminous in the latter. The 

 only other distinction of ordinal value between the two is the twice divided stigma of Cordia 



tiacece) 



Tournefi 

 The different position of the ovule is uot alone sufficient to con- 



, stitute an ordinal character, and neither, I apprehend, ought so much importance to be attached 

 to the circumstance of the 4 carpels cohering into a single nut in one, and continuing distinct 



in the other. 



Tournefortia to Helioti 



easy, though the latter differs nearly as niuch from the former as EhretiecB does from Cor- 



Heliot 



as 



seems but little reason for separating Cordia and Ehretia. 



The passage from Heliotropece to Borragece is not quite so smooth as in the preceding 

 examples ; the difference in the ovary and fruit, and especially in the hypogynous insertion of 

 the style of the latter being greater. But on the other hand the similarity of habit is more 

 marked and of a kind, for the most part, so conspicuous as to lead even a cursory observer to 

 detect the relationship. It therefore becomes a question of some nicety to decide, whether 

 relationships easily detected by external characters is to give place to others deduced from 

 microscopic research? which those indicated above certainly are. % impression is that the 

 verdict will be in favour of the former ; and to that extent, confirmatory of De Candolle's views 

 ii this instance. 



Mm 



