ILLUSTRATIONS OF INDIAN BOTANY. 73 



less concrete. Fruit a cremocarpium, or capsular, or baccate, or drupaceous, 2- or many-celled 

 Seeds one or many in each cell, in the former case attached to the apex, or more usually 

 to the base of the cell; m the latter to a central placenta. Albumen horny or Heshy 

 copious. Embryo straight or slightly curved, inclosed in the albumen : radicle turned to the 

 hilum : cotyledons fohaceous.— Leaves simple, entire, opposite (very rarely vcrticillate) : stipuleg 

 2 at the base of each leaf, entirely distinct, or cohering either with the leaf or with each 

 other, or both ways: their apex sometimes produced into bristles, sometimes into foliaceous 

 expansions resembhng vcrticillate leaves. 



Affinities. A very cursory perusal of the above ordinal character will show that, as 

 regards fructification, this is a very complex family. In certain points, however, there is much 

 unitormity. The calyx tube is adherent to the ovarium: the petals cohere, forming a monoi)e> 

 talous corolla with the stamens, more or less perfectly, adhering to the tube and, invariably 

 equaling them m number, the anthers introrse. So far a nearly uniform structure prevails' 

 Ihe ovary however vanes to a great extent both as to structure and contents. 

 - X " *n^ y'^-% ^i'e^.^w^"^^'^ it is 1-celled with a single, erect ovule. In Spermacocew it 

 IS two-celled with a single ovule in each, but its direction varies, sometimes it is pendulous 



Knoxia, 



(Borrerioy Meyers) occasionally i 

 and is then said to be peritropal. 



HedyotidecB 



with numerous ovules attached to a 



placenta, but in at least one species they are reduced to two or three, all except one of 

 which abort, leaving a 2-seeded capsule, as in Spermacocece . Hameliece differ from SpermacocecB 

 m having a many-celled ovary with numerous ovules in each, while they are themselves 

 separated from Guetardecs, which also have a many-celled ovary, by the latter having few or 



solitary ovules. •> >/ e> 



The Cinchoneas have two-celled capsules and numerous ovules, like Hedyotideae, but 

 their seed are winged. GardinetB have also 2-ceUed ovaries and many ovules, except in 

 Oardenia itself, which has a 1-celled ovarium, and numerous seed attached to several, usually 

 4, parietal placenta, and, as forming an exception to the structure of the tribe, ought not to 

 have been selected for the name, implying that it is the typical genus. 



In some of the genera of this section the seed are flattened and pendulous from the 

 apex of the cell lying over each other like tiles, or imbricated. In the CoffeecB tribe the ovary 

 18 2-celled with a single ovule in each, but the direction varies, being pendulous, erect, or 

 peltate, affording, according to the direction, good generic characters. From this imperfect 

 enumeration of variations in the structure and contents of the ovarium it will be seen that, 

 beyond its being adherent, good ordinal characters cannot be obtained from that organ, neither 

 can they be taken from the mature fruit, which is capsular, baccate, drupaceous or, like those 

 ot LmbellifercB, a cremocarp, the distinctive mark of the section PcederiecB, including Lygo- 

 S5ln!.f^^\^^ ^^°"^ ^^^' ^^ ^^^^ Botanists, been most improperly separated as the type of 



the ovary and 



a distinct order. 



Not 



mature fruit— this is admitted by all Botanists to be a "well-marked and strictly hmited order." 

 It 18 in truth a most natural one, readily distinguished from all others by three easily-observed 

 marks, but which must all be present to constitute a genuine Cinchonacious plant. These are 

 an adherent ovarium— monopetalous corolla— and opposite entire leaves with intermediate 



stipules. C aprifoliacece 



Cinchonads. Log 



acetB have a monopetalous corolla, opposite leaves and stipules, but a free ovarium; they 

 therefore are equally related, being only separated by a single distinctive mark, the free, not 



adherent ovarium, hence they, in like manner, may be looked upon as inferior flowered 

 Ctnchonads. 



This order, viewed as a whole, furnishes a strildng example of the value a constant charac- 

 ter, even when solitary, in the limitation of natural families. Here we have a family, including 



not fewer than 2500 species, kept together as one by the constant presence of opposite leaves 



