284 DICOTYLEDONS. VIOLACEiE. 



2. D. longifolia, leaves radical obovate tapering below into a 

 long footstalk erect, scape with a simple raceme, p. 98. — Bogs. 



3. D. anglica, leaves radical oblongo-spatbulate tapering down 

 into a long footstalk erect, scape with a simple raceme, p. 98. 

 — Boss. 



OrdeuLXII. CISTE^. Juss. 



Cat. 5 -parted. Petals 5. Stamens indefinite. Ovary I. Style 

 and Stigma 1. CV/p5?//(? many-seeded, many-valved; some- 

 times l-celled, with the p/ocew/oi in their middle; sometimes 

 many-ceiled, with dissepiments proceeding from the middle of 

 the valves. Seeds numerous, round. Embryo recurved or 

 spiral, surrounded with a somewhat fleshy albumen. {Radicle 

 at the opposite extremity of the seed to the umbilicus. 

 Br. MSS.^) 



Shrubs or Under-shrubs. Leaves opposite, with or without 5/i- 

 pules. Flowers in racemes or corymbs. Mirh. 



1. CISTUS. 



Cal. of 5 leaves, 2 smaller than the rest. Pet. 5. Caps, with 

 3 valves and many seeds, 



1. C. Heliantliemum, somewhat shrubby procumbent, leaves el- 

 liptical oblong somewhat hairy mostly white and pubescent be- 

 neath, tlie margin slightly revolute, stipules lanceolate, p. 170. 

 — Dry pastures. 



OrderLXIII. VIOLACE^. renL 



Cal. 5-cleft. Pet. 5, irregular. Stam. 5. Anthers often united. 

 Style and Stigina I. Capsule l-celled, 3-valved, many-seed- 

 ed. Seeds attached to the middle of the valves. Embryo 

 in the axis of a fleshy albumen. Radicle opposite the hilum. 



Herbs with alternate leaves and stipules. 



I.VIOLA. 

 Cal. 5-cleft, the segments produced at the base. Upper petal 

 spurred at the base. 



^ We are obliged to Mr. Brown for the knowledge of this important pe- 

 culiarity in the structure of the seeds of Cistus. The true direction of the 

 radicle had been previously overluoked by carpological botanists, (unless 

 Gnertner intended to indicate it by the term " radicula centripeta" which he 

 applies to Cistus and Helianthemum,) and, as Mr. Brown has remarked to us, 

 offers, by this curious anomaly, a character by which not only the limits of 

 Cistecp, can be positively defined, but which indicates the affinity of Lccliea 

 and Hudson'm to this order, in both which he has observed the same struc- 

 ture. 



