43 



XXXII. TERNSTROMIACEiE. 



TstLSSTRoMi-EJE, Mirb. Bull. Philom. SRI. (1813.) — Ternstromiacetf., Dec. Mem. 

 Soc. H. N. Genev. vol. 1. (1823); Prodr. 1.523. (1824); Camhesstdes Mtmoire 

 (1828.)--TnEACE^, Mirb. Bull. Phil. (1813.)— Camellie^, Dec. Theor.Elem. 

 ed. 1. (1813) ; Prodr. 1. 529. (1824.) 



Diagnosis. Polypetalous dicotyledons, with hypogynous, indefinite, 

 monadelphous, or polyadelphous stamens, concrete carpella, an ovarium of 

 several cells, with the placentse in the axis, a persistent imbricated many- 

 leaved calyx, alternate simple leaves, and definite seeds. 



Anomalies. Cochlospermum has the ovarium 1 -celled, with imperfect 

 septa, to the margins of which the ovula are attached. Leaves very rarely 

 opposite. Cambesstdes. 



Essential Character. — Flowers very rarely polygamous. Sepals 5 or 7, imbri- 

 cated in {estivation, concave, coriaceous, deciduous, the innermost often the largest. Petals 

 5, 6, or 9, equal in number to the sepals, often combined at the base. Stamens very nume- 

 rous ; _/f7tf/ne«<A- fihform, monadelphous, or polyadelphous; anthers versatile, or adnata. Ova- 

 rium superior, with several cells ; styles from 3 to 7, filiform, more or less combined ; ovules 

 pendulous, or erect, or peltate. Capsule 2-7-celled and capsular, with the dehiscence taking 

 place in various ways ; sometimes coriaceous and indehiscent ; usually with a central 

 column. Seeds large, attached to the axis, very few ; albumen none, or in very small 

 quantity ; embryo straight, bowed or folded back, the radicle turned to the hilum ; cotyledons 

 very large, often filled with oil, occasionally plaited lengthwise ; an arillus sometimes 



present Trees or shrubs. Leaves alternate, coriaceous, without stipulce, usually undivided, 



now and then with pellucid dots. Peduncles axillary or terminal, articulated at the base. 

 Flowers generally white, seldom pink or red, very rarely (in Cochlospermum) yellow. 



Affinities. This order originated in 1813, with M. Mirbel, who 

 separated some of its genera from Aurantiacese, where they had been 

 placed by Jussieu, and at the same time founded another closely allied 

 order, under the name of Theacese. These opinions were substantially 

 adopted by Messrs. Kunth and Decandolle, the latter of whom, moreover, 

 formed several sections among his Ternstromiaceae. It is, however, certain, 

 that no solid difference exists between this last order and Theacese or 

 Camelliese, as they were called by Decandolle ; and Cambessedes, after a 

 careful revision of the whole, has come to the conclusion, that even the sec- 

 tions proposed by Decandolle among Ternstromiacese are untenable. I shall 

 profit by M. Cambessedes' observations in all I have to say upon the order. 

 Ternstromiaceae may be compared, in the first place, with Guttiferae, 

 with which they accord more closely than with any thing else, and in the 

 affinities of which they entirely participate. They differ thus : in Tern- 

 stromiaceae the leaves are alternate, to which there are scarcely any excep- 

 tions ; they are always opposite in Guttiferae. In the former the normal 

 number of the parts of the flower appears to be 5 and its multiples; in 

 Guttiferae it is evidently two. In the former the calyx is always perfectly 

 distinct from the corolla ; these two organs are usually confounded in the 

 latter. TernstromiacetB have the petals generally united at the base, and a 

 twisted aestivation ; in Guttiferae they are distinct, with a convolute aestiva- 

 tion. The seeds of the former are almost always either destitute of albumen, 

 or furnished with a membranous wing; the latter have neither the one nor 

 the other. The first have the radicle always near the hilum; the second 

 have it either near the hilum or turned in an opposite direction. Finally, 

 in Guttiferae the cotyledons are very thick, and firmly glued together ; and 

 this character, which is not observed in Ternstromiaceae, is the more im- 

 portant, as it is not liable to any exception. Ternstromiaceae are allied 

 to Hypericinese through the medium of Carpodontos, a genus which, with 

 the foliage of the latter order, has the fruit of the former; and also of 

 certain plants of Hypericineae, which, according to Cambessedes, have 



