312 



XL. SAPINDACBiE. [Dodonm. 



Leaves flat, more or leas ouneate, entire or toothed at the end, rarely 

 exceeding IJin., and usually under lin. 

 Much-branched, erect or divaricate shrubs. Terminal flowers clustered 

 or shortly racemose. 

 Leaves broad-cuneate, rounded or truncate at the end ...... b- JJ- cuneata. _ 



Leaves narrow-cuneate, rather acute, acuminate or 3-toothed at the end 7. -D. pedimculans. 

 Leaves linear-filiform, heath-like or pine-like, 1 to Sin. long, not crowded 8. D.fihfoha. 

 Branches terete or nearly so. Leaves linear or linear-cuneate, obtuse, 



mostly under IJin. long 9- ^- lohulata. 



Sekies II. Platypterae.— icaues quiU entire, fiat. Wings of the capsule very divergent 

 or divaricate, not reaching to the style nor to the base, each carpel, including its wing, broader 

 than long, transversely ovate or oblong. 

 Leaves linear or lanceolate. Branches very angular. Dissepiments 



persisting on the axis 10. D. trumcatmUs. 



Dissepiments splitting and coming off with the valves. 



Leaves broadly, somewhat bluntly, lanceolate n. D. Hansma. 



Leaves narrow-linear W. B. stenophylla. 



Sekies III. Apterse.— iem'es entire or toothed. Capsule without wings, or the angles 

 slightly and irregularly dilated into very narrow wings. 

 Sepals lanceolate. Buds ovoid or globular. Dissepiments persisting on 



the axis of the fruit'. Branches scarcely angled. Leaves obovate, 



cuneate, or triangular, glabrous or pubescent. Flowers mostly axillary. 



Sepals narrow, short 13. D. triangularis. 



Series IV. FinnatS. — Leaves all pinnate or very rarely a few simple ones at the base of the 

 branches. Capsule of the CycloptersB, except in D. oxyptera and D. inoequifolia, where it 

 approaches that of the PlatypterEe, and in D. .humilis, where it is apterous. 

 Tall shrubs or small trees. Leaflets flat, oblong, lanceolate or obovate, not 

 coriaceous. Racemes or panicles terminal, loose. 

 Leaflets usually numerous, lanceolate or oblong. Capsule not inflated, 

 the wings broad. 

 Leaflets J to lin.; rhaphis broadly winged. Sepals 1 to IJ lines . . 14. D. megazyga. 

 Leaflets 5 to 27, flat, broad-linear, J to lin. long ; rhachis winged, 



almost f line long 15. D. macrozyga. 



Leaflets few, obovate or oblong. Capsule large and inflated . ... 16. X). physocarpa. 



Much-branched, leafy shrubs. Pedicels solitary or clustered (racemose in 

 jD. multijuga and D. pinnata). 

 Leaflets obovate, cuneate or oblong, often toothed, the margins usually 

 recurved or revolute. Plant usually pubescent or villous (except D. 

 humilis). 

 Capsule winged, hirsute at least when young. 

 Villous. Leaflets 7 to 20 or more ; rhaohis winged. Sepals 

 acuminate. Capsule-wings rounded. 



Pedicels long, clustered 17. Z). vestita. 



Pubescent. Leaflets 3 to 7 ; rhachis angular. Pedicels short. 



Sepals obtuse. Capsule wings acutangular 18. D. oxyptera. 



Plant shortly hairy. Leaves very small, terminal leaflet undivided 



or bifid 19. D. Macrossani. 



Capsule winged, glabrous or very sparingly pubescent. Plant 

 pubescent or rarely glabrous. Leaflets usually under 11. Pedicels 



short, clustered 20. Z). boroniafolia. 



Leaflets narrow-linear, convex underneath. 



Capsule-wings rounded ; dissepiments splitting and coming off with , 



the valves. Leaflets under 15_ 21. D. adenophora. 



Differing from the last in elongated pedicels, dissepiments never 



separating from their axes, cells longer and the wing narrower . . 22. D. tenuifolia. 



In the following 9 species, great as is the diversity in the size of the capsule and the precise 

 shape of the wings, these differences afford no specific characters, and are often very difficult to 

 class as varieties, even when perfectly ripe and well-formed capsules are obtained ; and the 

 shape of the wing often alters much during growth, or is apparently affected by the manner in 

 which the capsule has ripened. The very shining seeds distinguish two species, but where they 

 are usually opaque they sometimes are somewhat shining. There remains little but the very 

 uncertain characters derived from, foliage to separate all these species, which are yet much too 

 constantly dissimilar to be united into one.— Benth. 



