ILLUSTRATIONS OF INDIAN BOTANY. 211 
CARALLIA. Hl. Mal. v. t.13 (inti i i 
lL. C. Ceylanica, vee foliis il sme obin- delineatis.) Pai aie l aes: ee 
Sissimis subretusis latitudine subdimi io-longioribus, 3. C. Sinensis, (Arn.); foliis cuneato-obovati i 
Arn. l.c.; Wight Cat. n. 2446—C. obcordata, Wight ac obtuse eaten latitudine subdaplo Wnighactone, 
én Git, is Arn. 1. c. (cum, syn, 
2. C. corymbosa, (Wight); foliis oblongo-obovalibus 4. C. integerrima, (D.C.); foliis ovalibus subacuminatis 
obtusis vel obsoleto ac obtuse acuminatis latitudine latitudine duplo longioribus, 
2—24 plo longioribus.~— Wight Cat. n, 2447.—Rheed. 
EXPLANATION OF PLATE 89. 
1. Kandelia Rheedet, (W. & A.) Flowering branch—~ 5-6. Ovary cut transversely and vertically, 3 celled, 
natural size. with two pendulous ovules in each. 
2. A dissected flower to show the relative position of 7. A fruit at the commencement of germination, 
“ye - A fruit after germination has considerably ad- 
3. Stamens. . vanced nearly a foot in length—natural size. : 
4. Stigma. 9-10. Sections of the fruit and tigellus—natural size, 
with the exceptions mentioned, all more or less magnified. 
EXPLANATION OF PLATE 90. 
1. Flowering branch of Carailia Ceylanicus—natural 4. A en detached. 
, 5-G. Ovary cut transversely and vertically—all more 
2. A dissected flower. or less magnified. fe ph! 
8s, 
3. Stamen $ 
LX.—COMBRETACES. 
This is one of the most strictly tropical orders we have yet had to examine, for, though 
some species extend beyond these limits yet none go beyond the warmer latitudes on either 
side 
Arrinities. These are not easily defined according to DeCandolle, the polypetalous 
genera approach Myrtaceae while the apetalous ones have a closer affinity with Santalaceae and 
Elaeagneae, and even with Laurineae through Gyrocarpus. Combretum differs from the rest 
of the order in having quaternary flowers with 8 stamens, and folded not spirally convolute coty- 
Jedons. From Myrtaceae and Onagrarieae and Memecyleae, they differ in their 1-celled ovary 
and pendulous ovules. From Santalaceae and Elaeagneae they are distinguished by the folia- 
ceous convolute or plaited, not fleshy cotyledons. This last structure, which separates them 
from all other orders, allies them with Gyrocarpeae which has spirally convolute cotyledons, but 
from which they are easily distinguished by the longitudinal, not valvular, dehiscence of the cells. 
of the anthers. “ The solitary carpel of which the fruit consists is peculiar to these and to 
Alangieae, and neatly distinguishes those two orders from all others of the myrtal alliance” 
Lind. Upon the whole the weight of authority is in favour of the present station as all those 
‘Writers Sto tive given much of their attention to natural affinities agree in placing Combreta- 
_ €¢ae among the series of orders with which they are here associated. = ee 
