218 FLORA INDICA. {Berberkha . 



tliera broke down under every successive author's hands, so that each, thinking his 

 own species new, because not agreeing with the descriptions of his predecessors, de- 

 scribed them as such accordingly. Lastly, we have compared our notes and observa- 

 tions with the results arrived at by Madden, Strachey and Winterbottom, Wallich, 

 Edgeworth, Koyle, and others, and find that none of these botanists agree with one 

 another nor with us in their views of the limits of the forms. 



Under these circumstances we have felt it incumbent upon us to devote a great 

 deal of time to studying the variations of each organ, and the result has been to re- 

 duce the species to a few well-marked forms ; uuder these we have ranged the spu- 

 rious species as varieties, retaining, however, the specific names they bore, so that 

 they may be applied as such by those who take a different view of the value of spe- 

 cific characters to ourselves. "We have also pointed out, under each variety, its rela- 

 tions to the other varieties of the same species, and to those of other species. 



The following remarks on the variations of organs, etc., may be useful. 



As regards habit, the species, without exception, vary extremely, many of them 

 from tall bushes with twiggy branches to prostrate stunted shrubs, according to cold, 

 and the degree of exposure to winds and drought ; a reduction of leaves and stipules 

 to spines, of racemes to fascicles of flowers, a shortening of the peduncles and pedi- 

 cels, a reduction in the size of the flowers and of the leaves, with additional coria- 

 ceousness, and sometimes the development of glandular pubescence and glaucous 

 bloom, are all characters more or less directly attributable to elevation, exposure, 

 cold, or drought ; it is however to be remarked, that an increased size and fleshiness 

 of berry often accompanies these changes. The spines are more usually 5-fid in the 

 dry conutry forms than in those from humid localities. 



There is no natural or constant distinction into evergreen aud deciduous-leaved 

 species ; for, though some species or forms, as B. hycioides, Asiatica, and NepaleTisis 

 are always persistent-leaved, and the common form of B, vulgaris is always deci- 

 duous-leaved, the forms Crelica aud Sinensis of the latter have often very persistent 

 foliage, and the duration of the leaves of B. aristata entirely depends on the depth of 

 forest, and the amount of light, heat, and moisture to which it is consequently exposed. 

 The many forms of this plant which have been raised in Kew Gardens, from seeds 

 sent home by ourselves and others, we find to present every variety in amount of 

 persistence ; and after three years' observation we conclude, that in certain seasons 

 some are wholly deciduous which in others are quite persistent, and that the period 

 at which the cold arrives has a difierent effect on different varieties. We also observe 

 that much depends on the ago of the plant, and that different parts of the shrub are 

 very diflerently affected. 



The size, toothing, and cutting of the leaves, and of the opposite sides of each leaf, 

 vary extremely in all the species, as does the number of leaves in each fascicle, in 

 all parts of the individuals. The rapidity with which they colour is equally variable ; 

 those alpine species which are in the upper temperate Himalayan regions exposed 

 to sudden frosts, redden rapidly, converting green mountain-slopes into bright-red 

 in two nights. The racemes of flowers are often more or less cymose, the pedicels 

 being more or less fasciculate ; these and the peduncles vary extremely in robustness, 

 aud are sometimes almost fleshy and very glaucous. We have been unable to connect 

 the varions forms of inflorescence with habit, further than that, as stated above, there 

 is a reduction of all parts in alpine forms. Though the extreme states of B. aris- 

 tata, with racemose and cymose inflorescence, are extremely unlike, we have gathered 

 specimens on which these occm- on one and the same branch ; we have also found 

 stunted specimens of the same plant with solitary axillary pedicels, wholly resembling 

 B. angulosa in this respect, which is typically one-flowered. 



We have devoted especial attention to the variations of the flowers and fruit, be- 

 cause, in all polj'petalous genera, in which there is a gradual transition from bracts 

 to petals, the floral envelopes all vary extremely in relative size and form. The 

 petals themselves are notched, entire, or bifid sometimes in the same species, specimen, 

 aud even flower, aud vary from being larger than, to smaller than the sepals. The 

 size, position, and prominence of the glands at the base of the petals is a most falJa- 



