Berbem.] flora indica. 217 



sperma, seminibus erectis. Embryo majusculus. — Fratices Ugnoflavo, 

 foliis pinnatis v. suppressione pinnarum lateralium simplicibus, foliolis 

 stipulisque sa?pe in spinas abeuntibus, ftoribus Jlavis. 



Berberis, including Mahonia, is a perfectly natural and veil-defined genus, whose 

 species, however, are so singularly sportive in habit and all characters, that it is im- 

 possible to form any accurate estimate of its extent. One hundred have been enume- 

 rated, which number may no doubt be reduced by one-half. Both botanical authors 

 and horticulturists have long been aware of the extreme difficulty of limiting the 

 species of this genus. Of its sportive character the European B. vulgaris is a good 

 example, upon which we are the more anxious to dwell, both because this plant occurs 

 in its normal English form, and in many abnormal states in the Himalaya, and be- 

 cause it is of the utmost advantage to us, who press upon the attention of our fellow- 

 botanists an amount of variation in mountain and tropical plants which they are 

 slow to believe, to have such an example of variation in Europe to quote. With the 

 B. vulgaris, in its ordinary north of Europe form, most botanists are familiar ; but 

 this is so unlike the Mediterranean forms, that two were described as different, one 

 by Linnaeus and Sibthorp, under the name of B. Cretica, and another by Rcemer and 

 Schultes as B. JEtnensis, species that are now considered, by some of even the mos 

 critical European botanists (Boissier and Cosson and Gussone), as forms of B. vul- 

 garis; and it is this prominent fact to which we desire to draw attention at the out- 

 set, that none of the Himalayan forms we here reduce to B. vulgaris differ more from 

 the typical state of that plant than do B. Sinensis and B. Cretica. The B. crata- 



gina, DC., of Asia Minor, and B. emarginata, Willd., of Siberia, appear to us to 



*~ o^««:«« ri; Q + ;r»«+ir»r* tlinn ^Fjrt/^isis ATifl Cv pf) r.a. and indeed 



[aim 



they have been reduced by some authors already ; and if to these be added the B. 

 Canadensis of North America, the geographical range of the species will then be 



from Siberia westward to the lakes of Canada. . 



In the Himalaya Dr. Wallich distinguished nine species, all differing widely in 

 general appearance from one another, and from B. vulgaris; many of them also in 

 specific characters. To these (three of which are founded on error) others have been 

 added, whicb, being found further west than Dr. Wallich's species, approached nearer 

 to the European types, without, however, so resembling the common state of L. 

 vulgaris as to suggest a comparison with any of the varieties of that plant whicl 

 inhabit a similar climate ; these were consequently described as new. 



The first impression conveyed by reviewing the whole Himalayan genus, by laying 

 out our very large suites of specimens collected with a view to show variations was 

 the strong resemblance between the West Himalayan deciduous-leaved forms and the 

 European B. vulgaris, amounting, in Kashmir and Kishtwar specimens, to absolute 

 identity ; and that, proceeding eastwards and southwards, the more coriaceous-leaved 

 species prevailed, and soon replaced the others, in the form of B anstata and its 

 varieties; that in Tibet and in the drier regions of the lofty Himalayan valleys, we 

 evervwhere found small, stunted, excessively spinous species, with small extremely 

 coriaceous leaves, and racemes often reduced to umbels, and even to axillary single- 

 flowered pedicels ; and that, descending lower in the same valleys and to the foot oi 

 the bills along the whole length of the Himalaya, many of these appeared to pass 

 kv ; nMM «-vil ™^„*^«„ ;«*** *1ia lnro-P-lrnvpd hushv form of anstata, with cona- 



ceous foliage. 



dry lofty regions and in the lowei 

 ' marked forms and species, oftei 

 itermediate elevations, of interme 

 ihine all these into an inextricable 



plexus of species or forms that admitted of no absolute characters \ and the more com- 

 plete and extensive our materials, the more did the species blend. 



If from our collections we turn to the labours of others, we find that they 

 have terminated in an equally unsatisfactory manner. So long as botanists had few 

 -}'<<imens, these were easily divided into species; but the characters attributed to 



2 F 



