LINNEAN SOCIETY OP LONDON. xlvii 



This comparative paucity of species is probably real, but not to 

 the degree that Grisebach was led to suppose from the scanty data 

 he had access to. He does not appear, in making his calculations, 

 to have yet seen even the first volume of Oliver's ' Flora of Tropical 

 Africa;' and he judges chiefly from Hooker's 'Niger Flora' and 

 Achille Richard's ' Flore d'Abyssinie,' which he regards as tolerably 

 fair exponents of the vegetation of the two best-known districts of 

 the region, now termed by Oliver Upper Guinea and Nileland. 

 Comparing the plants of these two districts enumerated in the two 

 volumes now published of Oliver's 'Flora' with the corresponding 

 portions of the two above-mentioned works (the orders preceding 

 Umbelliferae), we find the Abyssinian or Nilelaud species increased 

 from 562 to 853, and those of Upper Guinea from 747 to 1091 ; and 

 Grisebach would probably have to raise his number of 1650 Abys- 

 sinian pheuogamous species to about 2500, and the 1870 from Upper 

 Guinea, to about 2800. The total pheuogamous species in our her- 

 baria now ready to be entered in the Tropical-African Flora cannot 

 be far short of 8000 ; and there is, I think, little doubt that 

 several thousands may be yet to be added to them from the vast 

 tracts of country entirely unknown to botanists. But even this 

 increased number may not be more than half of what could be sup- 

 plied from the much smaller area included in the Brazilian empire, 

 the extraordinary richness of whose natural productions, animal as 

 well as vegetable, has been frequently commented upon ; it may 

 also, as stated by Grisebach, fall considerably short of the probable 

 number in his Indian Monsoon region, which, from the Himalaya to 

 the north coast of Australia, has an extent in latitude about equal 

 to that of the Soudan region, with a few more degrees of longitude, 

 from the Indian peninsula and Ceylon to the extreme east of New 

 Guinea. But might not this difference be in some measure accounted 

 for by some of those considerations which he so positively rejects as 

 irrelevant? If it be true that in plants the production through 

 natural selection of new races from variation is favoured by changes 

 in cUmatological and other physical conditions, whilst a long con- 

 tinued uniformity of these conditions enables races once acclimatized 

 through a long course of generations by that same natural selection 

 to hold their own even long after they have become reduced or 

 weakened by age — if we may further consider the number of highly 

 differentiated, monotypic, or sparingly varied races endemic in Africa, 

 and especially those which are intermediate between subgenera, 

 genera, tribes, &c. which in all other countries are well defined, to 

 be remnants of races of the highest antiquity, may we not regard 



