ANNOTATIONS AND ADDITIONS. 307 



perceive that even in the great families our knowledge of which 

 has been of late most strikingly enriched by the descriptions of bota- 

 nists we are still acquainted with only a small part of existing 

 plants. The Repertorium of Walpers completes Decandolle's Pro- 

 dromus of 1825, up to 1846: we find in it, in the family of Legu- 

 minosa3, 8068 species. We may assume the ratio, or relative 

 numerical proportion of this family to all phaenogamous plants, to 

 be -Jr as we find it ^ within the tropics, T V in the middle tempe- 

 rate, and 3^3 in the cold northern zone. The described Leguminosae 

 would thus lead us to assume only 169,400 existing phaenogamous 

 species on the whole surface of the earth; whereas, as we have 

 shown, the Compositsei indicate more than 160,000 already known 

 species. The discordance is instructive, and may be further eluci- 

 dated and illustrated by the following analogous considerations. 



The major part of the Composite, of which Linnaeus knew only 

 785 species, and which has now grown to 12,000, appear to belong to 

 the Old Continent; at least Decaiidolle described only 3590 Ameri- 

 can, whilst the European, Asiatic, and African species amounted to 

 5093. This apparent richness in Compositae is, however, illusive, 

 and considerable only in appearance; the ratio or quotient of the 

 family ( r ] T between the tropics, \ in the temperate zone, and -fj in 

 the cold zone), shows that even more species of Compositas than 

 Leguminosse must hitherto have escaped the researches of travellers; 

 for a multiplication by 12 would give us only the improbably low 

 number of 144,000 Phaenogamous species. The families of Grasses 

 and Cyperacese give still lower results, because comparatively still 

 fewer of their species have been described and collected. We have 

 only to cast our eyes on the map of South America, remembering 

 the wide extent of territory occupied by grassy plains, not only in 

 Venezuela and on the banks of the Apure and the Meta, but also 

 to the south of the forest-covered regions of the Amazons, in Chaco, 

 Eastern Tucuman, and the Pampas of Buenos Ayres and Patagonia, 

 bearing in mind that of all these extensive regions the greater part 

 have never been explored by botanists, and the remainder only im- 

 perfectly and incompletely so. Northern and Central Asia offer an 

 almost equal extent of Steppes, but in which, however, dicotyledo- 

 nous herbaceous plants are more largely mingled with the Grammeae. 



