ANNOTATIONS AND ADDITIONS. 163 



well as of Torreya, Salisburia adiantifolia, and Cephalo- 

 taxus from among the Taxinese, recalls forcibly the obscurity 

 which still prevails in the conditions which have determined 

 the original distribution of vegetable forms, a distribution 

 which cannot be sufficiently and satisfactorily explained solely 

 by similarity or diversity of soil, thermic relations, or me- 

 teorological phenomena. I remarked long ago that the 

 Southern Hemisphere for example has many plants belonging 

 to the natural family of Rosacese, but not a single species 

 of the genus Rosa. We learn from Claude Gay that the 

 Rosa chilensis described by Meyen is only a wild variety 

 of the Rosa centifolia (Linn.), which has been for thousands 

 of years a European plant. Such wild varieties, (i. e. varie- 

 ties which have become wild) occupy large tracts of ground 

 in Chili, near Valdivia and Osorno. (Gay, Flora Chilensis, 

 p. 340.) 



In the tropical region of the Northern hemisphere we 

 also found only one single native rose, our Rosa montezumse, 

 in the Mexican highlands near Moran, at an elevation of 

 8760 (9336 Engl.) feet. It is one of the singular phe- 

 nomena in the distribution of plants, that Chili, which has 

 Palms, Pourretias, and many species of Cactus, has no 

 Agave ; although A. americana grows luxuriantly in 

 Roussillon, near Nice, near Botzen and in Istria, having 

 probably been introduced from the' New Continent since the 

 end of the 16th century, and in America itself forms a con- 

 tinuous tract of vegetation from Northern Mexico across the 

 isthmus of Panama to the Southern part of Peru. I have long 

 believed that Calceolarias were limited like Roses exclusively 

 to one side of the Equator ; of the 22 species which we 



