17G ORIGIN OF CULTIVATED PLANTS. 



trees in the ancient plantations of the country, does not 

 .speak of it.^ 



In conclus'on, I consider it most probable that the 

 species is indigenous in Ecuador, and perhaps in the 

 neighbouring part of Peru. 



Oranges and Lemons Cltrwi, L'nn^us. 



Tlie ditierent varieties of citrons, lemons, oranges, 

 shaddocks, etc., cultivated in gardens have been the 

 subject of remarkable works by several horticultuiists, 

 among which Gallesio and Risso^ hold the lirst rank. 

 The difficulty of observing and classifying so many 

 varieties was very great. Fair results have been 

 obtained, but it must be owned that the method was 

 wrong from the beginning, since the plants from which 

 the observations were taken were all cultivated, that is 

 to say, more or less artificial, and perhaps in some cases 

 hybrids. Botanists are now more fortunate. Thanks to 

 the discoveries of travellers in British India, they are 

 able to distinguish the wild and therefore the true and 

 natural species. According to Sir Joseph Hooker,^ who 

 was himself a collector in India, the work of Brandis* is 

 the best on the Citrus of this region, and he follows it 

 in his flora. I shall do likewise in default of a mono- 

 graph of the genus, remarking also that the multitude 

 of garden varieties which have been described and 

 figured for centuries, ought to be identified as far as 

 possible with the wild species.^ 



The same species, and perhaps others also, probably 

 grow wild in Cochin-China and in China ; but this has 

 not been proved in the country itself, nor by means of 

 specimens examined by botanists. Perhaps the im- 

 portant works of Pierre, now in course of publication, will 



' Molina, FrencTi trans. 



' Gallesio, Traite da Cifrvs, in 8vo, Paris, ISll ; 771330 and Poiteau, 

 llistoire Naturelle des Oramjers, 1818, in fulio, 109 plates. 

 ' Hooker, Fl. of Brit. Ind., i. p. 515. 



* Brandis, Forest Flora, p. 50. 



* For a work of this nature, the first step would be to publish good 

 figures of wild species, showing particularly the fruir, wtiic-h is not seen 

 in herbaria. It would then be seen which forms represent 'd in tha 

 plates of Risso, Duhamel, and others, are nearest to the wikl types. 



