Morphology of the Citrus. 199 



ance of cultivated places, with their primitive huts and 

 inhabitants.* Who can tell how often a tract of 

 country has had its forest cleared, either by fire or 

 other means, for purposes of cultivation and villaging, 

 and then become reforested by new forms, to be at 

 some future time recleared and reforested. It is evi- 

 dent that, in addition to natural causes, man may have 

 had a hand in extinguishing many species. I see no 

 good reason for insisting on there being a wild form 

 of the pummelo somewhere, especially as the existence 

 of the cultivated present large forms can, I think, be 

 accounted for without much difficulty, by fusion of two 

 ovaries, as in plates 223 and 224. See also plate 83, fig. b. 



When we say that the carpels forming the fruit, 

 and the floral apparatus, are modifications of leaves, 

 we do not mean of ti\z present leaves, but of some 

 former expansion of the bark, which eventually was 

 modified into the present leaves, the present flowers, 

 and the present fruit. For all we know to the con- 

 trary, the original expansion of the bark in the citrus 

 may have been of the nature of the modern green 

 rind of the fruit. 



In the changura or digitate citrons and lemons, 

 we probably have a glimpse of how the citrus fruit 

 began to evolve. In the " limonier digite " of Risso, 

 fig. a, pi. 178, we appear to have, as I said, a single 

 whorl of carpels, united at their base, and free at the 

 distal ends, withoiit any inner whorl of pulp carpels. 

 The whole fruit is rind. In the digitate citrons of 

 India, fig. a, pis. 139 and 140, we have a further step 

 a double fruit, with an outer whorl of rind carpels, 

 and an inner one, also of rind carpels. The double 

 digitate citrus is a step in advance of the single one, 



* It is enough to read the history of Babylon and Assyria to 

 realize what can be effected by this means. 



