1 88 Oranges and Lemons of India. 



Sometimes there are many more whorls, as in Risso's 

 " Bergamottier mellarose a fleur double," shown in pi. 

 241, figs, a and b, of the accompanying Atlas. 



It has been objected that, if the hesperidium con- 

 sisted of two whorls, the law of phyllotaxis would re- 

 quire that their segments should be alternate instead 

 of opposite. To this objection I would reply (a) 

 That the rind whorl has been so altered by selection, 

 that in most cases it has become one uniform casing, 

 without any trace whatever of the carpel sutures, and 

 where the sutures of the rind carpels are still visible, 

 the fruit has been so altered by increase in the number 

 of its parts, that it is impossible to say now what the 

 order of the two sets of carpels may originally have 

 been (if there were two' separate whorls). Le Maout 

 and Decaisne, "System of Botany" (Hooker), at p. 

 3 1 8, give a diagram of the orange flower, showing 5 

 sepals, 5 petals, 20 stamens (in five groups of i, 3, 4, 

 5 and 7 stamens each), and with 8 pulp carpels ; and 

 at p. 319 is given a transverse section of an orange, 

 with 9 pulp carpels* Unless we come across a wild 

 citrus with 5 pulp carpels and 5 distinct rind carpels 

 (if such a thing ever existed), we cannot say what the 

 phyllotaxis of the citrus fruit may have been, viz.: 

 whether its parts were alternate or opposite, through 

 torsion, or suppression of an intermediate whorl. The 

 original wild citrus may be extinct, and may have es- 

 caped our observation for ever. I have seen only one 

 orange the kumquat which had 5 pulp carpels (pre- 

 sumably the original number), and the rind had no sign 

 of carpel sutures. 



(b) There are many instances in which the law of 

 phyllotaxis does not hold good, such as in Dipsacus, 



* Vide all the sections of all the citrus throughout the accom- 

 panying Atlas. 



