484 NATURAL HISTORY OP PLANTS. 



one, containing nearly three hundred and twenty species all belong- 

 ing to the -warm regions of the world. One Melia, it is true, 

 extends from North China to the northern banks of the Mediter- 

 ranean. To the south, the genus Epicharis is also represented by a 

 New Zealand species, the Cape of Good Hope again possesses a 

 species of Ekehergia and two species of Turrcca. But the majority 

 of the genera are tropical. The genus Trlchllia, existing in the 

 warmest parts of Africa and South America, does not advance in 

 North America beyond the warm parts of Mexico. The Lansiiims in 

 tropical Asia stop at the Himalayas. Except Jfclia, met with iu 

 both worlds, and, as we have seen, even in the temperate parts, all 

 the genera of the 3Ieliecv. series are peculiar to the tropical regions 

 of the old world. The genus Turrcea, the most widely-extended 

 of these genera, is found at the Cape, Madagascar, and the Mascariue 

 Islands, in tropical Asia and Oceania, and in tropical eastern and 

 western Africa. Australia possesses the genera Synoum, Ilearnia, 

 and Owenia. Tropical America presents as special genera Guarea, 

 Cabralea, Swietenia, and Elutheria ; she divides with the old world 

 the genera Ifelia, Trichilia, and Cedrela^ to say nothing of Carapa, 

 consisting of coast plants which, like the mangrove, are met with on 

 all the tropical shores of both worlds. We may estimate the species 

 peculiar to the new world at almost a hundred (nearly a tliird of the 

 family). 



The characters up to the present constant in this family are : 

 alternation of the leaves, absence of stipules, regularity of the 

 ilowers and primitive direction of the descendent ovules, with the 

 micropyle turned upwards and outwards.^ Other characteristics, 

 which, although not constant, are at least very rarely wanting, 

 being : the hermaphrodism of the flowers, the independence (between 

 themselves) of the parts of the corolla, the definite number of the 

 pieces of the androceum (isostemonous or diplostcmouovis). The 

 characters which, on the contrary, vary the most, arc : the union or 

 independence of the staminal filaments, the number of the ovules 

 and seeds, the presence or absence of a wing on the seeds and of 

 albumen inside them ; then the simple or compoimd character of the 

 leaves, the relations of the sepals to each other and the mode of 



' Xot to sjicak of the union of the ovarian part of the carpels to whieh we shall refer presently. 



I 



