DISTRIBUTION OF THE MELIACE. 235 
have, independently of my own researches, and after careful scrutiny, been also admitted 
by the highest authorities on the subject. Consequently I do not hesitate to consider 
them as being thoroughly natural. 
To return to the figures contained in the Table, it is evident that each of the horizontal 
series formed by them furnishes us with a complete statistical representation of the geo- 
graphical distribution of one genus ; at the same time a continuous thick, dark line drawn 
round the extremities of all such horizontal series enables us to understand clearly the 
relation which exists between that geographical distribution and the morphological 
affinities of the various genera. 
A first result becomes thus manifest. We are at once struck by the fact that the 
greatest number by far of the genera of Meliacez are to be found in the central part of the 
Table, whilst оп the sides there are only а few сепегіс forms presenting but a slight 
mutual affinity. In other words, we may say that the American and East-Polynesian 
genera, rare and very distinct, are morphologically connected by a series of numerous 
intermediate forms, all concentrated iu the Asiatie regions or in their neighbouring 
archipelagoes. 16 is, for instance, a positive fact, that out of the 35 genera composing the 
whole order, no less than 26 are represented within the limits of the western coast of 
India on one side, and of the eastern shore of Australia on the other. 
We also observe that, with the exception of Melia and Cedrela, these various genera 
are very much localized, the figures corresponding with each of them forming only short 
horizontal series in the synoptic Table. With respect to the genus Мейа, its exceptionally 
large area is more apparent than real, being due, for the most part, to the wide spread- 
ing of a single species, namely, Мейа Azedarach, which is not to be considered 
indigenous out of Tropical Africa and India. I have consequently left all the other 
stations of that species outside the continuous line including the genus, which is thus 
reduced to a very moderate area. The genus Cedrela, on the other hand, extends in 
reality over a vast surface ; but it must be observed, at the same time, that there exists 
a striking morphological difference between the American and Indian or Australian 
species, a fact which could not, however, be made apparent in the Table. 
Another general conclusion to be derived from the synoptic Table is the great analogy 
Which exists between the Polynesian and the Asiatic Мейасее. In fact, we see that all 
the Polynesian species, excepting Vavea Amicorum, belong to what may be termed the 
Indian type of Meliacess—that is to say, to genera which are represented in either one 
or the other of the Indian peninsulas. : 
With regard to Australia, the case is somewhat different. Besides may — be- 
longing to the Indian type, that continent contains also two genera ( Owenia and Lansium) 
which belong exclusively to it; and moreover ten out of the eleven species composing 
the genus Flindersia are hitherto known only in Australia. Although the whole tribe 
of Meliez, so abundant in Tropical Africa, is entirely missing in South America, there 
_ 655, nevertheless, a striking analogy between the types of those two distant con- 
_ tinents. Not only are the three genera Trichilia, Сагара, and even Guarea represented 
in Africa as well as in South America, but, moreover, Carapa procera is actually common 
to the west coast of Africa and the Guianas. 
