Ce ope 

1891.] | Zoology. 837 
Known for some years past by a single species referred to one of the 
most widely distributed forms in Africa, Chromis niloticus, the Chro- 
mididæ are to-day represented in Madagascar by nine series belonging 
to four genera, the affinities of which are as follows: 
The Paretroplus are distinguished from the Hemichromis, which 
belong to Africa and to Asia Minor, by a greater number of anal spines. 
The Paretroplus belong to an African type. 
If the Paratilapia have certain affinities with the Hemichromis, they 
are still more closely allied to the Acara, from which they are distin- 
guished only by the indentations of the bony projections of the external 
branchial arch ; but the Acara belong to tropical South America. 
The Paracara also have close affinities with the Acara. 
As to the Ptychochromis, Steindacher has shown that they are 
separated from the Chromis by the presence of a lemelliform, com- 
pressed projection from the upper part of the first branchial arch. 
This character is found in the genus Geophagus, from South America. 
The genera of the Chromidide peculiar to South America have 
ctenoid scales, while those of Africa and Western Asia have cycloid. 
Among the Chromididg of Madagascar, the Paretroplus, we may say, 
belong to the African type, as they have cycloid scales. All the other 
Chromididze of Madagascar have ctenoid scales. Thus the Chromidide 
of Madagascar are more closely related to the fresh-water species of 
tropical South America than to the African species. 
From a study of the herpetological and ichthyological fauna of the 
fresh waters of Madagascar, it appears that this island, Southern 
Africa, and South America formed in a pre-Tertiary epoch, parts of the 
same continent, which had a fauna of the same origin and charac- 
ter. While South America and Africa have received since the Pliocene 
epoch invasions of animals of another creation, Madagascar has had, 
on the contrary, no connection since that epoch with any other land, 
and presents to-day the same fresh-water fish fauna as it did in the 
Middle Tertiary, without any addition save that of the Carassius 
recently introduced. The division between Madagascar and Southern 
Africa must have happened before the commencement of the Miocene, 
for neither Cyprinide or Characinide are found in Madagascar. Com- 
_ munication must have lasted for a much longer time between South 
Americaand the South of Africa, which, toward the Pliocene epoch, had 
received an influx of Characins; whilst the Cyprins, introduced into 
Africa from the European continent, or rather from the South Asiatic, 
had hardly penetrated South America. (Dr. H. E. Sauvage, Bull. Soc 
Zool. de France, 1891, p. 190.) 
