AUSTRALIAN HONEY-ANTS. 263 



view we take, we are bound to assume that the genus could only have 

 reached its present distribution at a time when land connections existed 

 between New Guinea, Australia and New Caledonia, if I am right 

 in maintaining that winged females do not occur in the genus. 



There are, however, still other considerations that point to the 

 origin of the genus Leptomyrmex in Australia and the subsequent 

 migration of some of its species to Papua, viz. the development of the 

 honey-storing habit in the different species. I have found repletes 

 in all the Australian species except unicolor. L. pallens of New Cale- 

 donia is known to have repletes, and the same is probably true of 

 7iiger and fragilis, although nothing is known concerning the haljits 

 of these two Papuan forms. I regard L. unicolor as the most primitive 

 species of the genus, because its larva is least modified, because the 

 adult has not acquired the honey-storing habit and because it is con- 

 fined to the moist coastal "scrub." Nearly all the other forms live 

 in much dryer situations and have evidently developed repletes as a 

 means of tiding over the dry season. We know that this habit has 

 been independently developed in several other Australian ant genera. 

 It has long been known in Melophorus species and in a Camponotus 

 (C. inflatus) . I have observed it in several species of Pheidolc and in 

 an undescribed Pheidologeton from Queensland, a genus not hitherto 

 recorded from Australia. In the species of these two genera the 

 soldiers have become repletes, with a very noticeable though not 

 spherical distension of the gaster. Now the Australian botanists have 

 concluded from the differences between the juvenile and adult foliage 

 of the Eucalypti and phyllodineous Acacias that the Australian cli- 

 mate was originally much more humid than it is at the present time. 

 The honey-ants point to the same conclusion, and in Leptomyrmex we 

 still find one species, unicolor which preserves the ancient habits of the 

 ^enus and lives only in the coastal rain-forests, while all the other 

 forms have become modified for storing liquid food during protracted 

 periods of heat and drought. 



Like most other Dolichoderinte the species of Leptomyrmex are poor 

 in plastic characters. As Emery has shown, they are most easily dis- 

 tinguished by the shape of the head, though some of them present 

 useful characters in the shape of the thorax, petiole, gaster and tibije. 

 Sculpture and pilosity are very monotonous throughout the genus. 

 According to my observations, there is very little variation in color 

 among the members of a colony, but several and perhaps all of the 

 species exhibit one or more color varieties. These sometimes form a 

 series from pale to dark forms which may be repeated in another spe- 



