DISCUSSION. 427 



MS.). These two forms are allied by certain structural 

 characters, such as, the staminal filaments shortly but distinctly 

 united at the base into a tube, a character, possibly, of sufficient 

 importance to justify the establishment of a Section l^ibulosce.) 

 The matured fruits of both species open out very broadly at the 

 orifice and are similar in general outline. There are certain 

 other characters, however, by which they may be easily 

 sepai'ated. The length of the filaments, the thickness of the 

 rim of the fruits and its stiff upright habit readily distinguish aS'. 

 speciosus, which is common in West Australia. The short 

 filaments, thin-rimmed fruits, and its tall, slender, and pendulous 

 habit as readily distinguish C. vwiinalis, common in North- 

 Eastern New South Wales, and throughout Queensland. C. 

 rigidus R.Br., and C. linearis DC, are both sufficiently distinct 

 to be regarded as good species, especially as the seedlings 

 develop the characteristics of their respective parents; but these 

 are both found in the Port Jackson District. 



Dr. J. B. Cleland. — The avifauna of Australia seems to be well- 

 fitted, in some ways, for the study of the relationships of species. 

 The typically Australian family of the Honey-eaters( J/eZi/?^rt^/c/fe) 

 affords examples of what seem to be two independent ways in 

 which new species may be evolved. Two species of Meliomis, 

 M. novce-hollandice and M. sericea, are common in the neighbour- 

 hood of Sydney. They closely resemble each other in general, 

 but the former has a white iris, when adult; and the latter, a 

 dark brown iris, and a larger patch of white feathers near the 

 ears. Their habits are almost identical, but there is a slight 

 difl'erence in the notes of the two species. The geographical 

 range of AI. novce-hollandice, however, is more extensive than 

 that of M. sericea. In West Australia, two similarly related 

 species are found, M. longirostris corresponding to 31. novce- 

 hollandice, and M. mysticalis to M. sericea. In a young specimen 

 of M. novce-hollandice obtained recently in the neighbourhood of 

 Sydney, the irides were found to be of the tint of cafe au lait 

 (the cotFee-colour predominating), instead of enamel-white. On 

 the universally admitted rule that, in its development, each 



