BY E. C. ANDREWS. 567 



ters, from the generalised type of the fertile tropics, and that they 

 have been evolved in response to some particular, physical revolu- 

 tion, such as a change to a sandy, a clay, a dry, or a cold environ- 

 ment. 



Similarly, the fruits of the earlier types of Myrtaceae were 

 fleshy, and the capsular-types, in this instance, appear to be a 

 response to less fertile and less sheltered habitats. The depauper- 

 ate fruit of the Chamaelaucieae here suggests a much more recent 

 response again to conditions more generally severe. 



The habit of the individual is again instructive. The typical 

 Myrteae either may be the largest of forest-trees or they may be 

 elegant shrubs, with full habit and glossy foliage. The typical 

 Chamaelaucieae, and many of the Leptospermeae, are very depaup- 

 erate in form, and not at all suggestive of the stately and hand- 

 some Myrteae and Metrosidereae. Such depauperate forms, more- 

 over, abound in the subarid and barren, sandy areas of Australia. 

 From this consideration, also, such forms appear to be much more 

 recent developments. Furthermore, almost without exception, such 

 depauperate and apparently aberrant types have a limited range, 

 being endemic to Australia, frequently Western Australia. This 

 restricted geographical range, moreover, suggests a much more 

 recent origin than that of the widely-spread Myrteae. 



The distinctive characters, however, of the majority of Western 

 Australian from Eastern Australian species of Myrtaceae, in view 

 of the fact that the areas considered are mainly sandy in nature, 

 and the centre of Australia arid, strongly suggest that the Central 

 Australian desert was a hindrance rather than an impetus to Myr- 

 taceous differentiation ; and that the peculiar sandy soil of Eastern 

 and Western Australia, formed one of the most potent influences 

 in the production of new species, and that, whereas, probably in 

 earlier Tertiary time, the Eastern and Western species commingled, 

 they have more recently developed along divergent lines, since the 

 interposition of an arid barrier of less sandy soil, in Central Aus- 

 tralia. 



From these considerations, it would appear that the Myrteae are 

 much the oldest branch of the family, that Euleptospermeae and 



