54 GENERAL REMARKS ON THE 



abundance, and exists, tlioiigli much more sparingly, within 

 the tropic. 



RestiaccEc are ahnost equally numerous at the Cape of 

 Good Hope as in the principal parallel of New Holland. 

 One species only of the order has been observed in New 

 Zealand, and hitherto none in South America. 



CYPERACEt^. In Terra Australis this is a very ex- 

 tensive order, consisting already of more than .200 species. 

 It contains, however, fewer peculiarities in structure than 

 several other orders that are much less numerous. Its 

 maxim nm ajjpears to be in the principal parallel ; but the 

 species observed solely vvithin the tropic exceed one third 

 of the whole number. Cyperacese, in many respects, are 

 nearly related to Restiaceec, and when furnished with a 

 true perianthium are distinguishable from the monosper- 

 mous genera of that order, solely by the different position 

 of the embryo in the seed.. But in the greater part of the 

 order the perianthium is either entirely wanting or merely 

 setaceous. Puirena, Lepidosperma and Orcobolus, all of 

 them natives of New Holland, are almost the only genera 

 in which it is found of nearly the usual appearance. 



What I have formerly termed perianthium in Carex, 

 Diplacrum, and Schoenus nemorum, ought, perhaps, rather 

 to be considered as hiternal bracteae, analogous to those of 

 Lepyrodia, of Irideas, and, perhaps, to the upper valve of 

 the inner envelope of grasses. 



I have formerly remarked that the Perianthium of Hypoe- 

 580] lyptum consists merely of the squamae of a spicula, 

 similar to that of Kyllinga, but reduced to two valves. 



GPA]\IINE^. This order comprehends, at least, one 

 fourth of the whole of ]\Ionocotyledones, and in Terra 

 Australis, where upvrards of 200 species have already been 

 observed, it bears the same proportion to that primary 

 division. 



I have formerly, in arranging the Australian genera of 

 Gramiriece, endeavoured to explain wliat I conceived to be 

 the natural subdivision of nearly the whole order into tvro 



