146 MR. G. BEXTITAir ON MTKTACE^. 



r 



some minor ones, might have constituted quite as good a mono- 

 typic genus as any of those separated from Metrosiderois. 2. Lysi- 

 CAE,PTJS,P. MuelL, another single Australian species, the Tristania 

 angustifolia^ Hook., referred to the latter genus on accovmt of the 

 stamens being slightly interrupted opposite to the sepals, and thus 

 showing an approach to the section Neriophjllum of Tristania^ 

 but really much nearer to Metrosideros, differing in habit and in 

 the polygamous flowers, the outer stamens of the hermaphrodite 



L 



ones bearing moreover abortive reniform anthers. 3. Cloezia, 

 Brongn. and Gris, to which are probably referable Mooria and 

 Ballardia of Montrouzier, and comprising several New- Caledonian 

 species, only differing from Metrosideros in the shorter stamens, 

 fewer ovules, and more basal placentation, 4. Tepitaxia, G-riseb., 

 a single Chilian species, removed from Metrosideros on account of 

 its habit, inflorescence, and geographical station, with the few 

 ovules and basal placentation of the preceding genera. 5. SpekHO- 

 LEPis, Brongn. and Grris, comprising tw^o New Caledonian species 

 unknown to me. Besides the4-merous not 5-merous flowers, the 

 seeds are described as solitary by abortion, bearing near the 

 hilum a kind of involucre or ring of six membranous scales — a most 

 remarkable anomaly, unless it should prove that these scales are 

 abortive ovules adhering to the perfect seed. 6. Nai^ia, Miq-j 

 comprising one Malayan and one Australian species, which we 

 have restored to Metrosideros as a section, as being more generally 



known under that genus, and scarcely differing, except in the 

 broad flat, not linear-cuneate seed. 



Metkosidekos, Banks, was long the repository for all capsular 

 Myrtacese with numerous free ej sorted stamens, and as now li- 

 mited, besides the two Nani(S and a rather anomalous species from 

 South Africa, contains a considerable number of forms ranging 

 over the Pacific, from New Zealand and New Caledonia to the 

 Sandwich Islands, some of them very variable in foliage and in- 

 dumentum, and exceedingly difficult to distribute into well-defined 

 species. All are distinguished from the neighbouring genera, 

 chiefly by the nun\erous ovules covering the whole surface of toe 

 peltate or laterally adnate placenta. Some of the species difi^r 

 considerably from each other in the ovary, wholly inferior or more 

 or less oi tlie upper portion free, sometimes after flowering re- 

 maining adnate only by the broad base. 



XANTHOSTKMoy, F. Muell., was first proposed for a ^ori • 

 Australian tree or shrub, remarkable for its long, erect, somew 

 rigid stamens, with peculiar anthers, and differing also iro 



^ — 



