KTi. G. EEKTiiA^r o:n" myktaceje. 115 



single unilateral stamen-bearing appendage, as in several Lecy- 

 thidet'e, or into as many regular stamen-bearing appendages (or 

 claws of staminal bundles) as there are petals or calyx-lobes, as 

 in Ilelaleiica and other polyadelphous Le2)tospermeap, is in most 

 cases taken as an absolute generic character, being constant in a 

 very large number of species otherwise allied. In Astartea these 

 staminal bundles are opposite to the calyx-lobes, or alternate with 

 tlie petals ; in Melaleuca and its allies tliey are, on the contrarj^, 

 opposite to the petals — a difference more important than the actual 

 prominence or non-prominence of the disk (polyadelphy or free- 

 dom of the stamens) ; for the former distinction, or a tendency to- 

 wards it, when it exists, generally separates the natural subtribe 

 Bieekeea> from the other Leptospermea?, whilst the passage from 

 the tmiou of the stamens in phalanges or bundles to their more 

 or less conspicuous juxtaposition in clusters, and thence to the 

 regular uninterrupted ring, is sometimes gradual. This polya- 

 delphy or freedom separates Astartea very artificially from 

 Ji(Echea\ it is quite constant in LomaTcTcea^ Conothammts^ JJeati- 

 fortia, Jiegelia, and Calotliamnus^ nearly constant in separating the 

 large genus Jlelaleitca from CaUistemon^ but becom.es vague or 

 specific only in Pity matocar pus ^ Eremcea, and Eucalyptus {EiiJes- 

 onia)^ and scarcely constant in Tristania, 



In point of numbers of stamens, the general character of Myr- 

 tacea? as compared with Melastomacese is their indefiniteness ; and 

 this is perhaps quite constant in Myrtea^ and Lecytliideae, even 

 where, as in Myrrldnium^ their total number is usually below tliat 

 of twice the petals; it is also constant in most genera of Lepto- 

 sperme?e, and in Cahjtlirix^ Lliotz'kya^ and Ilomaloearpus among 

 Chama?lauciea?. In the rcmaiiiing Chama)laucie?e the number is 

 as constantly definite, twice or four times that of the petals ; whilst 

 in the very natural genera Sclioltzia and BcccJica the character 

 entirely breaks down : constant in species, it barely serves to make 

 artificial sections. 



The inflection of the stamens in the bud is a very general cha- 

 racter, not only in Myrtaccae, but in Melastomacege and other 

 allied Orders, but it is not quite constant O. Berg found them 

 erect from the first in a Brazilian shrub allied to Psidium, which 

 he not only established as a genus under the name of Feijoa, but 

 raised to the rank of a distinct tribe of Myrtacea^, believing it to 

 be also exceptional in the presence of albumen. We have re- 

 tained it as a genus, as there appear to be other minor points of 

 distinction; but the stamens erect in the bud appear occasionally 



i2 



