12G MU. G. BENTHIM OS MTRTACEiE. 



more or less altered in size, shape, or colour j but wlien tlie flowers 

 are so crowded as to form a head or spike, the floral leaves are 

 usually reduced to bracts, either small and concealed by the 

 enlarged bud when ready to open, or larger and imbricate in the 

 bud, and falling off as the flov/cr expands. In some cases the 

 lower bracts or floral leaves of the spike, together with a few 

 empty ones immediately below the spike, are enlarged, imbricate^ 

 dry, and scale-like, or coloured and petal-like, forming an invo- 

 lucre under or enclosing the spike. Occasionally a few of the 

 upper floral leaves are coloured and enlarged into a crest crown- 

 ing the spike. And in the great majority of those Myrtacese which 

 have a capitate or spicate inflorescence, the axis, after the flowering 

 is over, or even at an earlier period, continues to grow beyond the 

 spike, producing, instead of floral leaves, stem-leaves like those 

 below it. This explains the numerous casea where the flower- 

 heads are all strictly terminal, whilst the fruits form clusters 

 surroundino; the base of the vear's shoots. 



In these simple inflorescences the flowers are constantly closely 

 sessile in Melaleuca and several allied genera, more or less pedi- 

 cellate in Sceckea and others ; but the same character is variable 

 in Leptos2)€7^mum, Knnzea, &c., where, indeed, it often happens m 

 one species that some flowers, especially females, are closely 

 sessile, Avhilst others, especially males, are pedicellate. 



The racemose inflorescence of American Eugotice must also be 

 regarded as simple. The axillary raceme Is a reduced branch, 

 and often grows out, like the spike of Euleptospermcse, into a 

 normal leafy branch, — a circumstance upon which Berg's genus 

 StenocaJyx was chiefly founded, but which forms but a very vague 

 generic character. The racemes of Lecythideae are more deter- 

 minate ; and although their ramification be the same in principle 

 as that of the leaf-branches, there is this difl*erence, which appears 

 constant, that the rachis never grows out into a leafy brancli. 



In the truly compound inflorescences of Myrtacere, buds are 

 produced in the axils of the bracteoles, developing each a single 

 flower or a several-flowered branch, producing thus the determi- 

 nate centrifugal cyme. This inflorescence occurs in Myrtea*, 

 JVIetrosidereae, and Ba^ckeese, but not in other tribes or subtribes, 

 except rarely in Thryptomene^ otherwise nearly related to JjcvcTcea. 

 AVithin the tribes the compound inflorescence is often charac- 

 teristic of large groups, and assists in the distinction of genera 

 (as, for example, in separating many-flowered JTyr^/ from American 

 Sugenioi), but it is rarely absolute throughout a genus. 



