356 Lii. COMBRETACE^. [HHgera 



with its lobes and scarcely exceeding them ; stamens 5, inserted with 

 and alternate to the petals ; filaments cylindrical, rather thick, free at 

 the base, with 2 shorter whitish fleshy sub-spathulate rather turgid 

 staminodes between them ; anthers large, 2-celled, inserted on both 

 sides of the dilatedly capitate connectives, lateral, dehiscing with 

 ellipsoidal yellow ear-shaped bent-back valves ; cells opposite, bursting 

 in an elastic manner by means of a membrane ; pollen large, golden- 

 tawny, globose ; ovary inferior, 1 -celled. 1-ovuled ; style central, a 

 little longer than the stamens, funnel-shaped ; stigma thickened with 

 a flexuous edge : fruit samaroid. Abundant but sporadic, at the 

 borders of primitive forests among the mountains of Serra de Alto 

 Queta and of Quilombo ; in fl. bud beginning of Dec. 1855, fl. Jan. 

 and Feb. 185G, in young fr. beginning of Oct. 1855, fr. March 1856, and 

 fully ripe fr. Sept. 1856. No. 1753. Coll. Carp. 552 and 073. 



By some misunderstanding these specimens were placed in the Flora 

 of Tropical Africa, ii. p. 436, under Gi/rocarpus Jacquini Roxb. 



LIII. MYRTACE^. 



Africa is but little favoured with the natural occurrence of 

 Myrtacese : Welwitsch reckoned that only from 1 to 1^ per cent, 

 of the whole number of species in the world had been found 

 native in that continent. This is the more surprising, since 

 several species grow with exuberance and produce excellent fruit, 

 while the soil and climate of tropical Africa have proved to 

 be very favourable to all the introduced species. In Angola the 

 indigenous species are by no means numei-ous, but it is probable 

 that they will be found more abundant in the countries lying to 

 the east. Most of the species are small trees belonging to the 

 genus Etogenia ; many occur in the form of low shrubs in places 

 cleared by the cutting down of forest trees for the purpose of 

 cultivation and afterwards abandoned. It is worthy of mention 

 that the vast quantity of the Guava trees Avhich are met with 

 constitute occasionally dense forests, as, for instance, on many 

 islands situate in the river Quanza, and principally on those of 

 Quitage, Bumba, and Calemba ; their fruits are equal in colour 

 and fragrance to the guavas actually cultivated, but they are 

 much larger and better flavoured. It will require a careful investi- 

 gation to be made in the more remote interior of the continent in 

 order to solve the problem whether these trees, which ai-e reported 

 to be really natives of America, are, however, indigenous in 

 Africa, or are only the results of previous cultivation, and 

 consequently are due to introduction in ancient times. 



Nearly the same doubt respecting its true origin pi;esents itself 

 in relation to an arboiescent species, not apparently distinct from 

 Eugenia Jamhos L., which is met with at the banks of brooks, 

 constituting extensive thickets in and about forest ravines in Alto 

 Queta, a high mountainous range of Golungo Alto, and which 

 grows in great quantity in neglected stations remote from dwell- 

 ings, and appears to be really indigenous in that region. The 

 typical form of the species, which is the Jamboeiro of the Portu- 

 guese colonists, is not very common, and Welwitsch was unable 



