364 ROBINSON. 



must be left in an unsatisfactory condition. E. subrotundifolia is undoubtedly 

 allied, but quite distinct. E. mindanacnsis seems separable in tbe herbarium by 

 its smaller flowers, but is probably closer to what is here named E. javanica. 



Regarding the remaining two, there seems to be no tangible method of separat- 

 ing them in the herbarium except by at least nearly mature fruit. When living 

 trees at that stage are examined, there is no doubt possible as to their distinctness. 

 The calyx-lobes of the "calubcob" have retained their size at flowering, are not 

 thickened, and are widespreading or reflexed ; the disk between them being nearly 

 flat: the other species, here locally called "macupa," has the calyx-lobes very 

 fleshy, incurved and nearly concealing the disk. The trees are different in habit, 

 and the two can be compared side by side in the Manila Botanical Gardens. 



In attempting to correlate these with the species of other countries, there is 

 considerable difficulty. E. javanica .and E. subglauca Koord. & Val. are constantly 

 separable by the leaves, those of the former having the odor of cinnamon, those 

 of the latter entirely odorless. This test is of no assistance here. Leaves taken 

 from the two trees above mentioned were at once compared, and both when 

 crushed, had the same odor, which did not suggest that of cinnamon or of cloves 

 to any of several persons present. To me, the odor most strongly suggests that 

 of similarly crushed leaves of Myrica carolincnsis of the Atlantic coast of North 

 America; tlie fruit, which rarely contains seeds, has a faintly coniferous taste. 



"Macupa" seems not to exist in the Philippine! apart from cultivation, "calub- 

 cob" is widely distributed, usually if not always along streams, and is in all 

 probability indigenous. All attempts to satisfactorily identify it with exotic 

 species having failed, it is here described as new. 



It has been the custom to reduce E. jambos Blanco to E. malacccnsis Linn., 

 but it much more likely belongs here. I have not been able to get any record of 

 E. malacccnsis anywhere in the neighborhood of Manila. 



34. Eugenia calubcob sp. nov. 



E. montana Naves in PI, Filip. ed. 3 (1877) pi. I'io, dubie Blanco 1. c. (1837) 

 416, non alioruni. 



E. javanica Auct. Philip., pro majore parte, non Lam. I. c. 



Indorescentiis saepissime terminalibus, sed etiam in ramis subtus foliis 

 suffultis, floribus sessilibus vel pedicellatis, spuric racemose vel subcorym- 

 bose dispositis : calyeis lobis 4, rotundatis, sub f ructu nee incrassatis nee 

 ineurvatis; petalis 4, liberis: foliis chartaceis, lanceolatis ad ovalibus, basi 

 cordatis vel obtusis, apice breviter obtuse acuminatis; venis utrinque 

 10-15. 



Inflorescences on any part of the branch from below the leaves to the 

 apex, 6-18 cm long, including the flowers, the individual flowers sessile 

 or with pedicels up to 5 mm long, usually in threes at the apices of the 

 blanches of the cyme, or often solitary on its lower branches, then falsely 

 longer-pedicelled : calyx-tube turbinate, about 1 cm in length to the bases 

 of the lobes and there of about 1 cm diameter; calyx-lobes 4, 3.5-5 mm 

 long, 4-7 mm wide at the base, the apex rounded; petals 4, free, white, 

 orbicular-ovate, 6-8 mm in diameter; disk thick; stamens numerous, the 

 filaments 10-15 mm long, the anthers about 0.7 mm long; style 2.5-3.5 

 cm long; ovary 2-celled, each cell with several ovules: fruit edible, 2-3 

 cm in diameter, the calyx-lobes withering-persisting, erect, spreading, or 

 recurved, not becoming fleshy, nor strongly incurved, the disk flat or 

 nearly so, 1-celled, not more than one perfect seed found. 



