A PRELIMINARY REVISION OF PHILIPPINE MYRTACEiE. 339 



often unequal, oblique or folded with the radicle included. Very few Philippine 

 species are yet known in fruit, and such results as have been obtained do not 

 warrant generalizations. 



Many of the species have wide limits of variation, while others may seem to 

 be separated by rather trivial characters. In the cases where a good series of 

 specimens made it possible to form a definite opinion, the differences are often 

 much greater than can be expressed in any key. Often the real basis of separation 

 lies in the nature of the venation, and defies brief description, verbal or numerical. 

 In doubtful cases, there is no way to obtain a satisfactory determination except 

 by direct comparison with authentic material. Even then, it is often very 

 difficult, as two series may be distinct but separated in such a way that it is 

 conceivable that additional material might close the gaps. 



In general the comparison of Philippine species with those of other countries 

 has proved disappointing. Close alliances occur in the case of nearly all species, 

 but only in a few do they seem close enough to warrant the inclusion of our plants 

 with the limits of the outside species. A disproportionally high percentage of 

 identical species falls within the limits of Syzygiwn. 



The number of endemic species is consequently large, and a considerable 

 proportion are here described as new. Several other species are represented in 

 this herbarium by insufficient material to warrant description, and the most 

 recent collections have added some of the most striking novelties. Even so, other 

 species are known only from Cuming's collections, though these are now few. With 

 several important regions of the islands not or very imperfectly worked, many 

 additions are still to be expected. 



Nearly all our Eugenias are trees, yielding hard, close-grained wood. They 

 are widely used for construction and internal work, but only in a few places are 

 they cut on a large scale. Yet, single species even, may form a measurable 

 proportion of the forest of a given locality, 3 and one of the purposes of the present 

 paper is to lay a basis for their separation and commercial classification. Nat- 

 urally, the local names are badly tangled, the same species having several names 

 in the same or different localities, while names like Malaruhat {Mala-duhat, 

 false Eugenia jambolana), are applied to many species. 



KEY TO THE PHILIPPINE SPECIES OF EUGENIA." 



1. Individual flowers subtended by one pair of bracteoles, these conspicuous or 

 inconspicuous. 

 2. Calyx-lobes conspicuous, petals free, flowers usually large. 

 3. Inflorescence and flowers more or less villose or silky. 

 4. Leaf -bases acute. 



5. Leaves at least 3.5 cm wide I . E . aherniana 



5. Leaves not attaining a width of 2.5 cm. 

 6. Leaves at least 2.5 cm long. 



7. Leaves membranaceous 2. E. loheri 



7. Leaves chartaceous or subcoriaceous 3. E. pasacaensis 



3 See Merritt Bur. Philip. For. Bur. 8 (1909) 38 and elsewhere under 

 Malaruhat. 



8 In this key and in the descriptions the calyx is understood as beginning at 

 the articulation of the flower and that upon which it is borne. Very often, the 

 lower part of the calyx is greatly contracted and seems to form a pedicel, but this 

 pseudostalk is really a part of the flower, and often in whole or part conspicuously 

 swollen in fruit. Frequently, flowers are here stated to be sessile, when apparently 

 stalked. True pedicels are also present in a majority of species. 



