64 MEMOIRS OF THE NEW YORK BOTANICAL GARDEN [VOL. 10 



further apparent that the line between Marlierea and Myrcia is not always easy 

 to draw. 



In certain species of Marlierea and Calyptranthes the branchlets are two- 

 winged or two-keeled; in these species the genera may be readily distinguished 

 even in sterile specimens because the orientation of the wing' differs with respect 

 to the position of the leaves (see key, below). Except in these species, however, 

 the genera are distinguishable by calyx characters only. The distinction is not 

 difficult if the calyptra dehisces cleanly, without further splitting of the calyx 

 or hypanthium; the undoubted species of Calyptranthes behave in this way. In 

 a " typical" species of Marlierea, likewise, the closed or nearly closed buds open 

 by irregular longitudinal fractures resulting in four or five nearly equal lobes. 

 In various other species, however, most of which have been assigned at one time 

 or another to Marlierea, there are irregularities in the opening of the calyx. 

 These irregularities are of two principal kinds : 



(1) The closed bud opens irregularly, but one of the divisions is somewhat 

 or much larger than the others, simulating a calyptra. This calyptra-like lobe 

 may remain attached at one side, as in Calyptranthes. The other, smaller lobes 

 may persist or may be deciduous soon after anthesis. The prolonged rim of the 

 hypanthium may remain entire, or (usually) become irregularly lobulate, as in 

 M. sub orbicularis. In M. salticola the calyx is definitely calyptrate, without ac- 

 companying small lobes, but the species is assigned to Marlierea because after 

 the abscission of the calyptra the margin of the hypanthium is recurved and 

 becomes irregularly lacerate-lobed. 



(2) The calyx is incompletely closed in bud, with short lobes which separate 

 as the flower opens. The difficulty here is to find a suitable distinction, on this 

 character alone, between Marlierea and Myrcia. When the calyx-lobes are ini- 

 tially very small and the buds are almost entirely closed, the opening of the 

 flower initiates a major transformation in w 7 hich the splitting of the calyx and 

 hypanthium extends nearly the whole length of the bud, and the hypanthium 

 is often much distorted and flattened. These species can be assigned without 

 question to Marlierea. In certain other species, how r ever, e.g. in M. bipennis, 

 M. guildingiana, and two newdy described species, the calyx-lobes are relatively 

 large in proportion to the size of the bud ; the splits between the lobes are not 

 extensive, so that the lobes are not w T idely separated ; the tube of the hypanthium 

 does not become explanate but remains cylindrical in flower and in fruit ; the 

 fruit, in fact, crowned by the ring of the hypanthium-tube and the persistent 

 rounded calyx-lobes, resembles the fruit of various species of Myrcia. 



I do not think, therefore, that in doubtful cases a distinction should be made 

 between Myrcia and Marlierea on the basis of calyx-characters alone. A proper 

 decision may depend upon study and correlation of the combined characters of 

 calyx, inflorescence and pubescence. 



The distinction between Marlierea and Calyptranthes is even more difficult, 

 and often impossible unless opening buds are available for study. It is there- 

 fore difficult to construct a practical key for the identification of the species of 

 these two genera, and in the present paper the two genera have been keyed out 

 together, with minimal dependence placed upon the technical characters which 

 separate the groups. 



From the taxonomic standpoint a botanist with tendencies toward "lumping" 

 might be tempted to combine Marlierea and Calyptranthes. As a practical mat- 



