17 



Britton & Brown 6995, Portorico; (2) Britton & Boynton 8201, 

 Portorico; (3) Duss 47, Martinique, French W. I., and are true to 

 the isotype which I have seen in the Parisian Museum. 



Phyllanthus nummidariaefolius ( =P. tenellus Roxb. ; P. corco- 

 vadensis Muell. Arg., syn. nov. ; P. minor Fawc, syn. nov.) is 

 represented by the following collections : (1) Ball s.n., 1882, Tijuca, 

 Rio de Janeiro, Brazil; (2) Duss 2442-3557 [duplicate sheet], 

 Guadeloupe, French W. I. ; (3) Harris 12157, 12208, 12123 [three 

 sheets, including isotype of P. minor] , Jamaica. 



The best characters of identification of P. nummidariaefolius 

 from P. lathyroides and the species or forms in the latter's vicinity 

 (e.g., P. diffusus KL, well represented by: /. vS. De La Cruz 3662, 

 British Guiana, in the herbarium of the N. Y. Botanical Garden) 

 are the following: (a) Shape of the leaf. In P. nimimulariaef olius 

 the leaf, regardless of its size, is more or less gradually narrowed 

 from the center towards the extremities, being ovate to obovate. 

 In P. lathyroides and P. diffusus the leaf is essentially elliptic, 

 with the sides tending to run more or less parallel, (b) Length of 

 the pedicel. In P. nummulariaef olius the -pedicel, especially that 

 of a fruiting flower, is subcapillary but stiffly produced, always 

 manifestly elongate. In P. lathyroides and P. diffusus the pedicel is 

 much shorter. In P. niruri the pedicel is very short, so that the 

 female flower can here be described as subsessile. (c) Size of the 

 lobes of the calyx of the female flower. In P. nummulariaef olius the 

 lobes' are small, narrowly triangular-acuminate, showing like a 

 minute "star" at the tip of the pedicel. In P. lathyroides the lobes 

 are definitely large and subpetaloid. In P. diffusus and P. Niruri 

 the lobes are much smaller than in P. lathyroides and thus tend to 

 approach the size if not the shape of those of P. nummidariaefolius, 

 but the length of the pedicel is much shorter, as noticed above. 



The seed furnishes good characters of determination in Phyl- 

 lanthus, but only mature seeds can be usefully compared for critical 

 identifications and it is unfortunate that there are all too few speci- 

 mens in herbaria which have a complement of seeds fit to be used. 

 The vegetative characters listed above will be found adequate, I 

 believe, at least for provisional determinations. 



The Arnold Arboretum 

 Jamaica Plain, Mass. 



