138 POLYGAL.i: AMERICANiE. 



abundantly distributed through every portion of the Continent, from 

 Canada to Patagonia. Tlie herbaria at my disposal, besides those 

 at Kew and the British Museum, have been those belonging to the 

 Berlin, Vienna, Munich, and Brussels Collections, and that of 

 M. Warming. Dr. J. Miiller of Geneva has also kindly identified 

 for me some of the species described in DeCandolle's ' Prodromus.' 

 In these herbaria I have found materials for the description 

 of fourteen new species, and of several others hitherto imperfectly 

 known or described. Species named in Watson's ' Bibliographical 

 Index' or in Martius's ' Flora Brasiliensis ' are only referred to 

 when some fresh fact has come to hglit respecting them. In the 

 case of all others, either a full diagnosis is given, or reference is 

 made to a sufficient one already published. The sign * prefixed to 

 the name of a species signifies that it is a Brazilian species not 

 named in the former, the sign f that it is a United States species 

 not named in the latter publication. The greater number of the 

 new species are the result of Balansa's visit to Paraguay in 

 1874-77 ; and as this collection has been very recently distributed, 

 I have thought it might be useful to append at the end of the paper 

 a complete nomination of the species obtained hi that journey, as 

 far as I have been able to determine them. If w^e add to the 

 35 species named in Watson's ' Index,' and the 86 in the ' Flora 

 Brasiliensis' (one species only, P. panicidata, L., being common to 

 the two*), the 34 additional ones referred to in this paper, we get 

 154 as the total number of species of Polygala at present known as 

 natives of the American Continent. This does not include five 

 species found in the West Indian Islands but not growing on the 

 Continent, viz., P. hraclnjptera, Gris., P. spathulata, Gris., P. squami- 

 folia, Wr., P. saginoides, Gris., and P. erioptera, DC. With the 

 exception of P. tenuis, DC, of Brazil, which appears to be identical 

 with P. paludosa, St. Hil., of Tropical Africa, and P. erioptpra, DC, 

 w^hich has probably been introduced into the West Indies from 

 India, there is no species of Polygala common to the Eastern and 

 Western hemispheres. 



The subdivision of the genus is that adopted by me in my mono- 

 graph of the Brazilian species in AJartius's ' Flora Brasiliensis ; ' 

 the older subdivisions, especially that of DeCandolle's 'Prodromus,' 

 being altogether useless. It is not pretended that the six divisions 

 have all equal value, or that they will all maintain their place 

 when a complete monograph of the genus comes to be written ; the 

 5th and 6th especially are not satisfactory. But in the meantime 

 it is the best for my present purpose. 



Sectio a. — Carina triloba, lobus medius integer, nee cristatus 

 nee fimbriatus ; sepala exteriora discreta ; semina eximie strophio- 

 lata vel rarius estriophiolata. Frutices vel sufiVutices, fioribus 

 magnis et nonnunquam speciosis. 



This section belongs especially to Tropical America ; most of 

 the species are good-sized shrubs ; and, in the case of the Brazilian 



* Unless, as I. suggest below, P. grandijiora, Walt, is identical with 

 P. hebeclada, DC, and P. angiistifolia, H.B. K. with P.Jlahellata, Shuttl. 



