THE SALTX LISTS IN THE * LONDON CATALOGUE.' 463 of the former, and the two must go together. It afterwards came under my notice that Wimmer had remaikel the similarity, saying (Sal. Etw. p. 35) the two were so Hke in their inflorescence that they could not be separated, adding that in the fertile soil of West Bothnia S. Lapponum formed osier-like shoots. If S. Lappoinnn had stood next to -S'. viminalis in the list, changing places with S. laiiata, there would have been little objection to this arrangement ; but since iS'. viminalis has so many hybrids with the Capreae, which are mostly conveniently arranged under this species (see Explanations of the London Catalogue, ed. 9), this group fits in best in a linear arrangement just before the Capreae, as in former editions of the London Catalogue ; the " Smithiana hybrids forming a succession of connecting links, which is hardly broken by the insertion of S. Lapponum; for, as Wimmer observes [I.e.), this species undoubtedly shows the transition from S. viminalis to the cinerea group. S. purpurea and its hybrids are placed by Andersson at the end of the list, in pursuance of his division of the Salices into the three tribes, Pleiandrce, Diandrce, and Synandrm, The union of the two filaments, in whole or in part, is the main character of the Synandr(B, which include two groups, Iiicana and PurpureiB. The objection to the arrangement by which these two groups terminate the list is that a sudden jump is made from a series of dwarf alpine willows to some lowland species that have no apparent connection with them. This objection is easily removed. The Synandrce only differ from the DiandrcB by the connation of the filaments. Place the Synandra at the beginning instead of the end of the Diandrm ; let the Incanm (not represented in Britain) follow and not precede the Purpurea, and they form a natural passage from the Purpurea to the Viminales, which S. incana Schrank greatly resembles in foliage. The S. purpurea hybrids also fall into a natural position, and lead on to the Viminales and Capreae, to a union with which they most of them owe their existence. Then the very distinct willow, S. reticulata L., which Kerner at one time thought worthy of generic distinction, forms (with its immediate allies from Lap- land and N. America) a fitting termination to the list. Andersson's classification is open also to some other objections (see Dr. White's criticisms for instance, Journ. Linn. Soc. xxvii. 337), which, however, are in great part due to the inherent diffi- culties of the genus ; but it is on the whole the best arrangement published, subject to some such changes as I have suggested above and carried out in my list. My farther remarks will follow the order of the list embodied in the Catalogue, for the convenience of quoting the Catalogue numbers. 1395 triandra Linn. All three hybrids need a query as to their parentage. S. suhdola F. B. W. is assigned to triandra x alha with much hesitation in the Revision, and it is frankly admitted that the only bush known may be a form of S. viridis Fr. The name subdola, too, can hardly stand for S. alba x triandra; Wimmer declares S. undulata Ehrh. to be a hybrid of S. triandra and S. alba, though he had fully considered and rejected the triandra x viminalis theory (Sal, Eur, 145) ; and the only objection to Wimmer's theory
