THE SALIX LISTS IN THE 'LONDON CATALOGUE.' 469 this modification of it, is best shown by the quotation of a paragraph on this group from the Bevision (p. 415) : — " Before proceeding to notice each of these varieties, I may say that, as the result of the study of a very large series of specimens (many of them authenti- cally named), both British and Continental, I have failed to find such a permanence of characters as will serve to definitely separate one from another. Certain examples can be, without much hesi- tation, placed satisfactorily under one or other of the varietal names. There are many, however, that cannot really be referred to one variety more than to another, and which, combining to some extent the characteristics of each, form connecting links. Of some others little more can be said than that they are modifications of S. Smithiana.^ With this confession of the failure of the arrangement to give satisfactory results, why should so keen an interpreter of willow hybrids have departed from his usual order, and inserted an aggre- gate hybrid with five varieties ? How is this any better than the old arrangement of S. amhigua Ehrh., for instance, with four varieties — an arrangement of former Catalogues now wisely dropped? Surely an attempt to differentiate the hybrids of S, viminalis and the CaprecB is not so hopeless as to justify the retention of this relic of a bygone classification.* At the same time S. stipuJaris Sm. and 8. acuminata Sm. must stand in our list as unsolved hybrids of the viminalis group, together with the three which are more easily recognised and understood, aurita x viminalis^ Caprea x viminalis, and cinerea X viminalis; for the present this arrangement, which is practically Wimmer's, can hardly be improved. 1406 lanata x herbacea (Stephania F. B. White). The Bev. E. S. Marshall's observations on the origin of S. Sadleri Boswell- Syme (Journ. Bot. 1894, 212), taken in conjunction with the accurate description of the same willow by Dr. White in the Revision (p. 422), have led me to the conclusion that 8. 8adleri and S. 8tephania are names for different forms of the same hybrid, the former nearer 8. lanata in general appearance, and the latter nearer 8. herbacea. I have yet another form in the garden from Glen Fiagh, Forfar, which is more exactly intermediate, and helps to connect these two described forms. 1407 Lapponum L. b. helvetica Vill. Forfar and Mid-Perth are the two counties from which specimens of this variety are supposed to have come, but it will be seen on reference to the Revision (p. 428) that the evidence for their being of British origin does not amount to certainty. Hence the ? after the comital number in the alternative list.
- In a parallel case, viz. the hybridization of S. purpurea with Caprece^ Dr.
White criticizes Andersson for retaining the name Pontederana, which "he makes" to "include the hybrids formed not only with S. cinerea, but with S. Caprea, S. grandifolia, and S. aurita, since he thinks that there is no sure method of separating them. As, however, he has not united these species, it is scarcely justifiable to unite the hybrids if it is at all possible to distinguish them ; and he himself has kept them separate as varieties " {Journ. Linn. Soc. xxvii. 450). Dr. White, in retaining the name S. Smithiana for the aggregate viminalis x Caprece, has done exactly what he regarded as "scarcely justifiable" in the case of Andersson with the name Pontederana.
