PBOFESSOR BABINGTON ON RUBUS IN 1891. 287 my view, it is therefore enoneous to take permanency from seed as a decided criterion of species." — Foclce, pp. 89, 90. He also justly remarks that " it is only by means of minute descriptions that we are able to recognize with certainty the various forms of plants. Those who rely too much on single characters for the recognition of species in very short diagnoses or tabular forms will only too often find themselves in a maze of error, for there is not one single character that can be considered as absolutely permanent and reliable." — Focke, p. 91. Introductory. After much consideration I have arrived at the conclusion that Dr. Focke's arrangement is more satisfactory than that of Genevier, for it does not separate allied plants so much. Genevier seems to have wished to use an artificial arrangement, which he probably believed to be more convenient for the readers of his book, than a more natural one. Although he has to some extent succeeded, he has far from wholly done so. I have therefore chiefly followed Focke in this essay; merely deviating from him in those cases where our views do not quite agree. M. Camus, in his recently published Catalogue des plantes de France, de Suisse, et de Behjiqiie (1888), has made a bold attempt, with some success, to form what may be called aggregate species. I fear that we can only approach to the formation of such definite and natural collections of named forms at present. I have endeavoured so to arrange our forms, as far as they are yet determined, for there may probably be many more than we know at present, in as con- venient and at the same time natural a manner as is in my power. It will be seen that the present arrangement is fundamentally the same as I have always followed, although it will be new to our botanists in some few points. I do not see how to improve it. It must be always remembered that a linear arrangement is necessarily un- natural ; for the affinities of the different plants do not lie in only two, but in many directions. We must therefore not be surprised by finding plants, which are manifestly allied, placed in distinct groups, when they seem, taking all the characters into account, to be more fitly there placed than with the others to which they show a relationship. Of course this adds much to the difficulty 'of ar- ranging them upon anything approaching to a natural system ; but we are obliged to employ a linear arrangement. Gandoger, in his remarkable Flora Europaa, tom. viii., divides the genus into three, and has taken much pains to reduce the number of species by arranging under each of his species those of other authors which he combines with them severally. To this attempt I have paid much attention, but have not thought it desirable to adopt the new genera into which he divides Paibus. Unfortunately he gives no definition of these genera, nor of the species, although he points out innumerable varieties under each of the latter. As Dr. Focke remarks, there seems to be endless variation amongst brambles, and therefore endless forms which may and perhaps ought to be named and defined. It matters little whether
