no independent thinker can for a moment suppose that the progress of the analysis or division of types has reached its limit. M. Jordan's work shows how far it has already been carried, and he admits that it must be carried further. It is as certain as anything in the future, that the number of book species in local Floras will go on increasing year after year, till the current of thought among naturalists has completely changed its direction.

This want of agreement among botanists as to what is to be considered a species, and what a variety, is a matter of common notoriety, as well as of the greatest importance. It goes far to prove that the forms, about which there is so much uncertainty, are of trivial importance in comparison with the general types of which they are a part. The subject may be illustrated by the comparison of any two local Floras, and for our purpose it may be well to select as examples the two British Floras, which take the most opposite views regarding species. Bentham's Handbook approaches, we conceive, as nearly as possible to the Linnean canon in its estimate of the value of species, while Babington in his Manual, taking a narrower view of the definition of a species, has a considerably greater number, though not nearly so many as some continental botanists. According to Bentham the number of British flowering plants is 1228, of which 1057 correspond exactly to the same number of species in Babington. It is only with respect to the remaining 171 species of Bentham's estimate that there is any difference of opinion. Of that number, 116 are divided by Babington into two, twenty-six into three, thirteen into four, eight into five, two into six, three into seven, one into eight, one into twelve, and one into forty. In no case does a single type in Babington represent two in Bentham. The tendency towards synthesis in the Handbook, and towards analysis in the Manual, is without exception. We may further remark that there is never, or so seldom as to be as good as never, any interlacement of forms, but that a single type of Bentham corresponds to two or more of Babington's. One or two exceptions in Ulmus and Rumex may be explained by the fact that Bentham has in these cases not gone far enough in the direction of synthesis to arrive at the true specific type.

If our space would permit of our extending this comparison to other local Floras, we should find everywhere the same want of agreement among botanists. The more the species are what is called critical, the less the chance of unanimity. In every case too it will

be found that there is a point, symbolised by the Linnean type, beyond which synthesis cannot be carried. The number of elements into which these varieties are divided varies indefinitely from the extreme sub-division advocated by Jordan, to the smallest possible number of parts, but whatever the number, it is (it may be said) invariably co-extensive with the Linnean type, the existence of which is recognised, if not as a species, at least as a group of species, or as a section or sub-division of a genus. The conclusion to which we are thus led has already been arrived at in another way by Dr. Hooker, who of all botanists at home or abroad, has given most thought to the question of the limitation of species. In his essay on the Arctic Flora in the 23rd volume of the Linnean Transactions, he tells us that every attempt to carry out his investigations regarding the distribution of Arctic plants was unsatisfactory so long as he kept before him the critical species, which are perhaps more numerous in that Flora than in any other, in consequence of the wide area over which most of the species are spread. It was only by having recourse to synthesis, by bringing together as much as possible, into one whole, all the closely allied forms and regarding them as one, that he was able to arrive at any clear views on the distribution of Arctic plants.

The considerations which we thus urge upon naturalists do not apply to systematic botany alone, but with equal force to Zoology. Till naturalists have acquired a thorough conviction that the increase of sub-division is necessarily progressive, and almost infinite, there will be no change in their mode of working. A complete change in the mode of looking at the general question of species is therefore an essential preliminary to a change of practice. For the present anything like unanimity in such matters is out of the question. It is, however, worthy of enquiry whether it be equally impossible to make some alteration in the mode of naming plants (and animals) which shall enable all schools of naturalists to compare their results more readily than they now can. At present there is an absolute want of a common medium of communication between those botanists, who take a comprehensive view of the value of a species, and those who restrict its meaning. The binomial system of Linneus was undoubtedly an enormous boon to science. The cumbrous nomenclature in previous use made it difficult to talk of plants. The phrases which took the place of a specific name varied according to the caprice of the namer, and when

alike did not necessarily denote in the book of one naturalist the same plant as in that of another. Linneus for the first time gave precision to the terminology of the science, and made progress possible. It is evident, however, that the precision thus imparted to science depends on the fixity of the meaning applied to the term species. With the gradual change in our views on this subject, and still more in consequence of their uncertainty, no one now knows in any case what meaning to attach to a name. It has lost its connection with a definite subject, and means now one thing and now another. The inconvenience of this uncertainty has been felt by all working botanists, and there can be no doubt that its removal would be a greater benefit to science than anything since the introduction of the binomial system of nomenclature.

In the case of genera the baneful effects of the gradually increasing tendency to sub-division have been well pointed out by Bentham. As the number of known species increased with the advance of science, and as each individual species became better known, it became possible to form groups lower in value than the original Linnean genera. For a time these groups were defined as subgenera; but this has been done more effectively, or, at least, more systematically in Zoology than in Botany. In the latter science the sub-division of gencra, notwithstanding the efforts of a few farseeing men, has been carried to so great an extent that a reaction has now set in. Each sub-division, of greater than specific value, being considered a genus, is thought entitled to a name, and in conformity with the principles of the binomial nomenclature, the old generic name disappears, or remains only as the name of a section or sub-order, intermediate between the genus and the order, and as such is soon forgotten by all but a few systematists.

The case is exactly the same with regard to species. Mr. Bentham has argued, and we quite agree with him, that the old genera are far more natural than the modern sub-divisions. He has further dwelt upon the importance of making the names in common use conform to the most natural sub-divisions. In like manner, if it be admitted that the old Linnean species, whether rightly or wrongly so called, be groups natural in themselves, and much more natural than the sub-species into which they are divided, it must be a matter of regret that their names should disappear entirely from our nomenclature. In the case of the fifty-three Drabas, which M. Jordan describes, all of which he regards as dismemberments of the

old Linnean D. verna, but to each of which he gives a distinct specific name, all sense of proportion, when compared with the other species of Draba, is entirely lost with the disappearance of the original name. Even if we refer them all to Erophila, as a genus or sub-genus, unless we regard that genus as monotypic we have no clue to the relative value of the specific name. So long as the section Batrachium of the genus Ranunculus consisted of one species, it was fairly comparable with R. muricatus, but now that it is sub-divided in Britain into twelve species, each of the sub-species is to appearance commensurate with the whole of R. muricatus, which in the British Flora is kept entire, whereas it is in fact only comparable with one of several forms into which that common European and Asiatic species may be analysed, and into which it is in fact divided in Oriental Floras.

To a certain extent Linneus had foreseen and provided for this difficulty by sub-dividing his species when necessary into varieties, marked by one of the letters of the Greek alphabet, and generally also with a special Latin name. In this manner the relative importance of the species and the variety was made manifest, the name of the species being most prominent. The Linnean nomenclature is still retained in theory, but it has been more or less abandoned in practice; forms which he regarded as varieties, being now very frequently regarded as species.

There have not been wanting here and there naturalists who have been sensible of the inconvenience of this state of things, and who have attempted to provide a remedy. Mr. Hewett Watson, in particular, in the course of his valuable investigations into the distribution of British plants, was of course a continual sufferer from the present lax system of nomenclature, and has recorded in the pages of his Cybele a vigorous protest against it. As is well known he has proposed to recognise three classes of species, each representing a fundamentally different idea, viz:-aggregate species, true species, and segregate species; designating by the first those species which are divisible into a certain number of types capable of recognition and definition, and presumably capable of transmitting their characters for a limited number of generations to their offspring. The third term, segregate or subspecies, he applies to the units of which the aggregate species is made up. Modifying these definitions a very little so as to regard the second or true species as equal in rank with the first, except in so far that it has not yet been resolved

into sub-species, though capable of being so, when examined with sufficient minuteness, we find these terms quite in accordance with ur views.

In the new edition of English Botany too we are glad to see at Mr. Syme has directed his attention to this subject, and like ery one who does so is fully sensible of its difficulties. He very uly observes that the real point of difference among botanists is hat some give the name of species to Hewett Watson's two first roups, namely, the aggregate species and true species, while others apply it to the true species and segregate or sub-species. In a few cases Mr. Syme has gone some way in the direction of synthesis, though not so far as we should be glad to see him do. For instance, he divides Bentham's Ranunculus aquatilis, not into twelve, but into seven species, six of them true species, that is to say having no subspecies, while one only, to which he restricts the name R. aquatilis, is an aggregate species, and is divided into sub-species. If it were at all likely that this mode of grouping the forms would meet with general acceptance it would be satisfactory to adopt it, but it will assuredly find little favour in the eyes of others, and will at the best be only one of many ways of grouping together a multitude of forms. Mr. Syme may depend upon it that when once he has turned his steps into the direction of synthesis, he will find no fixed resting place out of the pale of the Linnean canon. Indeed, we think that in assigning the rank of species to so many forms of the section Batrachium, he goes against his own excellent definition of a sub-species, as a form characterised by slight but constant differences, which are transmitted by descent for an indefinite period. In the case of Ranunculus circinatus, he tells us that it can be distinguished by the practised eye without hesitation from all the other forms. This is an expression familiar to us from the pen of M. Jordan, from whose example we at once learn that it may be used to prove anything. Surely it is a fallacy to assume that the eye can only detect differences, and that it may not also be applied to the discovery of resemblances. The same eye can be applied to the most minute micrometrical measurement, and to the study of the widest panorama. Pity that it should be restricted in its use to one or the other only.

In Mr. Syme's book, however, we are glad to see the introduction of a mode of naming plants, which will perhaps be available as a remedy for the present state of chaos. It is used, indeed, vaguely