184 SYSTEMATIC BOTANY. 



allowance must be made for these ; and this gives a laxity to the family 

 character which is puzzling- to the beginner. For example, the family 

 Ranunculaceae is very natural ; but we find in its character a certain range 

 of difference allowed for in the sepals, petals, pistils, and fruit ; the insertion 

 of all these, however, and that of the stamens, is fixed, and so is the cha- 

 racter of the seed. Similar conditions occur in most other families. The 

 decision as to what family a genus is to be referred to is made according 

 to the principle of majorities : whichever it agrees with in most of its cha- 

 racters (say, even three out of five), to that family it belongs. Great diffi- 

 culty, however, exists in certain cases from a vast series of genera running 

 into one another by almost imperceptible gradations, and this in different 

 directions. A considerable number of these agreeing closely are associated 

 into a family ; another similar group forms another family, and so on ; and 

 then, in the course of time, sundry intermediate genera present themselves, 

 which connect the established families, and which it is difficult to place by 

 the usual choice in either one or the other, the characters being balanced. 

 Thus the Natural family Loganiaceae is connected by " aberrant " genera 

 with Rubiaceae, Gentianaceae, Scrophulariaceae, and other families which 

 are truly natural, but which in this way come to be separated by somewhat 

 indefinite boundary-lines. The fact is, that the Vegetable Kingdom is a 

 whole, the families having seldom, a distinct isolated existence, except in 

 the minds of botanists. It may be presumed that they are all variations 

 from one or a few original stocks, and thus have numerous intermediate 

 or connecting links ; and we must regard them as analogous to countries 

 on the globe, which are parcelled out under distinct names, but most often 

 adjoin and run into one another, being only separated by an arbitrary 

 boundary-line. Some, indeed, lie off from the rest, like islands, the inter- 

 vening links being extinct ; but these are the exceptions. Such excep- 

 tions are found among the families which were established by the older 

 botanists, in which the essential agreements are accompanied by a striking 

 character of external habit, as in the Grasses, the Umbelliferae, the Com- 

 positae, the Leguminosae, the Coniferae, the Palms, &c. Such remarkable 

 peculiarities as these families possess mostly prevent them being broken up 

 Into smaller groups, as has occurred to many of the earlier orders of large 

 extent ; and most botanists prefer to distribute these genera into suborders 

 rather than discard the characteristic general name. Examples of these 

 are found especially in the Leguniinosae, RosaceaB, and Compositae. 



The Families or Orders are for the most part the same, in all 

 essential respects, in all existing "Natural Systems." A conside- 

 rable diversity presents itself in the modes in which different 

 authors have grouped these into Classes or Alliances. These, how- 

 ever, are still Natural groups, as are also those of still higher gene- 

 rality indicated in the chapter on General Morphology. But all 

 writers on Systematic Botany have found it requisite to group the 

 Orders or Classes of Flowering Plants into sections of somewhat 

 less generality than Dicotyledons and Monocotyledons, as these 

 respectively include series of families so extensive as to be incon- 

 venient in practice if left undivided. The members of these series, 

 however, are so intimately connected together by their natural 



