Families, to be natural divisions, should stand in the same relationship to genera as species do, but from the opposite side, whatever the subdivisions are into which they may be separated, for the sake of convenience, and as descending grades whereby to arrive with greater facility at their genera, just as the species of the latter are also sometimes grouped, that they may be reached with greater ease. These subdivisions of families have no analogy with the varieties which species occasionally throw off, although they may be as irregular in their occurrence; that is to say, in the association of a group of families arranged in their series of most proximate affinities, the first may present subdivisions, others, in irregular occurrence, may not require them,—just as in the species of a genus, arranged also in the series of their closest resemblances, one will present a stringent adherence to the specific type, or all may do so, or all or some may have a tendency to vary. Groupings of species are, however, of a less natural character usually than are those of families, and generally are artificial, being capriciously made to break down long genera, that the required species may be more readily arrived at. The characters which group families differ _inter se_. Thus in the Order _Hymenoptera_, the family of the bees is essentially framed upon their most distinguishing peculiarity—the tongue,—which in other families becomes of secondary importance. In some the neuration of the wings, their mode of folding, the form of the eyes, conjunctively with other peculiarities of general structure, etc. etc., which point to the differences in the economy that accompany all these, have successively the same prominent position which the trophi take in the family of the bees. I have already recently alluded to the relations of affinity and analogy, and it is desirable that some notion of the meaning and bearing of these terms should be given, as, in the majority of modern works on natural history, use is frequently made of them. On carefully surveying any class or order of creatures, the mind speedily becomes impressed by observing certain similitudes out of the direct line of continuous connection, and therefore remote from the strongest connecting links of positive relationship in the methodical series. Induced thence to inspect them more closely, we presently ascertain that what we at first conceived might be an error in their collocation, arises from very strong resemblances in certain particular features, but which are less important than those which directly unite them, and may not be permitted to interrupt the order established. It is, however, equally evident that they indicate relations which may not be neglected.