The fatal venom which many of these Reptiles possess has so impressed the mind of men, even the scientific, that for a long time snakes were primarily divided into poisonous and non-poisonous, a classification in which the more important characters, derived from the general structure, and especially from the skull, were subordinated to the physiological. Such a system was far from reflecting natural relationships. Besides, as our knowledge progressed, drawing a distinction between poisonous and harmless snakes became more and more difficult, so many snakes previously regarded as harmless proving to be poisonous in various degrees--at least enough to paralyze the small prey on which they subsist, if not to be of serious danger to man. In the division into families, as followed in this work, the presence or absence of a poison organ is left out of consideration. Further, in this as in many other groups of the animal kingdom, external characters do not furnish trustworthy indications for higher divisions, and the definitions of the families are therefore based exclusively on osteological characters. For those who wish to name snakes with facility, the key which concludes the chapter on External Characters will, however, remedy this defect, and suffice for the identification of all the European species without any reference to their anatomy. Many attempts have been made to furnish an easy criterion for the distinction of harmless from poisonous snakes, but the characters hitherto suggested with this object can only be applied successfully to the small number of representatives in a limited area. Thus, in Southern Australia it might be stated that all snakes showing the regular nine large shields on the upper surface of the head are dangerous to man, whilst those with small shields or scales are harmless; but in most parts of Europe this criterion would have to be reversed. In some countries the shape of the pupil might be used for the purpose, in others the size of the ventral shields, or the presence or absence of a loreal shield, between the nasal and the preocular, and so on. But when we have to deal with the snakes of the whole world, about 2,000 species, of which nearly one-third are poisonous to a greater or less degree, every attempt at a definition of the two categories without regard to the dentition breaks down. Only those who have made a study of the snakes of the world can make a guess from the general appearance as to an unknown form being poisonous or not, and even they may sometimes feel embarrassed, unless the dentition be examined; the mistakes which have occasionally been made by some experienced herpetologists are proof sufficient of the fallacy of external characters for this purpose. The Ophidia are divided into nine families, the first, third, seventh, and ninth of which have representatives in Europe: