In the second place, amongst habitual criminals who are not of unsound mind, however little the inmates of prisons may have been observed with adequate ideas and experience, there is a clear indication of a class of individuals, physically or mentally abnormal, induced to crime by inborn tendencies, which are manifest from their birth, and accompanied by symptoms of extreme moral insensibility. Side by side with these, another class challenges attention, of individuals who have also been criminals from childhood, and who continue to be so, but who are in a special degree a product of physical and social environment, which has persistently driven them into the criminal life, by their abandonment before and after the first offence, and which, especially in the great towns, is very often forced upon them by the actual incitement of their parents. Amongst occasional criminals, again, a special category is created by a kind of exaggeration of the characteristics, mainly psychological, of the type itself. In the case of all occasional criminals, the crime is brought about rather by the effects of environment than by the active tendencies of the individual; but whilst in most of these individuals the deciding cause is only a circumstance affecting all alike, with a few it is an exceptional constraint of passion, a sort of psychological tempest, which drives them into crime. Thus, then, the entire body of criminals may be classed in five categories, which as early as 1880 I described as criminal madmen, born criminals, <p 25>criminals by contracted habits, occasional criminals, and criminals of passion. As already observed, criminal anthropology will not finally establish itself until it has been developed by biological, psychological, and statistical monographs on each of these categories, in such a manner as to present their anthropological characteristics with greater precision than they have hitherto attained. So far, observers continue to give us the same characteristics for a large aggregate of criminals, classifying them according to the form of their crime rather than according to their bio-social type. In Lombroso's work, for instance, or in that of Marro (and to some extent even in my work on homicide), the characteristics are stated for a total, or for legal categories of criminals, such as murderers, thieves, forgers, and so on, which include born criminals, occasional and habitual criminals, and madmen. The result is a certain measure of inconsistency, according to the predominance of one type or the other in the aggregate of criminals under observation. This also contributes to render the conclusions of criminal anthropology less evident.