But it is evident that even when a criminal cannot be classed precisely in one or the other category, and <p 46>stands between the two, this is in itself a sufficiently definite classification, especially from a sociological point of view. There is consequently no weight in the objection of those who, basing their argument on an abstract and nebulous idea of the criminal in general, and judging him merely according to the crime which has been committed, without knowing his personal characteristics and the circumstances of his environment, affirm that criminal anthropology cannot classify all who are detained and accused. In my experience, however, as a counsel and as an observer, I have never had any difficulty in classifying all persons detained or condemned for crimes and offences, by relying upon organic, and especially upon psychological symptoms. Thus, as Garofalo recently said, whilst the accepted criminal science recognises only two terms, the offence and the punishment, criminal sociology on the other hand recognises three: the crime, the criminal, and the means best calculated for social self- defence. And it may be concluded that up to this time, science, legislation, and, in a minor degree, but without any scientific method, the administration of justice, have judged and punished crime in the person of the criminal, but that hereafter it will be necessary to judge the criminal as well as the crime. After these general observations on the anthropological classes of criminals, it might seem necessary to establish their respective numerical proportions. But as there is no absolute separation between one <p 47>and another, and as the frequency of the several criminal types varies according to the crimes or offences, natural or otherwise, against persons or property, no precise account can be rendered of the criminal world as a whole. By way of approximation, however, it may be said in the first place that the classes of mad criminals and criminals of passion are the least numerous, and represent something like 5 or 10 per cent. of the total. On the other hand, we have seen that born and habitual criminals are about 40 or 50 per cent.; so that the occasional criminals would also be between 40 and 50 per cent. These are figures which naturally vary according to the different groups of crime and of criminals which come under observation, and which cannot be more accurately determined without a series of special studies in criminal anthropology, as I said when answering the objections which have been raised against the methods of this novel science.