Outside of these conditions of the moral sense, which is no special sentiment, but an expression of the entire moral constitution of the individual, as the temperament is of his physiological constitution, other sentiments, of selfishness or even of unselfishness, are not wanting in the majority of criminals. Hence arise many illusions for superficial observers of criminal life. But these latter sentiments are either excessive, as hate, cupidity, vanity and the like, and are thus stimulants to crime, or else, as with religion, love, honour, loyalty, and so on, they cease to be forces antagonistic to crime, because they have no foundation in a normal moral sense. From this fundamental inferiority of sentiment there follows an inferiority of intelligence, which, however, does not exclude certain forms of craftiness, though it tends to inability to foresee the consequences of crime, far in excess of what is observed in the average members of the classes of society to which the several criminals belong. <p 11> Thus the psychology of the criminal is summed up in a defective resistance to criminal tendencies and temptations, due to that ill-balanced impulsiveness which characterises children and savages. II. I have long been convinced, by my study of works on criminal anthropology, but especially by direct and continuous observation from a physiological or a psychological point of view of a large number of criminals, whether mad or of normal intelligence, that the data of criminal anthropology are not entirely applicable, in their complete and essential form, to all who commit crimes. They are to be confined to a certain number, who may be called congenital, incorrigible, and habitual criminals. But apart from these there is a class of occasional criminals, who do not exhibit, or who exhibit in slighter degrees, the anatomical, physiological, and psychological characteristics which constitute the type described by Lombroso as ``the criminal man.'' Before further defining these two main classes of criminals, in their natural and descriptive characterisation, I must add a positive demonstration, which can be attested under two distinct forms--(1) by the results of anthropological observation of criminals, and (2) by statistics of relapse, and of the manifestations of crime which anthropologists have hitherto chiefly studied. As for organic anomalies, as I cannot here treat <p 12>the whole matter in detail, I will simply reproduce from my study of homicide a summary of results for a single category of these anomalies, which a methodical observation of every class of criminals will carry further and render more precise, as Lombroso has already shown (see the fourth edition of his work, 1889, p. 273).