The authors treated of in this chapter, and many others whom I have had to put under other headings as belonging to some special school, all have this point in common, that they try to find the causes of crime by means of statistics. The first question that arises is this: do criminal statistics give a real and complete picture of criminality? The answer is categorically, No. To give only a few reasons, there are a great many crimes, naturally insignificant, which remain unknown even to the person injured; there are others of which justice never takes cognizance, because the injured party has filed no complaint, either because he has pardoned the offender, or fears the trouble of a criminal trial, etc. In general, not all the cases known to justice are included in the criminal statistics but only those in which sentence is pronounced. An exception is found in Mayr’s book (§ VI of this chapter), in which all the crimes known to justice appear. In the first place the public officer dismisses many cases because of their insignificance (Germany is an exception—there the officer must prosecute); and in the second place there are a number of crimes whose authors remain unknown. Finally, only a part of those arraigned are convicted. Criminal statistics, therefore, cover only a part of the crime that exists. [83] The enemies of statistics have drawn from the preceding fact the conclusion that criminal statistics are useless for the study of the etiology of crime. This is an absolutely false conclusion, as false as it would be to claim that doctors cannot find the cause of a disease, because besides the thousands of cases known to them, there are at least as many that remain unknown. The only question that arises is the following: is the number known sufficient? is there a sufficiently large body of facts in hand for inductive studies? As far as criminality is concerned the answer can be perfectly positive. The number of offenses that do not appear in the criminal statistics is certainly large; they are, however, chiefly insignificant misdemeanors, such as insults, trifling assaults, petty thefts, etc. These would have little value in determining the etiology of crime, even if the statistics included them all. Serious crimes, on the other hand, in the majority of cases, without any doubt appear in the statistics. The great reason why criminal statistics suffice for etiological investigations is that the ratio of known crime to unknown remains relatively constant. The proofs of this are to be found in the criminal and judicial statistics themselves. The ratio between cases dismissed and those prosecuted, between convictions and acquittals, etc., etc., remains practically the same from one year to another. Further, every statistician of any experience is convinced that the law of averages rules more absolutely than any despot.