The first is that it is incorrect to assert a mechanical regularity of crime, which from Quetelet's time has been much exaggerated. There has been a too literal insistance on his famous declaration that ``the budget of crime is an annual taxation paid with more preciseness than any other''; and that it is possible to calculate beforehand how many homicides, poisoners, and forgers we shall have, because ``crimes are generated every year in the same number, with the same punishments, in the same proportions.'' And one constantly meets with this echo of the statisticians, that ``from year to year crimes against the person vary at the <p 81>most by one in twenty-five, and those against property by one in fifty''; or, again, that there is ``a law of limitation in crime, which does not vary by more than one in ten.'' This opinion, originated by Quetelet and other statisticians after an inquiry confined to the more serious crimes, and to a very short succession of years, has already been refuted, in part by Maury and Rhenisch, and more plainly by Aberdare, Mayr, Messedaglia and Minzloff. In fact, if the level of criminality is of necessity determined by the physical and social environment, how could it remain constant in spite of the continual variations, sometimes very considerable, of this same environment? That which does remain fixed is the proportion between a given environment and the number of crimes: and this is precisely the law of criminal saturation. But the statistics of criminality will never be constant to one rule from year to year. There will be a dynamical but not a statical regularity. Thus the element of fixity in criminal sociology consists in asserting, not the fatality or predestination of human actions, including crimes, but only their necessary dependence upon their natural causes, and therewith the possibility of modifying effects by modifying the activity of these causes. And, indeed, even Quetelet himself recognised this when he said, ``If we change the social order we shall see an immediate change in the facts which have been so constantly reproduced. Statisticians will then have to consider whether the changes have been useful or injurious. These studies therefore show how <p 82>important is the mission of the legislator, and how responsible he is in his own sphere for all the phenomena of the social order.'' The second consequence of the law of criminal saturation, one of great theoretical importance, is that the penalties hitherto regarded, save for a few platonic declarations, as the best remedies for crime, are less effectual than they are supposed to be. For crimes and offences increase and diminish by a combination of other causes, which are far from being identical with the punishments lightly written out by legislators and awarded by judges. History affords us various impressive examples.