ONE:But these periods of time will not be lengthened in exact proportion to the atrocity of crimes, since the probability of a crime is in inverse ratio to its atrocity. It will, then, be necessary to shorten the period for inquiry and to increase that of prescription; which[159] may appear to contradict what I said before, namely, that it is possible to inflict equal penalties on unequal crimes, by counting as a penalty that period of imprisonment or of prescription which precedes the verdict. To explain to the reader my idea: I distinguish two kinds of crimes��the first, atrocious crimes, beginning with homicide and including all the excessive forms of wickedness; the second comprising less considerable crimes. This distinction is founded in human nature. Personal security is a natural right, the security of property a social one. The number of motives which impel men to violate their natural affections is far smaller than those which impel them, by their natural longing for happiness, to violate a right which they do not find written in their hearts but only in the conventions of society. The very great difference between the probability of these two kinds of crime respectively makes it necessary that they should be ruled by different principles. In cases of the more atrocious crimes, because they are more uncommon, the time for inquiry ought to be so much the less as the probability of the innocence of the accused is greater; and the time of prescription ought to be longer, as on an ultimate definite sentence of guilt or innocence depends the destruction of the hope of impunity, the harm of which is proportioned to the atrocity of the crime. But in cases of lesser criminality, where the presumption in favour of a man��s[160] innocence is less, the time for inquiry should be longer; and as the harm of impunity is less, the time of prescription should be shorter. But such a division of crimes ought, indeed, not to be admitted, if the danger of impunity decreased exactly in proportion to the greater probability of the crime. One should remember that an accused man, whose guilt or innocence is uncertain, may, though acquitted for lack of proofs, be subjected for the same crime to a fresh imprisonment and inquiry, in the event of fresh legal proofs rising up against him, so long as the time of prescription accorded by the laws has not been past. Such at least is the compromise that I think best fitted to preserve both the liberty and the security of the subject, it being only too easy so to favour the one at the expense of the other, that these two blessings, the inalienable and equal patrimony of every citizen, are left unprotected and undefended, the one from declared or veiled despotism, the other from the turbulence of civil anarchy.That the punishments of long custody by which we[103] now defend our lives and properties are out of all proportion to the real needs of social existence is indicated by such a fact as that no increase of crime used to attend the periodical release of prisoners which was for long, if it is not still, customary in Russia at the beginning of each reign. Neither in India, when on the Queen��s assumption of the title of Empress, a pardon was granted to about one-tenth of the prison population, did any increase of crime ensue, as, according to all criminal reasoning, it should have done, if the safety of society depends on the custody of the criminal class.[60] In Sweden a low rate of crime seems to be a direct consequence of a low scale of punishment. Of those condemned to travaux forc��s, which may vary from a period of two months to a period for life, 64 per cent. are condemned for one year, and only 3 per cent. are condemned for seven years;[61] whilst sentences to the latter period in England form between 50 and 60 per cent. of the sentences to penal servitude.