I have, however, published an essay on international statistics of relapsed criminals, from which I drew the following general conclusion: that even in prison statistics, which often give higher totals of relapsed cases than are given by judicial statistics, because they are more personal, and therefore less uncertain, we never obtain the full number of relapses, though the totals given vary from country to country, from district to district, and from prison to prison. It <p 14>would be impossible to state accurately what proportion the numbers given bear to the actual number; but I am justified in saying, from all the materials which I have collected and compared in the aforesaid essay, that the number of relapses in Europe is generally between 50 and 60 per cent., and certainly rather above than below this limit. Whilst the Italian statistics, for instance, give 14 per cent. of relapses amongst prisoners sentenced to penal servitude, I found by experience 37 per cent; out of 346 who admitted to me that they had relapsed; and, amongst those who had been sentenced to simple imprisonment, I found 60 per cent. out of 363, in place of the 33 per cent. recorded in the prison statistics. The difference may be due to the particular conditions of the prisons which I visited; but in any case it establishes the inadequacy of the official figures dealing with relapse. After this statement of a general fact, which proves, as Lombroso and Espinas said, that ``the relapsed criminal is the rule rather than the exception,'' we can proceed to set down the special proportions of relapse for each particular crime, so as to obtain an indication of the forms of crime which are most frequently resorted to by habitual criminals. For Italy I have found that the highest percentages of relapse are afforded by persons convicted of theft and petty larceny, forgery, rape, manslaughter, conspiracy, and, at the correctional courts, vagrancy and mendicity. The lowest percentages are amongst those convicted of assault and bodily harm, murders, and infanticide. <p 15> For France, where legal statistics are remarkably adapted for the most minute inquiry, I have drawn up the following table of statistics from the lists of persons convicted at the assize courts and correctional tribunals, taking an average of the years 1877-81, which is not sensibly affected by the results of succeeding years.