We should thus have a self-acting test, not only of the length of restraint required for social safety, but also of that apprenticeship to labour which many convicts need; while there would be supplied a means of rectifying sundry failures and excesses of our present system. The plan would practically amount to an extension of trial by jury. At present, the State calls in certain of a prisoner’s fellow-citizens to decide whether he is guilty or not guilty: the judge, under guidance of the penal laws, being left to decide what punishment he deserves, if guilty. Under the arrangement we have described, the judge’s decision would admit of modification by a jury of the convict’s neighbours. And this natural jury, while it would be best fitted by previous knowledge of the man to form an opinion, would be rendered cautious by the sense of grave responsibility; inasmuch as {186} any one of its number who gave a conditional release, would do so at his own peril. And now mark that all the evidence forthcoming to prove the safety and advantages of the “intermediate system,” proves, still more conclusively, the safety and advantages of this system which we would substitute for it. What we have described, is nothing more than an intermediate system reduced to a natural instead of an artificial form—carried out with natural checks instead of artificial checks. If, as Captain Crofton has experimentally shown, it is safe to give a prisoner conditional liberation, on the strength of good conduct during a certain period of prison-discipline; it is evidently safer to let his conditional liberation depend not alone on good conduct while under the eyes of his jailors, but also on the character he had earned during his previous life. If it is safe to act on the judgments of officials whose experience of a convict’s behaviour is comparatively limited, and who do not suffer penalties when their judgments are mistaken; then, manifestly, it is safer (when such officials can show no reason to the contrary), to act on the additional judgment of one who has not only had better opportunities of knowing the convict, but who will be a serious loser if his judgment proves erroneous. Further, that surveillance over each conditionally-liberated prisoner, which the “intermediate system” exercises, would be still better exercised when, instead of going to a strange master in a strange district, the prisoner went to some master in his own district; and, under such circumstances, it would be easier to get information respecting his after-career. There is every reason to think that this method would be workable. If, on the recommendation of the officers, Captain Crofton’s prisoners obtain employers “who have on many occasions returned for others, in consequence of the good conduct of those at first engaged;” still better would be the action of the system when, instead of the employers having “every {187} facility placed at their disposal for