The problem, therefore, is to devise some means by which these benefits of a searching cross-examination may be retained and yet the abuse got rid of. The only feasible way of meeting the difficulty yet proposed is that of drawing up a series of rules or general directions as to evidence which shall not attempt to prescribe formal limits for cross-examination, but shall lay down in explicit words the general principles which should govern a judge in such cases. These rules would practically be a definition of the "discretion" he is now supposed to exercise. They would, for example, direct him not to allow an examination into matters so remote in time from the case in hand that they can have no bearing on the credibility of the witness; not to allow questions to be put which are plainly malicious and asked for the purpose of irritating the witness; and not to allow any examination into transactions which, though they may have a bearing on the character of a witness, have none on his credibility, _e.g._, an inquiry, in a murder case, of a witness in good standing, as to domestic difficulties with a deceased wife. It is not easy to lay down beforehand any rules by which we can discriminate the kind of evidence as to transactions involving moral character which ought not to affect credibility, but every one can easily imagine instances of such evidence. General directions of the kind we have just suggested are no more than a formal enunciation of the manner in which the "discretion" of a good judge would be and is exercised. They do not change the law, but they remind judges of what they may forget, and they may be appealed to by a persecuted witness with far more certainty than judicial "discretion." In the Indian Code, which is probably the best body of law that the legal reform movement begun by Bentham in the last century has yet produced, rules of this kind have been laid down, and we believe have been found to work with success. "THE DEBTOR CLASS"