1 9 o PROBLEMS OF THE ENGLISH as well. Deimling made no attempt to prove his scheme in this manner. I do not know whether any of my audience have ever attempted to do so on the basis of Deimling's collations, but I do know that if they tried they failed. The thing cannot be done, for Deimling's collations are both incom- plete and, even so far as they go, often inaccurate. It does not seem to be always realised that when you are recording the readings in which a number of manuscripts differ from one selected text, to be incomplete is not merely to give information that is defective, but information that is actually false. It has, therefore, been necessary to make a fresh collation with a view to determining the relation- ship of the manuscripts, and for this purpose I have naturally sele&ed the play of Antichrist, for which we have a sixth independent text whereby to check the readings of the rest. For this play I have made a careful collation of all the manu- scripts, and though I cannot, of course, say that no variant has escaped me to do so would be to stamp myself a charlatan I think it is unlikely that I have overlooked any reading occurring in more than one manuscript, and I do claim with some confidence what is really the important thing, namely, that wherever I have recorded any variant I have recorded all the manuscripts in which it occurs. If that confidence is justified, then, and only then, does my collation afford a sound basis for argument. Two classes of variants I have excluded from purview : those in stage directions, and those which merely affect linguistic forms such as 'ye' and 'you,' 'has' and 'hath,' &c. Inspection
