Such, then, is the ultimate conclusion to which our inquiry leads us. And if it be asked why Stephen should confer an earldom on Geoffrey at this particular time, the reply is at hand in the condition of affairs, which had now become sufficiently critical for Geoffrey to begin the game he had made up his mind to play. For Stephen could not with prudence refuse his demand for an earldom.[139] The first corollary of this conclusion is that "the second type" of Stephen's great seal (which is that appended to this charter) must have been already in use in the year 1140, that is to say, before his fall in 1141. Mr. Birch, who, I need hardly say, is the recognized authority on the subject, has devoted one of his learned essays on the Great Seals of the Kings of England to those of Stephen.[140] He has appended to it photographs of the two types in use under this sovereign, and has given the text of nineteen original sealed charters, which he has divided into two classes according to the types of their seals. The conclusion at which he arrived as the result of this classification was that the existence of "two distinctly variant types" is proved (all traces of a third, if it ever existed, being now lost), one of which represents the earlier, and the other the later, portion of the reign.[141] To the former belong nine, and to the latter ten of the charters which he quotes in his paper. The only point on which a question can arise is the date at which the earlier was replaced by the later type. Mr. Birch is of opinion that— "the consideration of the second seal tends to indicate the alteration of the type subsequent to his liberation from the hands of the Empress, and it is most natural to suppose that this alteration is owing to the destruction or loss of his seal consequent to his own capture and incarceration" (p. 15).