Still, on the whole, I incline to think that though there is not conclusive proof that Justin used a lost Gospel besides the present Canonical Gospels, it is the more probable hypothesis of the two that he did.  The explanations given above seem to me reasonable and possible; they are enough, I think, to remove the necessity for assuming a lost document, but perhaps not quite enough to destroy the greater probability.  This conclusion, we shall find, will be confirmed when we pass from considering the substance of Justin’s Gospel to its form.

But now if we ask ourselves what was this hypothetical lost document, all we can say is, I believe, that the suggestions hitherto offered are insufficient.  The Gospels according to the Hebrews or according to Peter and the Protevangelium of James have been most in favour.  The Gospel according to the Hebrews in the form in which it was used by the Nazarenes contained the fire upon Jordan, and as used by the Ebionites it had also the voice, ’This day have I begotten Thee.’  Credner [Endnote 110:1], and after him Hilgenfeld [Endnote 110:2], thought that the Gospel according to Peter was used.  But we know next to nothing about this Gospel, except that it was nearly related to the Gospel according to the Hebrews, that it made the ‘brethren of the Lord’ sons of Joseph by a former wife, that it was found by Serapion in the churches of his diocese, Rhossus in Cilicia, that its use was at first permitted but afterwards forbidden, as it was found to favour Docetism, and that its contents were in the main orthodox though in some respects perverted [Endnote 110:3].  Obviously these facts and the name (which falls in with the theory—­itself also somewhat unsubstantial—­that Justin’s Gospel must have a ‘Petrine’ character) are quite insufficient to build upon.  The Protevangelium of James, which it is thought might have been used in an earlier form than that which has come down to us, contains the legend of the cave, and has apparently a similar view to the Gospel last mentioned as to the perpetual virginity of Mary.  The kindred Evangelium Thomae has the ‘ploughs and yokes.’  And there are some similarities of language between the Protevangelium and Justin’s Gospel, which will come under review later [Endnote 110:4].

It does not, however, appear to have been noticed that these Gospels satisfy most imperfectly the conditions of the problem.  We know that the Gospel according to the Hebrews in its Nazarene form omitted the whole section Matt. i. 18—­ii. 23, containing the conception, the nativity, the visit of the Magi, and the flight into Egypt, all of which were found in Justin’s Gospel; while in its Ebionite form it left out the first two chapters altogether.  There is not a tittle of evidence to show that the Gospel according to Peter was any more complete; in proportion as it resembled the Gospel according to the Hebrews the presumption is that it was not.  And the Protevangelium of James makes no mention of Arabia, while it expressly says that the star appeared ’in the East’ (instead of ‘in the heaven’ as Justin); it also omits, and rather seems to exclude, the flight into Egypt.