without reference to the context; but, as we have seen, context counts for little or nothing in these early quotations,—­verbal resemblance is much more important.  The supposition of a common earlier source for both the Canonical and the Clementine text seems to me quite out of the question.  There can be little doubt that the reference to the Psalm is due to the first Evangelist himself.  Precisely up to this point he goes hand in hand with St. Mark, and the quotation is introduced in his own peculiar style and with his own peculiar formula, [Greek:  hopos plaerothae to rhaethen].

I must, however, again repeat that the surest criterion of the use of a Gospel is to be sought in the presence of phrases or turns of expression which are shown to be characteristic and distinctive of that Gospel by a comparison with the synopsis of the other Gospels.  This criterion can be abundantly applied in the case of the Clementine Homilies and St. Matthew.  I will notice a little more at length some of the instances that have been marked in the above table.  Let us first take the passage which has a parallel in Matt. v. 18 and in Luke xvi. 17.  The three versions will stand thus:—­

Matt. v. 18.

[Greek:  Amaen gar lego humin; heos an parelthae ho ouranos kai hae gae iota en ae mia keraia ou mae parelthae apo tou nomou, heos an panta genaetai.]

Clem.  Hom. iii. 51. Ep.  Pet. c. 2.

[Greek:  Ho ouranos kai hae gae pareleusontai, iota en ae mia keraia ou mae parelthae apo tou nomou] [Ep.  Pet. adds [Greek:  touto de eiraeken, hina ta panta genaetai]].

Luke xvi. 17.

[Greek:  Eukopoteron de esti, ton ouranon kai taen gaen parelthein, ae tou nomou mian keraian pesein.]

It will be seen that in the Clementines the passage is quoted twice over, and each time with the variation [Greek pareleusontai] for [Greek:  heos an parelthae].  The author of ’Supernatural Religion’ argues from this that he is quoting from another Gospel [Endnote 172:1].  No doubt the fact does tell, so far as it goes, in that direction, but it is easy to attach too much weight to it.  The phenomenon of repeated variation may be even said to be a common one in some writers.  Dr. Westcott [Endnote 172:2] has adduced examples from Chrysostom, and they would be as easy to find in Epiphanius or Clement of Alexandria, where we can have no doubt that the canonical Gospels are being quoted.  A slight and natural turn of expression such as this easily fixes itself in the memory.  The author also insists that the passage in the Gospel quoted in the Clementines ended with the word [Green:  nomou]; but I think it may be left to any impartial person to say whether the addition in the Epistle of Peter does not naturally point to a termination such as is found in the first canonical Gospel.  Our critic seems unable to free himself from the standpoint (which he represents ably enough) of the modern Englishman, or else is little familiar with the fantastic trains and connections of reasoning which are characteristic of the Clementines.