An inference in which there are two or three steps may be often quite as strong as that in which there is only one, and probabilities may mount up to a high degree of what is called moral or practical certainty.  I cannot but think that many of those which have been already obtained are of this character.  I cannot but regard it as morally or practically certain that Marcion used our third Gospel; as morally or practically certain that all four Gospels were used in the Clementine Homilies; as morally or practically certain that the existence of three at least out of our four Gospels is implied in the writings of Justin; as probable in a lower degree that the four were used by Basilides; as not really disputable (apart from the presumption afforded by earlier writers) that they were widely used in the interval which separates the writings of Justin from those of Irenaeus.

All of these seem to me to be tolerably clear propositions.  But outside these there seems to be a considerable amount of convergent evidence, the separate items of which are less convincing, but which yet derive a certain force from the mere fact that they are convergent.  In the Apostolic Fathers, for example, there are instances of various kinds, some stronger and some weaker; but the important point to notice is that they confirm each other.  Every new case adds to the total weight of the evidence, and helps to determine the bearing of those which seem ambiguous.

It cannot be too much borne in mind that the evidence with which we have been dealing is cumulative; and as in all other cases of cumulative evidence the subtraction of any single item is of less importance than the addition of a new one.  Supposing it to be shown that some of the allusions which are thought to be taken from our Gospels were merely accidental coincidences of language, this would not materially affect the part of the evidence which could not be so explained.  Supposing even that some of these allusions could be definitely referred to an apocryphal source, the possibility would be somewhat, but not so very much, increased that other instances which bear resemblance to our Gospels were also in their origin apocryphal.  But on the other hand, if a single instance of the use of a canonical Gospel really holds good, it is proof of the existence of that Gospel, and every new instance renders the conclusion more probable, and makes it more and more difficult to account for the phenomena in any other way.

The author of ‘Supernatural Religion’ seems to have overlooked this.  He does not seem to have considered the mutual support which the different instances taken together lend to each other.  He summons them up one by one, and if any sort of possibility can be shown of accounting for them in any other way than by the use of our Gospels he dismisses them altogether.  He makes no allowance for any residual weight they may have.  He does not ask which is the more probable hypothesis.  If the authentication of a document