Undoubtedly, if the operation is stated in its most general form; if it is to be supposed that a certain quantity, or magnitude, was divided into many more parts than the whole contained; and that, after the subtraction of several thousands of such parts, the magnitude of the remainder amounted to more than the original magnitude, there does seem to be an _à priori_ difficulty about accepting the proposition, seeing that it appears to be contradictory of the senses which we attach to the words "whole" and "parts" respectively. But this difficulty is removed if we reflect that we are not, in this case, dealing with magnitude in the abstract, or with "whole" and "parts" in their mathematical sense, but with concrete things, many of which are known to possess the power of growing, or increasing in magnitude. They thus furnish us with a conception of growth which we may, in imagination, apply to loaves and fishes; just as we may, in imagination, apply the idea of wings to the idea of a man. It must be admitted that a number of sheep might be fed on a pasture, and yet there might be more grass on the pasture, when the sheep left it, than there was at first. We may generalise this and other such facts into a perfectly definite conception of the increase of food in excess of consumption; which thus becomes a possibility, the limitations of which are to be discovered only by experience. Therefore, if it is asserted that cooked food has been made to grow in excess of rapid consumption, that statement cannot logically be rejected as an _à priori_ impossibility, however improbable experience of the capabilities of cooked food may justify us in holding it to be. On the strength of this undeniable improbability, however, we not only have a right to demand, but are morally bound to require, strong evidence in its favour before we even take it into serious consideration. But what is the evidence in this case? It is merely that of those three books,[48] which also concur in testifying to the truth of the monstrous legend of the herd of swine. In these three books, there are five accounts of a "miraculous feeding," which fall into two groups. Three of the stories, obviously derived from some common source, state that five loaves and two fishes sufficed to feed five thousand persons, and that twelve baskets of fragments remained over. In the two others, also obviously derived from a common source, distinct from the preceding, seven loaves and a few small fishes are distributed to four thousand persons, and seven baskets of fragments are left.