with each successive generation.  Now, the former of these may be said to be well established, and we recognise it as a law of life that all plants and animals propagate their own kind.  But the latter has, as yet, been hardly examined at all.  Each new generation shows special slight variations.  But what causes these variations? and what determines what they shall be?  In Darwin’s investigations these questions are not touched.  The variations are treated as if they were quite indefinite in number and in nature.  He concerns himself only with the effect of these variations after they have appeared.  Some have the effect of giving the plant or animal an advantage in the struggle of life; some give no such advantage; some are hurtful.  And hence follows the permanent preservation or speedy destruction of the plants and animals themselves.  But we are bound to look not only to their effects but to their causes, if the theory is to be completed.  And then we cannot fail to see that these variations in the progeny cannot be due to something in the progenitors, or otherwise the variations would be all alike, which they certainly are not.  They must, therefore, be due to external circumstances.  These slight variations are produced by the action of the surroundings, by the food, by the temperature, by the various accidents of life in the progenitors.  Now, when we see this, we see also how gravely it modifies the conclusions which we have to draw concerning the ancestry of any species now existing.  Let us take, for instance, the great order of vertebrate animals.  At first sight the Darwinian theory seems to indicate that all these animals are descended from one pair or one individual, and that their unity of construction is due to that fact; but if we go back in thought to the time at which the special peculiarities were introduced which really constituted the order and separated it from other animals, we see that it is by no means clear that it originated with one pair or with one individual, and that, on the contrary, the probabilities are the other way.  Although the separation of this order from the rest must have taken place very early, it cannot well have taken place until millions of animals had already come into existence.  The prodigality of nature in multiplying animal life is fully acknowledged by Darwin, and that prodigality is apparently greatest in the lowest and most formless type of animal.  There being, then, these many millions of living creatures in existence, the external surroundings introduce into them many variations, and among these the special variations to which the vertebrate type is due.  It is quite clear that wherever the external surroundings were the same or nearly the same, the variations introduced would be the same or nearly the same.  Now, it is far more probable that external surroundings should be the same or nearly the same in many places than that each spot should be absolutely unlike every other spot in these particulars.  The beginnings of