anything in the struggle for life, it must already be given in a state of working efficiency. So that, unless we invoke either the doctrine of "prophetic types" or the theory of sudden creations, I confess I do not see how we are to explain either the origin, or the development, of a reflex mechanism by means of natural selection alone. Lastly, in the third place, _even when reflex mechanisms have been fully formed_, it is often beyond the power of sober credence to believe that they now are, or ever can have been, of selective value in the struggle for existence, as I will show further on. And such cases go to fortify the preceding argument. For if not conceivably of selective value even when completely evolved, much less can they conceivably have been so through all the stages of their complex evolution back to their very origin. Therefore, supposing for the present that there are such cases of reflex action in nature, neither their origin nor their development can conceivably have been due to natural selection alone. The Lamarckian factors, however, have no reference to degrees of adaptation, any more than they have to degrees of complexity. No question of value, as selective or otherwise, can obtain in their case: neither in their case does any difficulty obtain as regards the co-adaptation of severally useless parts. Now, if all these distinctions between the Darwinian and Lamarckian principles are valid--and I cannot see any possibility of doubt upon this point--strong evidence in favour of the latter would be furnished by cases (if any occur) where structures, actions, instincts, &c., although of some adaptive value, are nevertheless plainly not of selective value. According to the ultra-Darwinian theory, no such cases ought ever to occur: according to the theory of Darwin himself, they ought frequently to occur. Therefore a good test, or criterion, as between these different theories of organic evolution is furnished by putting the simple question of fact--Can we, or can we not, show that there are cases of adaptation where the degree of adaptation is so small as to be incompatible with the supposition of its presenting a selective value? And if we put the wider question--Are there any cases where the co-adaptation of severally useless parts has been brought about, when even the resulting whole does not present a selective value?--then, of course, we impose a still more rigid test.