Gradual , Change , and by Guppy's Method of Differentiation. Go 7 
elaborate so deadly a poison as aconite, when a simpler one would do ; and 
on p. 194 he points out that change of climate does not change the chemistry 
of a plant, so that there is no opening for Natural Selection in a change of 
conditions. 
Coe (2, Ch. n) gives a very good series of extracts from Darwin, 
Wallace, and others, showing how they contradict themselves, e. g. in the 
£ Origin of Species *, pp. 65, 146, Natural Selection is said to be always acting, 
while on pp. 85, 169, it only acts at long intervals and under certain favour¬ 
able circumstances. On p. 25 he points out that it only comes into operation 
in case of adverse changes ; it is not wanted if the change is favourable. 
Also the changes that bring it into operation must not be too rapid, or the 
organisms would perish, nor too mild, or they would not involve a question 
of life and death. Just the right amount is wanted. Natural Selection 
must wait for favourable variations to turn up, and they may not turn up 
in time. Geometrical increase of a species does not occur. As the numbers 
keep roughly constant, though four give say eight, these do not give six¬ 
teen ; the number is again reduced to four every time; the eight do not 
survive to propagate. 
The great bulk of the research carried on was based upon Natural 
Selection, inasmuch as it seemed to give a satisfactory explanation of the 
facts of adaptation. The idea of adaptation was pushed to absurd extremes, 
and adaptations were found—as indeed was necessary if Natural Selection, 
which was essentially a theory of adaptation, were to hold its place as an 
active factor in evolution—in almost everything that was conceivable with 
the greatest stretches of imagination. In spite of desperate effort, however, 
in which one might say that the imagination was stretched beyond the 
limits of perfect elasticity, no one ever succeeded in finding adaptation in the 
enormous bulk of the characters that divide one species or genus from 
another, whether in plants or in animals. And not only so, but a more 
thoughtful analysis of the cases of adaptation actually described showed 
that nearly all of them involved correlated adaptation. For example, the 
possession of tendrils as climbing organs was always accompanied by 
flexibility of the stem, which was no longer able to stand by itself. It was 
therefore necessary to suppose that selection, picking out the very doubtfully 
advantageous beginning of tendrils, picked out also the less doubtfully dis¬ 
advantageous weak stems. This difficulty has always been a fatal weakness 
in the explanation of adaptation by selection. 
This same correlation difficulty repeats itself over and over again in 
other cases, and is a very formidable one to whose explanation the theory 
of Natural Selection has been unable to give any clue. Climbing plants 
occur over and over again among closely related erect plants, whether in the 
same genus or in closely allied genera, so that it is clear that the climbing 
habit has been independently acquired in hundreds of cases, and therefore 
S s 
