212 THE MECHANISM OF LIFE 



Now we cannot do any of these things. We cannot predict 

 what an individual man or woman will do at any time in the i 

 future, though we can sometimes say what a population will do. 

 (The reader must note this distinction between individual and 

 statistical effects; it is most important in speculative reasoning.) 

 We cannot say, merely from a knowledge of its structure and 

 behaviour, what was the past of the species to which an animal 

 / , belongs, and what is going to be its future. Let there be no 

 misunderstanding here: we do know a great deal about the j 

 phylogenies (or lines of descent) of some races of animals, but 

 that is because these evolutionary careers have left historical 

 records. Thus we know (or at least we believe) that the existing 

 one-toed horse has descended from a three-toed species which 

 had, in its turn, come down from a five-toed form, but this is 

 because we have fossil remains of the three- and five-toed horses, 

 because the living animal has vestiges of the second and fourth 

 toes, while the skeleton of the three-toed horse has vestiges of the 

 first and fifth digits. Evolutionary careers have therefore left 

 records in the form of fossil remains and vestigial structures, and 

 we are sometimes successful in reading these and so in tracing 

 out a line of descent. Obviously our success is not due to any 

 process of true deduction from a few fundamental concepts, as i 

 is the astronomer's when he calculates backwards to find when 

 eclipses occurred. Eclipses leave no records on the faces of sun, 

 moon, and earth like the evolution of the horse has done. 



And so the astronomer can calculate forward or predict, while 

 the biologist cannot do so, since the successes of the latter depend 

 upon records which cannot exist in the cases of events which 

 have not yet happened. Therefore we cannot predict what will 

 be the future evolutionary history of the horse or of any other 

 race of existing animals or at least no biologist has yet risked 

 his reputation by making such an attempt ! 



Now we may as well admit that this argument may not appear 

 to possess much force in itself. It may be that biology cannot 

 predict (and so supply the necessary test of the validity of its 

 theories) just because it has not yet attained the necessary 

 knowledge. No doubt Copernicus could not have predicted the 

 times at which equinoctial spring tides would occur at any 

 particular place in the North Sea, for his knowledge of planetary 

 theory was imperfect, and he had not obtained the concept of 

 gravitation. So it may be that biology cannot yet predict just 



