l62 



NATURE 



[September 30, 1920 



make James I. the son of Elizabeth Well, but 



that's just what palaeontologists are constantly doing. 

 The famous diagram of the evolution of the horse 

 which Huxley used has had to be corrected in the 

 light of fuller evidence. Palaeotherium, which Huxley 

 regarded as a direct ancestor of the horse, is now 

 held to be only a collateral, as the last of the Tudors 

 were collateral ancestors of the Stuarts. The later 

 Anchitheriuni must be eliminated from the true line 

 as a side-branch — a Young Pretender. Sometimes an 

 apparent succession is due to immigration of a distant 

 relative from some other region — " the glorious House 

 of Hanover and Protestant Succession." It was, you 

 will remember, by such migrations that Cuvier ex- 

 plained the renewal of life when a previous fauna 

 had become extinct. He admitted succession, but not 

 descent. H he rejected special creation, he did not 

 accept evolution. 



Descent, then, is not a corollary of succession; or, 

 to broaden the statement, history is not the same as 

 evolution. History is a succession of events. Evolu- 

 tion means that each event has sprung from the 

 preceding one. Not that the preceding event was the 

 active cause of its successor, but that it was a 

 necessary condition of it. For the evolutionary 

 biologist a species contains in itself and its environ- 

 ment the possibility of producing its successor. The 

 words "its environment" are necessary, because a 

 living organism cannot be conceived apart from its 

 environment. They are important because they 

 exclude from the idea of organic evolution the hypo- 

 thesis that all subsequent forms were implicit in the 

 primordial protoplast alone, and were manifested 

 either through a series of degradations, as when 

 thorium by successive disintegrations transmutes 

 itself to lead, or through fresh developments due to 

 the successive loss of inhibiting factors. I say " a 

 species contains the possibility " rather than " the 

 potentiality," because we cannot start by assuming 

 any kind of innate power. 



Huxley, then, forty years ago, claimed that 

 palaeontologists had proved an orderly succession. 

 To-day we claim to have proved evolution by descent. 

 But how do we prove it? The neontologist, for all 

 his experimental breeding, has scarcely demonstrated 

 the transmutation of a species. The palaeontologist 

 cannot assist at even a single birth. The evidence 

 remains circumstantial. 



Recapitulation as Proof of Descent. 

 Circumstantial evidence is convincing only if in- 

 explicable on any other admissible theory. Such evi- 

 dence is, I believe, afforded by palaeontological 

 instances of Haeckel's law, i.e. the recapitulation by 

 an individual during its growth of stages attained 

 by adults in the previous history of the race. You 

 ail know how this has been applied to the Ammonites; 

 but any creature with a shell or skeleton that grows 

 by successive additions and retains the earlier stages 

 unaltered can be studied by this method. If we take 

 a chronological series of apparently related species or 

 mutations, a', a', a', a', and if in a* we find that 

 the growth-stage immediately preceding the adult 

 resembles the adult a', and that the next preceding 

 stage resembles a', and so on; if this applies miutatis 

 mutandis to the other species of the series ; and if, 

 further, the old age of each species foreshadows the 

 adult character of its successor, then we are entitled 

 to infer that the relation between the species is one 

 of descent. Mistakes are liable to occur for various 

 reasons, which we are learning to guard against. 

 For example, the perennial desire of youth to attain 

 a semblance of maturity leads often to the omission 

 of some steps in the orderly process. But this and 



NO. 2657, VOL. 106] 



other eccentricities affect the earlier rather than the 

 later stages, so that it is always possible to identify 

 the immediate ancestor, if it can be found. An 

 admirable example of the successful search for a 

 father is provided by R. G. Carruthers in his paper 

 on the evolution of Zaphrentis delanouei. Surely 

 when we get a clear case of this kind we are entitled 

 to use the word "proof," and to say that we have 

 not merely observed the succession, but have pr-oved 



the filiation. 



» * » 



The " Line-upon-Line " Method of Palaeontology. 



You will have observed that the precise methods 

 of the modern palaeontologist, on which this proof 

 is based, are very different from the slap-dash con- 

 clusions of forty years ago. The discovery of 

 Archaeopteryx, for instance, was thought to prove the 

 evolution of birds from reptiles. No doubt it 

 rendered that conclusion extremely probable, especially 

 if the major premise— that evolution was the method 

 of Nature — were assumed. But the fact of evolution 

 is precisely what men were then trying to prove. 

 These jumpings from class to class or from era to 

 era by aid of a few isolated stepping-stones were what 

 Bacon calls anticipations, "hasty and premature," but 

 " very effective, because as they are collected from a 

 few instances, and mostly from those which are of 

 familiar occurrence, they immediately dazzle the intel- 

 lect and fill the imagination" ("Novum Organon," i., 

 28). No secure step was taken until the modern palaeon. 

 tologist began to affiliate mutation with mutation and 

 species with species, working his way back, literally 

 inch by inch, through a single small group of strata. 

 Onlv thus could he base on the laboriously collected 

 facts a single true interpretation ; and to those who 

 preferred the broad path of generality his interpreta- 

 tions seemed, as Bacon says they always "must seem, 

 harsh and discordant — almost like mysteries of faith." 



I have long believed that the only safe mode of 

 advance in palaeontology is that which Bacon coun- 

 selled, namely, "uniformly and step by step." Was 

 this not, indeed, the principle that guided Linnsus 

 himself? Not until we have linked species into 

 lineages can we group them into genera ; not until 

 we have unravelled the strands by which genus is 

 connected with genus can we draw the limits of 

 families ; nor until that has been accomplished can we 

 see how the lines of descent diverge or converge, so 

 as to warrant the establishment of orders. Thus by 

 degrees we reject the old slippery stepping-stones that 

 so often toppled us into the stream, and foot by foot 

 we build a secure bridge over the waters of ignorance. 



The work is slow, for the material is not always 

 to hand, but as we build we learn fresh principles 

 and test our current hypotheses. To some of these 

 I would now direct your attention. 



Continuity in Development. 



Let us look first at this question of continuity. 

 Does an evolving line change by discontinuous steps 

 (saltations), as when a man mounts a ladder ; or 

 does it change continuously, as when a wheel rolls 

 uphill? The mere question of fact is extraordinarily 

 difficult to determine. Considering the gaps in the 

 geological record, one would have expected palseonto- 

 logists to be the promulgators of the hypothesis of 

 discontinuity. They are its chief opponents. 



.\gain I must leave the facts and their interpreta- 

 tion, merely reminding you of such cases as the 

 heart-urchins or Micrasters of the Chalk. Here, 

 where we have a fairly continuous succession of many 

 hundred feet of similar rock, we do find a slow and 

 gradual change, such that no clean line can be drawn 

 between one form and its successor. 



