Arber — Atavism and Law Irreversibility. 29 



habitual orientation, not however by losing their usual 

 twist, but always by adding a second torsion above the 

 first. This shows, as Lindman points out, how inveter- 

 ate (eingewurzelt) this habit has become; in other words 

 it demonstrates that when the plant can reach the same 

 goal either by retracing its steps, or by pursuing its 

 adopted path to a further point, it has an overwhelming 

 bias towards the latter course. 



Dollo's Law has been subjected to considerable criti- 

 cism, especially by the late Professor Errera 6 on the 

 botanical side, and recently by Dr. Boulenger 7 as a zoolo- 

 gist. I am not competent to discuss Dr. Boulenger 's 

 arguments, since their appraisement demands a famil- 

 iarity with vertebrate morphology which I do not pos- 

 sess; but I wish here to consider Errera 's objections, as 

 well as certain general considerations relating to animals 

 and plants which have been held to militate against the 

 Law of Irreversibility. 



It may be stated, broadly, that the opponents of Dollo's 

 Law regard it as disproved by the facts of i reversion ' — 

 that is by cases in which a variation appears which is 

 interpreted as an atavistic 8 ' throw-back' to an hypo- 

 thetical ancestor, and in which some character since lost 

 by the species makes a renewed appearance. It will be 

 necessary to analyse Errera 's criticisms — most of which 

 conform to this type — in some little detail, since he claims 

 that the instances he cites "suffisent a mon sens a refuter 

 la theorie de 1 'irreversibilite. ' ' The first phenomenon 

 to which he points as evidence is not, however, a case of 

 varietal reversion; it is the apetalous character of cer- 

 tain Caryophyllaceae, Rosacese, etc., which he regards as 

 a recurrence of "l'apetalie primitive des Angiospermes 

 inf erieures. ' ' But this apetaly cannot be treated as fur- 

 nishing an exception to the Law of Irreversibility if the 

 more modern view be accepted which holds that the 



6 1 am indebted to Mr. C. Davies Sherborn and to Dr. G. A. Boulenger, 

 for drawing my attention to Errera 's criticism, which is contained in Une 

 lecon elementaire sur le Darwinisme, Recueil d'ceuvres de Leo Errera, Bot. 

 Gen., 2, pp. 163-268, 1909. 



7 Boulenger, G. A., L 'evolution est-elle reversible? Considerations au 

 sujet de certains poissons, Comptes rendus des seances de l'Academie des 

 Sciences, 168, p. 41 (seance du 6 Janvier 1919). 



s In the present paper the word ' atavism ' is used in a broad sense as 

 synonymous with ' reversion ' — a sense in which it is habitually used in both 

 French and English non-scientific literature. The attempt to restrict it in 

 genetics to those cases in which some character of a grandparent is repeated 

 in his grandchild seems indefensible when it is remembered that 'atavus' 

 means 'great-great-great-grandfather' and is also used in the general sense 

 of ' ancestor. ' 



