46 On MendeVs Categon'cs 



admitted variability of the races makes it possible for Mr Batcson to tcst his view 

 expcrimcntally ; for if a "recessive" plant, resnlting froni a cross, is " not nicrcly 

 liku but idontical with " one of the piire-bred plants used in the cross, then the 

 peas on the recessive phmt must cxhibit exactly the same series of colour-varieties 

 in their cotylcdons as thosc exhibited by the ptirc-brod plant which it resembles. 

 Before Mr Bateson can justify the view he has put foiwaid, he must therefore not 

 only find two races which obey Mcndel's laws when crossed, which he says he has 

 done, but he must determine 



(1) The variability of cotyledon colour in each race ; 



(2) The mean cotyledon colour and its variability in each plant used in 

 crossing ; 



(3) The inean cotyledon colour und its variability in each "recessive" and in 

 each "dominant" plant descended from the cross. 



Until the result of such determinations is known, it is impossible to distiuguish 

 between the rescmblance of a series of cross-bred plants to one of the anccstral races, 

 and their rescmblance to an individual plant of that race ; so that Mr Batesou's 

 contention cannot be supported by evidence. 



The confusion between rescmblance to a race and rcsemblance to an individual 

 involved in Mr Batcson's treatmcnt of Mendels work is one of the many unfoitunate 

 results which follow when Mendel's system of dividing a set of variable characters 

 into two categories, and of using these categories as Statistical units, is carried too 

 far. Unless the ränge of characters actnally includcd in each catcgory be constantly 

 borne in mind, the degrec of rcsemblance between two individuals, implied by 

 placing them in the same category, cannot be estimated ; and when, as constantly 

 happens, the ränge of Variation in one of the alternative categories differs widely 

 from that included in the othcr, the Mendeliaii System bccomes absolutely mis- 

 leading without some explanation (nearly always withhcld) of the real limits implied 

 by the terms used. Thus if two plants arc said to be glabrous, we know that thoy 

 arc absolutely similar in so far that they possess no hairs ; if they are said to be 

 hairy, we know that they both possess hairs, but one may, for anything we arc told 

 to the contrary, have ten times as many hairs, per unit of surface-ai"ea, as the 

 other. Again the tcndency to apply two categories, found suitable for a particular 

 race, to other races of the same or allied species leads to very harmful results; for 

 example the Classification of peas into thosc with green aud those with yellow 

 cotyledons leads to a (piitc crroncous conception of the distribution of cotyledon 

 colour in most existing races of peas, although it may have expressed the facts 

 observed by Mendel in the races which he used. 



The Report of Mr Bateson and Miss Saunders contains many statcnients which 

 would, I think, nevcr have been made if the authors had not been misled by the 

 use of Mcndelian categories. It is impossible to realise the mcaning of evidence, 

 brought forward to prove that particular hybrids behavc in the mauner described 

 by Mendel, unless the meaning of the categories employed in eacii special case is 



