October i8, 1906] 



NA rURE 



609 



I turn to the last point first. I venture to think that the 

 proljlem of homoganiy is essentially the same in relation to 

 both MetaEoan and Protozoan, when we are considering 

 its effect on the possible differentiation of species, and 

 endeavouring to surmount Huxley's difliculty. Having 

 shown myself that homoganiy certainly does exist in one 

 type of Metazoa, it was necessary that it should be shown 

 10 exist in the Protozoa, and for one type this is what 

 Dr. Pearl has achieved. 



Mr. Lister only obscures Dr. Pearl's statement as to 

 the persistency of type in cohjugant Paramajcia. Had he 

 understood the constants dealt with by Dr. Pearl in this 

 part of his paper he would have seen that they were what 

 in biometry are termed intra-racial and not inter-racial 



' values. The conclusions of Dr. Pearl have nothing lo do 

 with the inter-racial differentiation of gametes. Dr. Pearl 

 grew Paranijecia under much variety of environment, and 

 found that the non-conjugant type was highly correlated 

 with the environment and the conjugant type singularly 

 little affected by the environment. The whole inquiry was, 

 of course, undertaken to illustrate Weismann's position, 



' that while acquired characters are not inherited, the 

 environment can influence inheritance where one cell is 

 both soma and germ. In biology it has become almost 

 axiomatic to assume that the Protozoa can inherit 

 acquired characters on account of this identity, while in 

 the Metazoa the acquired character of the soma is at the 

 very least not usually inherited. Dr. Pearl brings out the 

 .ill-important point that the gamete in Paramoecium is 

 not, like the non-conjugant cell, markedly influenced by 

 the environment. If Mr. Lister assumes that the characters 

 acquired by the somatic cells are handed over to the 

 gametic ceils, this is, of course, to sweep away entirely 

 the Weismannian hypothesis, and we may reasonably ask 

 him for the quantitative proof of this assumption. The 

 proof will at any rate go to the basis of the current hypo- 

 thesis of "gametic purity." Mr. Lister asks for evidence 

 of any relation between external characters in man and 

 hi-i gamete. The problem is not this, but the relation 

 between the external characters in two phases of a cell 

 which can be watched in passing from its conjugant to 

 non-conjugant conditions, and that is an entirely different 

 matter. Even here biometricians will shortly be prepared 

 with an answer to the thus restated question,' although it 

 ha-i no relation whatever to Dr. Pearl's main point. 



The remainder of Mr. Lister's letter would never have 

 been written had he studied Dr. Pearl's paper or measured, 

 as the latter has done, five or six thousand Paramascia. 

 The passage in Maupas was sufficiently familiar to me, and 

 is actuallv referred to, together with the previous work 



I of Hertwig. Gruber. and others on differentiation by Dr. 

 Pearl himself in his paper. But the differentiation of two 

 populations can only be demonstrated by an accurate 

 quantitative investigation of the means, variabilities, and 

 correlations of those populations. It may bo rendered 

 " possible or probable," as Mr. Lister held in August by 

 a statement conveyed without detailed measurements in 

 seven lines of print. In fact, Maupas says he has never 

 found conjugants to exceed 225 /t, while Dr. Pearl, 

 measuring immensely larger numbers, has found individuals 

 up to 2S5 11, a value considerably in excess of what was 

 reached by the largest non-conjugant in his measure- 

 ments, 275 ^. It will be clear in the face of such results 

 that demonstration can onlv follow a study of large numbers 

 and their proper statistical treatment. 



Mr. Lister next proceeds to state that " every practical 

 biologist knows " that speciinens which have been pre- 

 <prved and fixed w^ill be distorted. Will the reader credit 

 the fact that pages of Dr. Pearl's memoir are devoted to 

 a discussion of the methods needful to avoid distortion? 

 Mr. Lister's statement only amounts to the confession that 

 he himself cannot prepare undistorled specimens. The 

 " practical biometrician " knows that the distortion can 

 he almost entirely avoided by instantaneous killing of the 

 Param^cia and the avoidance of diffusion currents in 



1 T will venture on a prophecy, rash as it may seem, but based upon a 

 general experience of cell division where the mother and daughter cells have 

 not been measured under like phases, that the correlation will not be found 

 to b; less than 0-5 or more than 0^7. 



changing to the higher grades of alcohol. The method of 

 attaining these results is amply discussed by Dr. Pearl, 

 but their application depends, as in all such matters, on 

 long practice, on training and on technique. It is open to 

 Mr. Lister to assert that the methods adopted by Prof. 

 Worcester and Dr. Pearl failed in their object ; it is not 

 open to him to insinuate that they have overlooked a dis- 

 tortion the danger of which is obvious to the merest tyro 

 in biology. But as he has not seen the preparations he 

 can only defend his assertion on the ground that the 

 measurements show marked evidences of the irregularities 

 which would be produced by such distortion ; and here 

 we see at once the absence of thorough examination of 

 Dr. Pearl's paper by Mr. Lister. Only three of the four- 

 teen series discussed by Dr. Pearl were from preserved 

 specimens. The methods employed in the other cases w^ere 

 different, but the many series were not mi.xed, as might 

 be inferred from Mr. Lister's statement. Even including 

 Dr. Simpson's measurements, made in a wholly different 

 manner, there is strilcing general agreement which is 

 absolutely inconsistent with the amount of variation which 

 would arise in the case of largely distorted forms. As 

 Dr. Pearl himself says, " The good agreement is some- 

 thing . . •. which probably no biologist would have pre- 

 dicted before the measurements were made. One has been 

 accustomed to think of Paramascium as a soft-bodied 

 creature likely to show great and altogether irregular 

 fluctuations." He then points out that Paramecium is 

 less variable than Arcella, Eupagurus prideauxi, or 

 Ophiocoma nigra, all of which organisms have a more or 

 less firm exo-skeleton. So much for the question of the 

 influence of distortion. 



Mr. Lister next proceeds to the assertion that Para- 

 ma^cium being a " slipper-shaped " ( !) animal, there 

 would be difficulty in measuring the breadth of the non- 

 conjugant as compared with the conjugant. He would 

 never have made this statement had he read Dr. Pearl's 

 paper, where the relative weight of length and breadth 

 measurements is considered at great length. The- 

 statement is further mere hypothesis, and not the experi- 

 ence of one who has learnt to measure Paramsecia. The 

 difTiciilty of measuring the breadth lies -Jiiith the conjugating 

 individuals and not with the non-coniugating individuals. 

 for reasons amply set forth by Dr. Pearl. In the next 

 place. Dr. Pearl's main argument is drawn, not from the 

 breadth, but from the length measurements, and, lastly, 

 had Mr. Lister followed the significance of biometric con- 

 stants he would at once have seen that his hypothesis was 

 invalid. If the measurement of the breadth were affected 

 bv a large source of error due to diversity of aspect when 

 the Paramasciuni is measured after death, there would be 

 little or no organic correlation between length and breadth. 

 Dr. Pearl shows that the actual correlation is markedly 

 higher for the non-coniugants than for the conjugants, and 

 is of an intensity which we might reasonably exqect from 

 previous investigations of similar organic correlations. 



Lastly, Mr. Lister proceeds gravely to inform us of 

 another source of error which he supposes to exist. He 

 savs that Dr. Pearl's non-coniugant population consists of 

 heterogeneous material, in which the variability would be 

 increased by the fact that it contained all stages of in- 

 dividuals in process of differentiation into gametes. It 

 seems astonishing to have to state it, but as a matter of 

 fact no less than six control series of Paramrecia, in which 

 no conjugation at all took place, and which each numbered 

 500 individuals, are dealt with by Dr. Pearl in his paper 

 and compared with the series of conjugant Paramsecia. 

 The comparison between conjugants in a preparation and 

 the nearest non-conjugants was made primarily to ascer- 

 tain whether the cultures were in any state of local hetero- 

 geneitv, so that a spurious correlation between conjugants 

 would necessarily arise from their being drawn from the 

 same part of the culture. It was precisely of the nature 

 of the test used hv Prof. Weldon and myself to ascert.ain 

 if locality influenced the value found for assortative mating 

 in man. 



Further, the slightest examination of Dr. Pearl's 

 diagrams would have shown Mr. Lister that the so-caljeJ 

 non-conjugant population is widely separated In distribution- 



NO. 1929, VOL. 74] 



