236 



.VA TURE 



[January 9, 150S 



the biological characters of the staphylococci pathogenic 

 for man, and tests that will be useful for their differentia- 

 tion are described. 



It will thus be seen that the report contains matter of 

 considerable scientific and practical interest, and it would 

 be a great pity if this work were to be discontinued in 

 the future, as has been rumoured it may be. 



R. T. Hewlett. 



A' 



MENDELISM AND SEX} 



LL science is founded on observed facts. All authenti- 

 cated facts, no matter how observed, are valuable 

 to science. Many invaluable facts cannot be observed 

 without the aid of some special method, for example, 

 experiment ; but, of the total mass of facts garnered by 

 science, data furnished by experiment form a very small 

 part. Therefore to rely solely on experiment is to put on 

 blinkers. 



Since species are able to exist in their environments, 

 they are adaptational forms. The more minute our know- 

 ledge of a species, the more certainly are we able to 

 assign past or present utility to nearly all its structures 

 and faculties. Adaptation extends deeper than structures 

 and functions. Variability itself is adaptive. A greater 

 or lesser degree of variability is a variation and material 

 for natural selection. There is satisfactory evidence that 

 the average degree of variability displayed by every species 

 and structure is controlled by selection. The mass of 

 variations are " spontaneous." Thus there are hundreds 

 of human races and diseases, and every race is resistant 

 to every lethal disease in proportion to its past experience 

 of it. Therefore in this case there is adaptation. There- 

 fore it is clear that the poisons of disease, no matter how 

 virulent or universally experienced by the race, do not 

 cause alteration in the germ-plasm, and consequent racial 

 degeneration. On the contrary, since adaptation has 

 occurred, it is plain that variations are spontaneous, and, 

 since diseases are so many, that they occur all round the 

 specific mean. But some races {e.g. European dogs in 

 India) have been known to degenerate when removed to 

 new environments, where the native races flourish. There- 

 fore the insusceptibility of the germ-plasm to the direct 

 action of the environment has been established by natural 

 selection, and this, combined with the facts that (i) species 

 tend to become more variable a few generations after 

 removal from ancestral environments to which they have 

 become closely adapted, and where, therefore, nature limits 

 variability ; (2) the degree of variability in functionally 

 correlated structures (e.g. pairs of limbs) tends to be corre- 

 lated ; and (3) the greater the need for adaptation the less 

 is the degree of variability when once adaptation has been 

 attained (e.g. head and fore-foot of squirrels as compared 

 to tail), affords plain evidence that variability is under the 

 control of natural selection. 



When cessation of selection as regards any character 

 occurs, that character tends to retrogress. Therefore 

 retrogressive variations tend to predominate over pro- 

 gressive variations, whereby, without an increase of 

 mortality, species are rid of redundancies, both useless 

 variations and old-established parts which have become 

 useless. Since this tendency to retrogression is highly 

 adaptive, the presumption is that it is an adaptation. 

 That the retrogression which follows panmixia is not due 

 to reversed selection is shown by the fact that, though 

 variations favourable against all diseases occur in every 

 human race, yet they retrogress unless preserved by selec- 

 tion, for races become resistant only to those diseases to 

 which they are exposed. 



The two central doctrines of Mendelism are : — (i) segre- 

 gation of units, and (2) independent inheritance of 

 ch.iracters. Taken by itself, the doctrine of segregation 

 assigns no function to conjugation. It merely controverts 

 the doctrine of blending. Taken with the doctrine of in- 

 dependent inheritance, it assigns to conjugation the func- 

 tion of effecting an exchange of germinal units between 

 the two sets of parental units. That much Mendelism 

 implies — that much and no more. Mendelians believe, 



> AVtract of a rarer read before the Linnean Society on DeCEmber 19, 

 1907, by G. Archdall Reid. 



NO 1993 VOL. 77] 



apparently, that they have found the key to all the 

 problems of heredity ; but obviously -Mendelism is con- 

 cerned with nothing more than the function of conjuga- 

 tion. No other problem of biology with which it is con- 

 cerned can be thought of. However grandiose the 

 language used by its adherents, they are quite unable, 

 when challenged, to indicate any other. 



Mendelian inheritance is common when varieties which 

 have arisen under artificial selection are crossed. It is 

 comparatively rare when natural varieties (e.g. human) 

 are crossed. Blending is then the rule. Latent traits, 

 also, are commonly revealed by the crossing of artificial 

 varieties. In the whole range of biological literature, no 

 instance is recorded of a latent trait being revealed by the 

 crossing of natural varieties. Even when artificial varie- 

 ties are crossed, they never revert beyond the wild variety ; 

 that is, they never reveal traits that were latent in the wild 

 variety. Presumably, therefore, characters become latent 

 only under artificial selection, and consequently Mendelism 

 is concerned, not with the main problem of conjugation, 

 but only with certain anomalies which occur under con- 

 ditions of artificial selection. 



It is admitted on all hands that artificial selection is 

 founded mainly on mutations, and that the inheritance of 

 mutations tends to be alternative. It is admitted that the 

 inheritance of fluctuations tends to be blended, and the 

 evidence is conclusive that natural selection builds on 

 fluctuations. Thus varieties are most numerous when 

 mating individuals (e.g. birds) are enabled by good powers 

 of locomotion to interbreed over a wide area. No 

 interpretation of these facts save that of blending can be 

 thought of. Human varieties, for example, arise only under 

 conditions of geographical isolation. It has been said, on 

 the evidence of half-a-dozen generations, that mutations 

 are stable, and having arisen can be eliminated only by 

 selection. This implies that only progressive variations 

 occur in nature, and therefore that no structures ever dis- 

 appear or retrogress except through reversed selection ; 

 but though variations favourable against all diseases occur 

 in all human races, only those which are selected are 

 preserved and contribute to evolution. Therefore it is clear 

 that the rest retrogress, though there can be no rever.sed 

 selection in this case. 



When species are sexually dimorphic, mating individuals 

 differ, as a rule, little in non-sexual characters, but much 

 in sexual characters. Offspring reproduce either the 

 paternal or the maternal sexual characters. That is, the 

 reproduction of sexual characters is alternative, the male 

 and female characters being " allelomorphic " to one 

 another. But the inheritance of them is not alternative, 

 for each sex inherits the characters of the other in a latent 

 state, .ts is proved by a mass of evidence. Therefore, 

 though in bi-parental reproduction there is apparently no 

 blending as regards the sexual traits, yet the fact is that 

 the patent characters of the one sex blend with the latent 

 characters of the other. Sometimes the male characters 

 are latent for a long series of generations, as in aphides, 

 or apparently permanentlv, as in Cypris reptans. A 

 mutation, like a sexual difference, is a large dififerenoe, 

 and when an individual mutates and mates with the parent 

 type, the reproduction of the mutation tends to be 

 alternative. But the evidence is massive that the inherit- 

 ance is not alternative, but, on the contrary, that the 

 mutation is latent in those lines of descent which follow 

 the parent type, whereas the ancestral trait is Intent in 

 those lines which follow the type of the mutant. Like the 

 sexual traits in bi-parental reproduction, the recessive is 

 temporarily latent in the impure dominant. Like the male 

 characters in aphides and Cypris reptans, it is more or 

 less permanently latent in " pure " dominants, . as is the 

 dominant character in the recessive. This is proved by the 

 occasional occurrence of recessives in lines of *' pure " 

 domin.ants, and vice versa. It is even more decisively 

 proved by the reproduction of latent ancestral characters, 

 especially when (artificial) varieties are crossed. Cu^nnt's 

 theory of colour factors attempts to interpret in Mendelian 

 terms this fact of the reproduction of latent ancestml, 

 traits, but his hvnotlipsis totallv fails to account for 

 the reappearance of latent ancestral traits in cure lines 

 of descent, as, for example, when an aged fem.-ile bantam 

 reproduces, not the secondary main characters of her own- 

 variety, but those of an ancestral type. Here there can 



