NA TURE 



THURSDAY, JUNE ii, 1896. 



ON BEHALF OF SELECTION. 

 Ueber Germinal- Selection ; eine Qui-llc hestimmt gerich- 



teter Varin/ion. Von August Weismann. Pp. xi + 79- 



(Jena : Gustav Fischer, 1896.) 

 'T'HE special purpose of the present treatise, the 

 substance of which was given as an address at the 

 International Congress of Zoologists at Leyden in 1895, 

 is stated by the author to be the rehabilitation of the 

 principle of selection. This principle, though many 

 writers now seek to minimise or to dispense with it, still 

 appears to him to be absolutely necessary for any 

 scientific explanation of the problem of life. The only 

 alternative would be to allow the existence of teleological 

 contrivances, and this in science is inadmissible. The 

 theory of natural selection, says Prof Weismann, has 

 been rated too highly, and is now suffering the effects of 

 an inevitable reaction. It has not been overrated in the 

 sense of having been credited with too wide a sphere of 

 action, but in the sense that investigators have believed 

 that they understood its whole method of operation, and 

 had a clear conception of all its factors. This, however, 

 is not the case. It has been generally left out of account 

 that besides the individual or personal selection recognised 

 by Uarwin, there is a selective process always at work 

 between the various parts of the individual organism 

 (Roux), and even between the ultimate vital units within 

 the germ itself. This conception had already been 

 partly propounded by the author in his Romanes lecture 

 delivered at Oxford in 1S94, and in his last rejoinder to 

 Herbert Spencer ; ' it is here stated with greater com- 

 pleteness, and brought into more intimate relation with 

 the doctrine of selection as commonly understood. By 

 its means he claims to have advanced a more satisfactory 

 explanation of the origin of variations and their direction 

 along appropriate lines of development than any as yet 

 proposed. 



It is quite impossible to do justice to the view here 

 stated within the limits of a short notice such as this. 

 Those interested in the evolution controversy must be 

 referred to the treatise itself, where they will find the 

 author's position fully explained and illustrated, and from 

 which they will also be able to judge for themselves how 

 far his new conclusions are borne out by the facts and 

 reasoning at his command. The main heads of the 

 argument may, however, be briefly sketched as follows. 



The laws of variation provide the stones for the build- 

 ing, which are laid in place by selection. Our knowledge 

 of the selection-value of variations is necessarily limited ; 

 we are able, however, to adduce many cases of trans- 

 formation that can only be accounted for on principles of 

 utility. One such instance is the distribution of colour 

 in butterflies as between upper and under surface, and 

 fore and hind wing. For e.xample, while the upper side 

 of ProtOi^onius resembles a Heliconius, the under side is 

 like a leaf ; this must be a consequence of adaptation. 

 So, too, must be the correspondence of the hind wing 

 with the apex of the fore wing on the under surface of 



1 *' Neue Gedanken zur Vererbungsfrage." Jena, 1895. " Heredity 

 once more," Contemporary Knu'ew, September, 1895. 



NO. 1389, VOL. 54] 



many butterflies. In view of the fact that the wing-areas 

 so coloured accord with the usual posture of each species 

 during rest, it is absurd to talk here of simple " corre- 

 lation." Can mere " laws of development " account for 

 the fact that all leaf-like butterflies are wood-haunting 

 species ? The case of Kalli7na by itself is decisive for 

 adaptation. 



But how have the suitable variations, which have 

 culminated in such perfect adjustment to needs, origin- 

 ated in the right situations, and in correlation with the 

 appropriate instincts? Herbert Spencer applies Lam- 

 arckian principles to the explanation of functional 

 adjustments ; but this will not meet the case of such 

 parts as are purely passive in function. The current 

 conception of selection (i.e. individual selection) is also 

 inadequate to e.xplain instances of this nature. The root 

 of the process must lie deeper ; the variations in question 

 nmst be determined in the germ. This is also shown b\' 

 the dwindling of disused organs, which disappear in a 

 manner not to be explained by individual selection. 

 Lamarckism {pace Lloyd Morgan) will not ser\-e, even 

 as a working hypothesis ; and if this be the case, there 

 must be, as Osborn says, a hitherto unrecognised factor 

 in transformation : i.e. the direction taken by the varia- 

 tion of a part must be determined by utility. Known 

 facts, as for instance those of artificial selection, will 

 carry us a certain distance towards an explanation. In 

 such a case as that of the long-tailed poultry of Japan, 

 the variation must have been enhanced by selection, 

 and the germ itself must have undergone progressive 

 alteration. For further steps we must have recourse to 

 hypothesis. Variations oscillate about a mean, and 

 selection raises the mean to a higher point. This, is 

 satisfactorily accounted for by the theory of " determin- 

 ants." The determinants are subject to the same con- 

 ditions of nutrition as body-constituents of a higher 

 degree, and will accordingly differ in size and strength. 

 Hence the opportunity for the progressive raising of the 

 mean by mdividual selection. But a more important 

 principle is yet to be introduced. The phenomena of 

 retrogression in a disused part show that, as the 

 advocates of Lamarckism have rightly alleged, the 

 simple raising or lowering of the mean by "personal" 

 selection is not adequate to explain the facts. Panmixia 

 will account for the degeneration of such a part up 

 to a certain point, but not for a gradual and continued 

 dwindling ending in complete disappearance. 



The really efficient cause is germinal selection. This 

 rests on Roux's conception of the " struggle of parts," a 

 principle which must apply to the most minute units of 

 life, not only in the somatic, but also in the germ cells ; 

 not, however, of course to " molecules " in the chemical 

 sense. Panmixia starts the determinants of a dwindling 

 organ on the inclined plane, down which they are impelled 

 by intra-germinal selection to their final disappearance. 

 The progressive increase, no less than the decrease of a 

 part, nmst also be assisted by a like selective process 

 taking place within the germ. 



But it is necessary to show how simultaneous useful 

 variations arise under the law of selection. This follows 

 from the fact that the alterations of determinants in the 

 germ, when they are once set going by individual 

 selection, continue without needing the help of that 



G 



