September 3, 1908] 



NA TURE 



423 



engrams in which lies the secret of the ontogenetic rhythm. 

 But the mode of action of the mnemic nucleus is com- 

 pletely different from that of Weismann. He assumes that 

 the nucleus is disintegrated in the course of development 

 by the dropping from it of the determinants which regulate 

 the manner of growth of successive groups of cells. But 

 if the potentiality of the germ nucleus depends on the 

 presence of engrams, if, in fact, its function is comparable 

 to that of a nerve-centre, its capacity is not diminished 

 by action ; it does not cast out engrams from its substance 

 as Weismann 's nucleus is assumed to drop armies of deter- 

 minants. The engrams are but cut deeper into the records, 

 and more closely bonded one with the next. The nucleus, 

 considered as a machine, does not lose its component parts 

 in the course of use. We shall see later on that the nuclei 

 of the whole body may, on the mnemic theory, be believed 

 to become alike. The fact that the mnemic theory allows 

 the nucleus to retain its repeating or reproductive or mnemic 

 quality supplies the element of continuity. The germ-cell 

 divides and its daughter cells form the tissues of the 

 embryo, and in this process the original nucleus has given 

 rise to a group of nuclei ; these, however, have not lost 

 their engrams, but retain the potentiality of the parent 

 nucleus. We need not therefore postulate the special form 

 of continuity which is characteristic of Weismann 's theory. 



We may say, therefore, that the mnemic hypothesis 

 harmonises with the facts of heredity and ontogenv. But 

 the real difficulties remain to be considered, and these, I 

 confess, are of a terrifying magnitude. 



The first difficulty is the question how the changes 

 arising in the soma are, so to speak, telegraphed to the 

 germ-cells. Hering allows that such communication must 

 at first seem highly mysterious.' He then proceeds to 

 show how by the essential unity and yet extreme ramifi- 

 cation of the nervous system " all parts of the body are 

 so connected that w-hat happens in one echoes through the 

 rest, so that from the disturbance occurring in any part 

 some notification, faint though it may be, is conveyed to 

 the most distant parts of the body." 



A similar explanation is given by Nageli. He supposes 

 that adaptive, in contradistinction to organic, characters 

 are produced by external causes ; and since these characters 

 are hereditary there must be communication between the 

 seat of adaptation and the germ-cells. This telegraphic 

 effect is supposed to be effected by the network of idio- 

 plasm which traverses the body, in the case of plants by 

 the intercellular protoplasmic threads. 



Semon faces the difficulty boldly. When a new character 

 appears in the body of an organism, in response to chang- 

 ing environment, .Semon assumes that a new engram is 

 added to the nuclei in the part affected ; and that, further, 

 the disturbance tends to spread to all the nuclei of the 

 body (including those of the germ-cells), and to produce 

 in them the same change. In plants the flow must be 

 conceived as travelling by intercellular plasmic threads, 

 but in animals primarily by nerve-trunks. Thus the re- 

 productive elements must be considered as having in some 

 degree the character of nerve-cells. So that, for instance, 

 if we are to believe that an individual habit may be 

 inherited and appear as an instinct, the repetition of the 

 habit will not merely mean changes in the central nervous 

 system, but also corresponding changes in the germ-cells. 

 These will be, according to Semon, excessively faint in 

 comparison to the nerve-engrams, and can only be made 

 efficient by prolonged action. Semon lays great stress on 

 the slowness of the process of building up efficient engrams 

 in the germ-cells. 



Weismann - speaks of the impossibilitv of germinal 

 engrams being formed in this way. He objects that nerve- 

 currents can only differ from each other in intensity, and 

 therefore there can be no communication of potentialities 

 to the germ-cell. He holds it to be impossible that somatic 

 changes should be telegraphed to the germ-cell and be 

 reproduced ontogenetically — a process which he compares 

 to a telegram despatched in German and arriving in 



1 E. Hering in Ostw.ild's Klasslkcr rle>- exaktcn Wts^ettschaften, No. 

 148, p. 14; see also S. Butler's translation in "Unconscious NIemory," 

 p. iiq. 



-Weismann, ''The Evolution Theory," 1904, vol. ii. p 63; also his 

 " Richard Semon "s ' Mneme ' und die Vererbung erworbener Eigenschaften," 

 in the Archiv fiir Rassen- und Gesellscha/ts-Biologie, 1906. Semon has 

 replied in the same journal for 1907. 



NO. 2027, VOL. 78] 



Chinese. According to Semon,' what radiates from the 

 point of stimulation in the soma is the primary excitation 

 set up in the somatic cells ; if this is so, the radiating 

 influence will produce the same effect on all the nuclei 

 of the organism. My own point of view is the following. 

 In a plant (as already pointed out) the ectoplasm may be 

 compared to the sense-organ of the cell, and the primary 

 excitation of the cell will be a change in the ectoplasm ; 

 but since cells are connected by ectoplasmic threads the 

 primary excitation will spread and produce in other cells 

 a faint copy of the engram impressed on the somatic cells 

 originally stimulated. But in all these assumptions we are 

 met by the question to which W'eismann has directed atten- 

 tion — namely, whether nervous impulses can differ from 

 one another in quality?- The general opinion of physio- 

 logists is undoubtedly to the opposite effect — namely, that 

 all nervous iinpulses are identical in quality. But there 

 are notable exceptions, for instance, Hering,^ who strongly 

 supports what may be called the qualitative theory. I am 

 not competent to form an opinion on the subject, but I 

 confess to being impressed by Hering's argument that 

 the nerve-cell and nerve-fibre, as parts of one individual 

 (the neuron), must have a common irritability. On the 

 other hand there is striking evidence, in Langley's^ experi- 

 ments on the cross-grafting of efferent nerves, that here 

 at least nerve impulses are interchangeable and therefore 

 identical in quality. The state of knowledge as regards 

 afferent nerves is, however, more favourable to my point 

 of view. For the difficulties that meet the physiologist — 

 especially as regards the nerves of smell and hearing — are 

 so great that it has been found simpler to assume differ- 

 ences in impulse-quality, rather than attempt an explana- 

 tion of the facts on the other hypothesis.' 



On the whole it may be said that, although the trend 

 of physiological opinion is against the general existence of 

 qualitative differences in nerve-impulses, yet the question 

 cannot be said to be settled either one way or the other. 



Another obvious difficulty is to imagine how within a 

 single cell the engrams or potentialities of a number of 

 actions can be locked up. We can only answer that the 

 nucleus is admittedly very complex in structure. It may 

 be added (but this not an answer) that in this respect it 

 claims no more than its neighbours ; it need not be more 

 complex than "Weismann 's germ-plasm. One conceivable 

 simplification seems to be in the direction of the pangenes 

 of De Vries. He imagines that these heritage-units are 

 relatively small in number, and that they produce compIe.x 

 results by combination, not by each being responsible for 

 a minute fraction of the total result." They may be coin- 

 pared to the letters of the alphabet which by combination 

 make an infinity of words.' Nageli " held a similar view. 

 " To understand heredity," he wrote, " w'e do not need a 

 special independent symbol for every difference conditioned 

 by space, time, or quality, but a substance which can re- 

 present every possible combination of differences by the 

 fitting together of a limited number of elements, and which 

 can be transformed by permutations into other combina- 

 tions." He applied (/or. cit,, p. 50) the idea of a com- 

 bination of symbols to the telegraphic quality of his idio- 

 plasm. He suggests that as the nerves convey the most 

 varied perceptions of external objects to the central nervous 

 system, and there create a coherent picture, so it is not 

 impossible that the idioplasm may convey a combination 

 of its local alterations to other parts of the organism. 



.'\nother theory of simplified telegraphy between soma 

 and germ-cell is given by Rignano.' I regret that the 



* Semon, " Mneme," ed. i. p. 142, does not. however, consider it proved 

 'hat the nucleus is necessarily the smallest element in which the whole in- 

 heritance resides. He refers espec'ally to the regeneration of sections of 

 Stentor which contain mere fragments of the nucleus. 



2 I use this word in the ordinary sense without reference to what is known 

 as iiit^ifnliiv. 



'■'• " Zur Theorie'der Nerventh.^tigkeit." .'Vkademische Vortrag, 1898 (Veit, 

 Leipzig). 



■* Proc. R. Soc, 1904. p. 99. Journal 0/ Physiology, .vsiii. p. 240. and 

 xxxi p. 365. 



'^' .See Nagel, " Handbuch der Physiologic des Menschen," iii. (1905), 

 pp. i-is. 



6 De Vriec, "Intracellular Pansenesis," p. 7. 



" I take this comparison from I .otsy's account of De Vries's theory. Lotsy, 

 " Vorlesuncen iiher Deszendenztheorien," tgo6, i. p. 98. 



^ Niigeli's " Abstammungslehre," 1884. p. 73. 



i* For what is here given I am partly indebted to Signor Rignanos 

 letters. 



