174 



NATURE 



\yune 27, 1872 



must be room for very tjreiit variety, because a single strain of 

 impure blood will reassert itself after more than eight generations; 



(2) an individual has 256 progenitors in tlie eighth degree, if 

 tliere have been no ancestral intermarriages, wfhile under tlie 

 ordinary conditions of social and neighbourly life, he will cer- 

 tainly have had a considerable, though a smaller, number of them ; 



(3) the gradual waning of the tendency to reversion as the gene- 

 rations increase, conforms to what would occur if each fresh 

 marriage contriliuted a competing element for the same place, 

 thus dikithig the impure strain until its relative importance was 

 reduced to an insignificant amonnt. It follows from these ar ni- 

 ments that for each place among the personal elements there 

 may exist, and probably often does exist, a great variety of latent 

 elements that formerly competed to fill it. 



I have spoken of the primary elements as they exist in the 

 newly-impregnated ovum, where they are structureless, but con- 

 tain the materials out of which structure is evolved. The em- 

 bryonic elements are segregated from among them. On wliit 

 principle are they segregated ? Clearly it is on some principle 

 whose effects are those of "Class Representation," using th.at 

 phrase in a perfectly general sense, as indicating a mere fact, ami 

 avoiding any hypothesis or affirmation on points of detail, about 

 most, if not all, of which we are profoundly ignorant. I give as 

 broad a meaning to the expression as a poliiician would give to 

 the kindred one, a " representative assemljly." By this he means 

 to say that the assembly consists of representatives from various 

 constituencies, which is a distinct piece of information so far as 

 it goes, and is a useful one, although it deals with no matter of 

 detail ; it says nothing aliout the number of electors, their quali- 

 fications, or the motives by which they are inlluenced ; it gives 

 no information as to the number of seats ; it does not tell us how 

 many candidates there are usually for each seat, nor whether the 

 same person is eligible for, or may represent at the same time, 

 more than one place, nor whether the result of the elections at 

 one place may or may not influence those at another (on the 

 principle of correlation). After these explanations there can, I 

 trust, be no difficulty in accepting my definition of the general 

 character of the relation between the embryonic and the structure- 

 less elements, that the former are the result of election from the 

 latter on some method of Class Representation. 



The embryonic elements are drudoped into the adult 

 person. "Development" is a word whose meaning is 

 quite as distinct in respect to form, and as vague in re- 

 spect to detail, as the phrase we have just been considering ; 

 it embraces the combined effects of growth and multiplication, 

 as well as those of modification in quality and proportion, under 

 both internal and external influences. If we were able to obtain 

 an approxunate knowledge of the original elements, statistical 

 experiences would no doubt enable us to predict the average 

 value of the form into which they would become developed, just 

 as a knowledge of the seeds that were sown would enable us to 

 predict in a general way the appearance of tlie garden when the 

 plants had grown up. But the individual variation in each case 

 would be great, owing to the large numljer of variable influences 

 concerned in the process of development. 



The latent elements in the embryonic stage must be developed 

 by a parallel, I do not say by an identical process, into those of 

 the adult stage. Therefore, to avoid all chance ot being misap- 

 prehended when I collate them, I will call, in the diagram I am 

 about to give, the one process " Development {a) " and the other 

 " Development (/')■" 



It is not intended to affirm, in making these subdivisions, that 

 the embryonic and adult stages are distinctly separated ; thev are 

 continuous, and it is impossible but that tliey should overlap, 

 some elements remaining embryonic while others are completely 

 formed. Nevertheless the embryo, speaking broadly, may fairly 

 be looked upon as consecutive. 



Again, the two processes are not wholly distinct ; on the con- 

 trary, the embryo, and even the adult in some degree, must re- 

 ceive supplementary contributions derived from their contempo- 

 rary latent elements, because ancestral qualities indicated in early 

 life frequently disappear and yield place to others. The reverse 

 process is doubtful ; it may exist in the embryonic stage, but it 

 certainly does not exist in a sensible degree in the adult stage, 

 else the later children of a union would resemble their parents 

 more nearly than the earlier ones. 



Lastly, I must guard myself against the objection, that though 

 structure is largely correlated, I have treated it too much as con- 

 sisting of separate elements. To this I answer, first, that in 

 describing how the embryonic were derived from the structureless 



elements, I expressly left room for a small degree of correlation ; 

 secondly, that in the development of the adult elements of the 

 embryonic, there is a perfectly open field for natural selection, 

 which is tlie agency by which correlation is mainly established ; 

 and thirdly, that correlation affects grou]is of elements, and not 

 the complete ]ierson, as is proved by the frequent occurrence of 

 small groups of persistent peculiarities, which do not affect tli-- 

 rest of the organism, so far as we know, in any way whatever. 



The ground we have already gained may be described as 

 follows : — 



Out at the stnictureless ovum the embryonic elements are 

 taken by Class Representation, and these are developed (n) into 

 the visible adult individual. On the other hand, returning to 

 our starting-point at the structureless ovum, we find, after the 

 embryonic elements have been segregated, the large residue is 

 developed (/') into the latent elements contained in the adult in- 

 dividual. All this is summarily expressed in the first two 

 columns of the diagrams below. I might have inserted vertical 

 arrows to show the minor connections between the corresponding 

 stages in the two parallel processes, but it would have com- 

 plicated the figure. 



In what way do the patent and latent adult elements respec- 

 tively contribute rejiresentatives towards the structureless stage 

 of the next generation ? We know that every quality they 

 possess may be transmitted to it, but it does not follow that they 

 are invariably transmitted. The contributions from the patent 

 elements cannot be by " Class," because their own original ele- 

 ments have been themselves specialised, and therefore can con- 

 tain no more than one or a few members of each class {which, it 

 is true, must have been somewhat developed, both in numbers 

 and variety). Their contributions may therefore be justly de- 

 scribed as being effected on some principle that has resulted in a 

 '• Family representation," though whether in a strictly universal 

 representation I do not profess to say. 



As regards the large variety of adult latent elements, they 

 cannot all be transmitted, for the following obvious reason ; the 

 corresponding qualities of no two parents can be considered 

 exactly alike ; therefore the accumulation of sub-varieties, if they 

 were all preserved, as the generations rolled onwards, would 

 exceed in multitude the wildest flights of rational theory. The 

 heritage of peculiarities through the contributions of 1,000 con- 

 secutive generations, even supposing a great deal of ancestral 

 intermarriage, must far exceed what could be packed into a 

 single ovum. The contributions from the latent adult elements 

 are therefore no more than representative ; but we know they 

 cannot be so on the broad principle of "class representation," if 

 the word " class " be applied to the same large orders as before, 

 and if the representatives are few in number, because it is incum- 

 bent on them to furnish all the various members of each Class 

 whence the representatives have to be drawn. Therefore, bear- 

 ing in mind what has been just argued, that it is impossible for 

 the elements of every individual quality to be contributed, we 

 are driven to suppose, as in the previous case, a " Family Repre- 

 sentation," the similar elements contributed by the two parents 

 r.anking, of course, as of the same family. It is most important 

 to bear in mind that this phrase slaves a fact and not an hypo- 

 thesis ; it does not mean that each and every Family has just one 

 representative, for it is absolutely reticent on all matters of de- 

 tail, such as those I enumerated, when speaking of Class Repre- 

 sentation. To show the importance which I attach to this dis- 

 claimer, I may lie permitted to mention what appears to me the 

 most probable modus operandi, namely, that it is in reality a large 

 selection made on a broader and not a narrower system than 

 that of classes, and similar to that obtained by an indiscriminate 

 conscription ; thus, if a large army be drawn fr mi the pro -incos 

 of a country by a general conscription, its constitution, accord- 

 ing to the laws of chance, will reflect with surprising precision 

 the qualities of the population whence it was taken ; each village 

 will lie found to furnish a contingent, and the composition of the 

 army will be sensibly the same as if it had been due to a system 

 of immediate representation from the several villages. 



The following diagram expresses the whole of the foregoing 

 results : — It begins with the structureless elements, whence the 

 parent individual was formed, and ends with its contributions to 

 the structureless elements, whence his offspring is formed. 



I will now inquire, what are, roughly speaking, the relative 

 ]5roportions of the contributions to the elements of the offspring 

 made respectively by the patent and latent elements of the ailult 

 p.arent? It is better not to complicate the inquiry by sijeaking, 

 at first, of these elements in their entirety, but rather of some 



