Mechanism of Creative Evolution (1932)
C. C. Hurst

Three distinct systems of the working of modifying genes have so far been recognised in genetical experiments.

First, in the original case of the discovery of multiple genes by Nihlsson Ehle in which the three pairs of genes R, S and T that produce red grain in wheat have an equal cumulative effect, so that RrSsTt produces a medium shade of red, RRSSTT the darkest shade of red and Rrsstt, rrSstt and rrssTt the lightest shade of red, while rrsstt is white. In this way the three pairs of genes produce a graded series of shades from white to deep red, the frequencies of which, 1, 6, 15, 20, 15, 6, 1, fit the curve of probability. For many years geneticists have utilised Nihisson Ehle's polymeric scheme of equally cumulative dominant factors to cover the genetics of graded and measurable characters of size, weight and speed which present a graded series of results. This scheme, however, has been applied chiefly to grades of dominant characters.

Second, the brilliant genetical analysis of the Dutch rabbit by Punnett and Pease (1925) provides a scheme for the analysis of modifying genes in recessive characters. We have already seen that this scheme involves a series of dominant genes which produce an unequal cumulative effect in the same direction, giving a complicated and overlapping series of phenotypes, but the genotypes are the same and follow the usual Mendelian system.

Third, a similar scheme for the analysis of modifying genes in recessive characters has been employed with signal success by Philiptschenko (1927), the Russian geneticist, in his remarkable analysis of the genetics of the famous "Marquis" wheat. This system differs from that of Punnett and Pease in so far that it involves an equal cumulative effect in the same direction of dominant (and recessive) genes modifying a recessive character. Philiptschenko rounds off his research by applying the results to an interpretation of the genetics of musical ability in man, which so far fits the known data and is certainly the most stimulating and encouraging work in human genetics that we have had for many years.

This important paper was originally published in Russian and came into the hands of Prof. Engledow of Cambridge, and, thanks to his interest and that of Prof. Sir Rowland Biffen of the Imperial Bureau of Genetics, Cambridge, an excellent English translation has been made by Dr Hudson, the Assistant Director of the Bureau.

After several years of experimental crossings and inter-breedings among the different sub-species, varieties and pure lines of the Bread Wheats (Triticum vulgare), Philiptschenko succeeded in identifying six pairs of genes which in combination produce the peculiar broad-shaped grains and glumes of the famous "Marquis" wheat. This super-wheat originated in Canada from a single grain selected by C. Saunders in 1903 from a hybrid obtained by his brother, A. Saunders in 1892, and has proved to be of international value, being cultivated with success over vast regions in Europe, Asia and America. Its chief characters are its broad grains and glumes, the length-breadth index of which provided the main material on which the genetical analyses of Philiptschenko were based. By breeding pure lines of the "Marquis" Wheat (T. v. lutescens) and the narrow-grained ordinary Bearded Wheat (T. v. ferrugineum rossicum) for several seasons, he first ascertained the range of fluctuating variations in the shape of the grains due to the environment, which served to define at the outset the precise genetical differences between the two pure lines. He then crossed the two pure lines and found the narrow grains and glumes to be incompletely dominant, although it was difficult to distinguish F1 from the narrow parent P1. In F2 pure narrow types were obtained, but in accordance with the principle of the Engledow "shift", the broad "Marquis" type was not recovered until F3. Owing to frequent overlapping of the various grades it was necessary to test the F3 plants in F4, when the situation became clear and the complex results reduced to Mendelian order. In order to secure a complete genic analysis it was necessary to cross the "Marquis" with several other races and varieties of the species, and finally Philiptschenko succeeded in identifying the six pairs of genes which in combination with one another interacted to produce the true "Marquis" wheat. Of these six pairs of genes the basic pair organising broad glumes and grains was the recessive aa, the narrow glumes and grains of the common bearded type being organised by the dominant AA.

On this recessive base aa, five other pairs of modifying genes were found which, interacting with aa, produced the true "Marquis" wheat. Of these, two pairs proved to be homozygous dominant genes BB and CC, while the remaining three pairs proved to be recessive genes dd, ee and ff. The genic formula for the broad grains and glumes of the "Marquis" wheat may therefore be written:

[aa + (BB + CC) + (dd + ee + ff)].

These exhaustive genetical experiments in wheats by Philiptschenko lead to results which are of considerable technical importance to geneticists and at the same time mark a definite step in the progress of studies in human genetics. First, the experiments have determined the genie formula of the broad-grained super-wheat "Marquis", and also of several different varieties of compactum and other wheats used in the tests, as well as of those of the ordinary narrow-grained bearded wheats. As a step to the further improvement of the grain in wheats this knowledge is invaluable. Second, the experiments have interpreted and satisfactorily explained the true nature of the Engledow "shift" met with in many experiments with plants and animals, where the grandparental characters are not precisely recovered in F2 but are represented by a "shifted" form less in degree or size than in the original grandparent. Third, and perhaps most important of all, Philiptschenko proceeds to compare his results in wheat with the considerable data on the inheritance of musical ability in man that have been accumulated during the last twenty years by himself and others. He finds that if the family data showing high, medium and low grades of musical ability in different individuals be compared with the wheat data showing high, medium and low indices of broadness of the grain, the results of the matings high x high, high x low, low x low, high x medium, low x medium, and medium x medium show ratios remarkably similar in wheat and man. In order to cover the music data Philiptschenko's genie formula for high musical ability may be expressed as follows:

[mm + (AA + BB + CC)],

where "musical" is represented by the recessive pair of genes mm (recessive to the dominant "unmusical" MM) plus the three dominant modifying genes AA, BB and CC, each of which produces an equal cumulative effect in the direction of increasing the natural musical capacity represented by mm, and which in its weakest expression would be of the constitution mmaabbcc. On this scheme the generally accepted hypothesis originally proposed by Hurst (1908) (fig. 156) and supported by data of Davenport (1911), Drinkwater (1916), Philiptschenko (1917), Diakonov and Luce (1922), Stanton (1922), Sacharov (1924), and Mjöen (1926), that in general, basic musicalness is a Mendelian recessive character, is confirmed, while at the same time the contentions of Haecker and Ziehen (1923) and Strogaya (1926), that musical ability is a Mendelian dominant character, are equally confirmed. Indeed it is now clear that general musical capacity and special musical ability are organised by different and independent pairs of genes, the former being recessive and the latter dominant. As Philiptschenko points out, the valuable data of J. Mjöen (1925, 1926) of the pedigrees of famous musicians, and other musical families in which at least ten children are found, are in close agreement with the genic formula proposed, and there is no doubt that it also covers satisfactorily the other published data.

Philiptschenko does not publish details of the working of his scheme, and it may be interesting to note that on his formula at least seven grades or degrees of musical ability may be expected, which may be set out as follows:

Grade Class (Phenotype) Genotype Dominant
7 Classic musicians of the calibre of Bach, Beethoven, Wagner and Mozart mmAABBCC 6
6 Highly talented musicians mmAaBBCC 5
5 Talented musicians of more than average ability mmaaBBCC 4
4 Average musicians rnmaaBbCC 3
3 Moderate musicians of less than average ability mmaabbCC 2
2 Indifferent musicians mmaabbCc 1
1 Weakly musical mmaabbcc 0
0 Unmusical MmAABBCC or

In the musical grades 2 to 6, the five phenotypes may be of twenty-five different genotypes. Thus, grade 6 of highly talented musicians with five dominant genes for ability might be of genotype AaBBCC, AABbCC or AABBCc, while grade 2 of indifferent musicians with one dominant gene might be Aabbcc, aaBbcc or aabbCc, and so on with the other grades except 1 and 7, in which only a single genotype is possible. A striking feature of this scheme, which also is not mentioned by Philiptschenko, is the possibility that the unmusical MM and Mm individuals may be carrying the dominant genes for high musical ability which are inactive when M is present and can only be expressed in the recessive state mm. This explains many puzzling cases met with, where heterozygous unmusical parents produce musical children of considerable ability, and it also clears up a number of difficulties found in the Bach and other pedigrees. (Hurst, 1930 A.)

In the absence of suitable experimental psychological investigations we do not know precisely what the cumulative genes A, B and C represent. Being associated with the recessive genes mm for musical capacity, they may be simply degrees of musical ability, including capacity for musical expression, musical memory, and sensibility to melody, harmony and rhythm. On the other hand, these independent pairs of genes, which appear to be carried in different pairs of chromosomes in the wheats and rabbits, may represent other psychological characters not necessarily musical, such as high powers of imagination, capacity for emotion and high dramatic sense, or in some cases one might represent simply a capacity for concentration which is necessary for success in any sphere of life. This would increase the number of musical phenotypes and make them qualitative.

In the great musical composers all these qualities seem to be present, and it may be that in them more than four pairs of genes are concerned, just as on another plane the super-wheat "Marquis" has been built up by a peculiar and rare combination of six pairs of dominant and recessive genes. When one considers that the population of Europe is only in part composed of musical people with mm genes and that the vast majority of these belong to grades 4 to 1 with from none to three dominant genes for musical ability, the occurrence of a super-musician must be very rare indeed and the extreme rarity of a musical genius is explained. Further, when it is realised that of the total genes carried by an individual only one-half can be transmitted to his offspring, so that a musical genius with, say, six dominant genes for musical ability cannot transmit more than three of them to each of his offspring, it is seen that the chances of a genius producing a genius are very remote though not impossible, as a glance at the genic formulae shows. The same formulae also explain the fact, familiar in all musical pedigrees, that the more gifted the parent the more likely are the offspring to be gifted, and that the higher grades of musicians produce offspring of higher musical grades than the lower grades of musicians do.

The successful application of Philiptschenko's genetical wheat formula to the published data of the incidence of musical ability in human families encourages the hope that by similar methods of research we may be able to approach the most difficult problem of genetics and certainly the one most important to man and the human race, that of the genetical basis of the human mind. Notwithstanding the remarkable output of various school and industrial "intelligence tests" and the numerous studies of genius, insanity and mental defects that have been made, so far little has been done on modern genetical lines to attempt to solve the problems of the genetical basis of ordinary intelligence. It seems likely, however, that a detailed study of the dominant modifying genes associated with the genes for musical capacity may lead eventually to a study of the genetical basis of mind1. Both are, in any case, of a psychological nature and may in some cases be identical genetically. A study of the genetics of normal intelligence is surely the only sound foundation for future studies of genius, insanity and mental defects. These important problems, however, will not be solved satisfactorily on a simple Mendelian basis, they are far too complicated for that, and it is urgently necessary that the most recent methods and conclusions of genetical research should be applied to these problems. Owing to the complexities of the subject and the many technical difficulties involved, considerable team work is required to bring the research to a successful issue, and close co-operation between experienced geneticists and experimental psychologists is a primary necessity. Once the characters and genes in a series of families are genetically established the statistician can then step in and apply the results to the general population. Once the material has been analysed genetically the use of advanced statistical methods will be of real service and is indeed indispensable.

1Since the above has been set up in type, a genetical formula for the inheritance of general intelligence in man has been constructed and tested on two widely different sets of data (Hurst, 1932): (1) 194 Leicestershire families consisting of 388 parents and their 812 offspring individually examined and graded by the author during the last 20 years; (2) 212 Royal families of Europe consisting of 424 parents and their 558 offspring studied and graded by Dr F. Adams Woods (1906). Woods' ten grades of intelligence, which are based on those of Galton (1869 and 1892), are adopted for both sets of data and may be characterised as follows: grade 10 (Illustrious), grade 9 (Eminent), grade 8 (Brilliant), grade 7 (Talented), grade 6 (Able), grade (Mediocre), grade 4 (Dull), grade 3 (Subnormal), grade 2 (Moron), grade 1 (Imbecile).

The genetical formula in its most heterozygous form is:

Nn + (Aa +Bb + Cc +Dd +Ee)

where N is a major gene determining mediocre intelligence of grade, and A...E and a...e are minor increaser and decreaser genes respectively, acting only in the presence of nn. In the presence of the major gene N the minor increaser and decreaser genes are inactive and inhibited.

729 genotypes are possible, e.g. grade 10 (Illustrious) is nnAABBCCDDEE, grade 0 (Idiot) is nnaabbccddee and so on for grades 1-9, each dominant minor gene increasing the grade by 1 and each recessive decreasing it. Grade (Mediocre) may be, e.g. nnAaBbCcDdEe, etc., or (NN or Nn) + any minor genes. Since the minor genes, so far as the evidence goes, are quantitative and not qualitative, only five gametic types are possible, and on this basis an Expectation Table has been made which predicts the grades of intelligence expected in the offspring of the matings of all the grades with one another.

The formula applied to the Leicestershire and the Royal families shows that out of 1370 offspring 1343 (98.1%) are of the grades expected, while 27 (%) are exceptions. A large proportion of the exceptions occur in the offspring of pathological parents of the highest grades 10 and 9, pointing to a pathological disturbance due to the environment or to other genes.

The next step is to use this genetical formula as a working hypothesis and to test it further and more fully in other families and populations.