64 As already explained, the super-gene controlling polymorphism may take the form of a long inversion. That situation has been investigated on a number of occasions and with valuable results, but the work of Dobzhansky and his colleagues on Drosophila pseudoobscura and D. persimilis has provided by far the most thorough analysis of it. Using the techniques of ecological genetics they have demonstrated that the inversion polymorphism of these species is maintained by heterozygous advantage, and that this is subject to powerful selection which, indeed, has been responsible for its evolution. Some aspects of their researches are particularly relevant to the present discussion and they must briefly be considered at this point.
In the first place, the existence of heterozygous advantage, detected in a number of Drosophila species, has been established by two distinct techniques: ecological and genetic. Numerous instances have been found in wild populations of D. pseudoobscura and D. persimilis in which the inversion heterozygotes exceed expectation, assuming equal viability for all three phases or 'genotypes' (Dobzhansky, 1947). Indeed in other species (D. willistoni, D. paulistorum and D. tropicalis) heterozygotes amount to more than half the individuals in some localities (Pavan, Dobzhansky and da Cunha, 1957). This they cannot do when the values are stable unless favoured by selection since, without differential viability, they take the value 2pq, where p2 and q2 are the frequencies of the respective homozygotes.
Dobzhansky (1951, p. 117) has also demonstrated such heterosis by an entirely different method, so providing exceptionally strong, because independent, proof of it. That is to say, when D. persimilis carrying the inversions known as 'Standard' (ST) and Chiricahua (CH) are reared in 'population cages' in optimum conditions, the three genotypes prove to be distributed according to the binomial square rule. Thus the polymorphism cannot be perpetuated by non-random mating, as in Panaxia dominula (pp. 42-3). 65 When, on the other hand, the larvae are in competition for a restricted food-supply, it is found that the proportion of heterozygotes exceeds expectation among the imagines to which they give rise. We here have a further, and entirely distant, demonstration that the polymorphism is maintained by heterozygous advantage.
An important question arises at this point. Is such 'single-gene heterosis' an essential property of the heterozygous state or does it evolve, on the lines indicated on pages 26-8, when a previously disadvantageous gene becomes an asset and begins to spread. The latter alternative has now been proved correct.
Dobzhansky (1950) has repeatedly bred stocks derived either from Mather or from Pi�on Flats, California, and has always found that the heterozygotes are superior to both homozygotes in crosses involving any combination of Standard (ST), Chiricahua (CH), Arrowhead (AR) and other types. For instance, the relative advantages of AR/AR, AR/CH and CH/CH in Pi�on Flats stocks kept at 25 deg.C prove to be 0.71 : 1 : 0.43. The situation is, however, very different when flies from these two localities, which are 300 miles apart, are interbred. For then the heterozygotes are no longer superior but intermediate in viability between the homozygotes. Dobzhansky also examined the survival of these inversion types in pairings between races that are geographically still more remote; that is to say, from Pi�on Flats and Chihuahua, Mexico. These are separated by 700 miles of country which includes extensive tracts of desert. In broods so obtained, and segregating for ST and CH, the viability of the three genotypes was in the ratio of 1.26 : 1 : 0.87 taking the heterozygotes as unity; but in the AR, CH cross, the heterozygotes were actually inferior to both homozygotes in the proportion 1.53 : 1 : 1.16.
The differing heterozygous advantages of the various inversions are adjusted to distinct environments, whether in space or in time. Thus we may consider the frequencies of the types in the gene-pool of Drosophila pseudoobscura (averaged, that is to say, throughout the season). Standard (ST) is the commonest form along the Pacific coast of California, where it varies from 50 to 60 per cent with Arrowhead (AR) at only 10 to 30 per cent. 66 Moving eastwards, across the Sierra Nevada and Cascade mountains, their frequencies are reversed: a tendency which continues until, in Utah and Arizona, AR occurs in 85 to 95 per cent of the chromosomes and all the others, including ST, are rarities (Dobzhansky, 1951, pp. 136-7).
Similarly the inversion-frequencies of this species are sharply adjusted to altitude. At the season when ST amounts to 46 per cent and AR to 25 per cent at 850 feet at the base of the Sierra Nevada, the ST phase drops to 25 per cent and AR rises to 44 per cent at 6,200 feet. That change is continued, so that at 10,000 feet 10 per cent of the chromosomes carry ST and 50 per cent carry AR.
It is of particular relevance to analyse the seasonal adjustments to which the inversions of Drosophila pseudoobscura are subject. Thus at Mount San Jacinto, Southern California, the ST type decreases from approximately 53 per cent in March to 28 per cent in June, and CH becomes commoner, changing from 24 to 40 per cent. Meanwhile the AR frequency fluctuates throughout the whole period. However, further north, at Mather near the latitude of San Francisco, where the seasonal variation in inversion-frequency is less, though still striking, ST becomes commoner (21 per cent in March to 30 per cent in October) AR becomes rarer (45 per cent in March to 33 per cent in October) while CH remains fairly constant (11 per cent in March to 9 per cent in October) throughout the whole season. Here, then, unlike those of the Mount San Jacinto populations, the proportions change during hibernation, returning from their autumnal to their spring values. Dobzhansky (1961) finds that this species passes through no more than six to eight generations annually in California, probably less elsewhere, so that very powerful natural selection is needed to produce such seasonal changes as these.
The superiority of the inversion-heterozygotes is expressed in various ways. Moreover, it becomes progressively more evident as the conditions deteriorate, and this includes overcrowding as well as adverse changes in the general environment. 67 These produce differential mortality as well as other effects, such as alterations in sexual activity and in rate of development, in which the heterozygotes are at an advantage.
Dobzhansky's work upon Drosophila pseudoobscura has borne fruit in another direction. He has detected (1958) a remarkable evolutionary change now in progress. This is related to the spread of the Pike's Peak (PP) inversion which, though always common in Texas, was a great rarity in California up to 1946: so much so that only 4 instances of it had been detected among 20,000 third chromosomes studied by that date, whereas about 8 per cent of them now carry it. This has, of course, necessitated adjustments in other inversion-types; the proportion of Chiricahua has greatly declined and so, to a lesser extent, has Arrowhead, while Standard has somewhat increased. A most curious, and so far unexplained, circumstance is that these changes have occurred to an approximately equal extent throughout the greater part of the State. Not only is this area immense (about 200,000 square miles) but it is exceedingly diversified; yet to the extreme regional differences involved, these inversion-frequencies, normally so sensitive to environmental change (pp. 65-6), respond not at all. They are the same in the great urban districts of San Francisco and Los Angeles as in the remote countryside, and the same also in coastal, inland-valley and mountain regions.
It seems, therefore, that we are not here faced with a change adjusted to smoke pollution (as with industrial melanism in the Lepidoptera) nor to insecticides, widespread and sinister as their use in the U.S.A. has become. The spread of the PP chromosome-type from Mexico, where it has always been established, seems excluded on two counts. First, its frequency has not increased in the intervening State of Arizona. Secondly, the amount to which D. pseudoobscura is able to scatter has been determined with accuracy and proves to be on an entirely wrong time-scale to account for this event (the flies extend to a radius of 1.76 kilometres from a releasing point in 10 months).
An obvious possibility is that the Californian climate has altered in a way that favours PP and is unsuited to CH. 68 As already mentioned, urbanization and the use of insecticides do not influence the frequency of the inversions; moreover, though there is a testing ground for atomic weapons at Charleston Peak, Dobzhansky points out that the prevailing winds tend to carry the fall-out away from California. It can only be said that if a recent climatic change has affected the area, this has not been detected. The whole U.S.A. must indeed be influenced by the slight universal trend to warmer weather indicated by the world-wide retreat of glaciers, but this does not at all accord with an event beginning in 1946 and subsequently following a rapid course, as in the spread of the PP phase of Drosophila pseudoobscura.
Dobzhansky is inclined to think that some genetic adjustment, due to mutation or to an appropriate chromosome-reconstruction, has occurred in the gene-complex of D. pseudoobscura in California to promote the observed alteration in its inversion-frequencies. This indeed seems the most probable explanation, but the vast area over which that occurrence has been recorded within a few years remains a quite mysterious feature of the situation. It is true that, as Remington has shown, air currents can draw insects much larger than Drosophila up to considerable heights, thence to scatter widely, but this hardly seems to account for the extraordinary effects under consideration. Here we have evidence for an evolutionary change in progress, but it requires further study before its cause can be elucidated.
Heterozygous advantage is evident also when the control of polymorphism involves whole chromosomes (pp. 23-4). 69 An instance of the kind, worked out in considerable detail both on the ecological and cytological side, is provided by the marine Gastropod Purpura (Nucella) lapillus.
Staiger (1954, 1957) has shown that two basic forms of this species exist, with n=13 and n=18 chromosomes. All the possible intermediates between them, amounting to 243, can be produced and seem to be viable and inter-fertile whether heterozygous or homozygous. That situation is the result of a delicate cytological adjustment. In the first place, all individuals possess eight chromosome-pairs that remain constant in structure. Furthermore, the n=13 set includes five V-shaped metacentrics each of which can break into two rod-shaped acrocentrics. Evidently the centromere is here a super-gene through which the metacentrics split.
Staiger finds that these different chromosome types are broadly adapted to the environment. The 2n=26 form inhabits exposed sites and that with 2n=36 is found in sheltered ones, while those with intermediate chromosome numbers colonize habitats that are intermediate also relative to wave action.
This distinction, though related to micro-habitats, is effective also upon a larger scale. In the Roscoff region, Brittany, Staiger showed that the 2n=26 and 2n=36 forms exist in appropriate conditions and form heterozygous colonies in intermediate places. Yet 15 kilometres further east, where the coast is relatively uniform and exposed, the 2n=36 type is absent and both exposed and intermediate localities maintain a pure 2n=26 population. Thus regional differences influence chromosome-numbers in habitats of an intermediate.
The shell-structure of Purpura lapillus is affected both by the cytological situation and the environment. The shells are largest and have the thinnest walls in the two pure populations (2n=26 and 36); they are the shortest and thickest in the most heterozygous ones. This is not a function merely of the amount of material used, for shells of identical dimensions have thicker walls when chromosomally heterozygous than when homozygous. Moreover, owing to the correlation between structure and habitat just mentioned, the shell-thickness decreases as we pass from places with an intermediate to an extreme degree of exposure.
70 Ebling et al (1964) working near Loch Ine in Scotland, have demonstrated one of the adaptive components of this situation. The thickest shells cannot be broken open by the crabs Carcinus maenas and Portunus puber, which are important potential predators of Purpura lapillus. They can, however, crush the type with the thinner shells but are deterred by wave action from invading the exposed localities where that form of the Gastropod is principally found. Moreover, there is some evidence that the relatively large opening associated with the thinner type gives superior adhesion to the rocks, valuable in rough seas.
Staiger draws attention to a further point of general interest. That is to say, the density of populations in relatively exposed places with a rich food supply is greater in heterozygous than in homozygous communities. Moreover, body-size increases as we pass from intermediate to very exposed places in the heterozygous groups but in homozygous ones the maximum dimensions are reached in localities with intermediate exposure. Here then we have evidence of heterozygous advantage, which in addition is associated with greater adaptive versatility.