| Galton Institute Home Page | December 2003 Newsletter Contents | Newsletter Index |
In a stable environment, which persists unchanged for a considerable period of time, stabilising selection will act to keep the members of a population as near the optimum adaptive condition for that environment as possible, and to maintain a balanced gene complex. It will ‘tend to favour either organisms clustering around mean values for all characters’ by selecting the appropriate polygene combinations (generally heterozygotes, for these will have intermediate values), ‘or those in which extreme deviations for one trait are compensated by some form of deviation in another’ (Lerner, 1954) ; for instance, ‘an increase in height may have to be followed by an increase in the strength of the bone of the leg’ (Sheppard, 1975).
It happens that the first observation of natural selection acting in nature was made on a case of stabilising selection. In 1899, Bumpus reported on 136 sparrows, Passer domesticus, caught by a snow-storm in New England; 72 survived, the rest died. (Berry, 1990) ‘The average deviation from the mean for 8 out of the 9 body measurements taken was smaller for the survivors than for the birds which perished’. (Lerner, 1954) Table 1 shows some examples of observed stabilising selection favouring intermediate phenotypes, collected by Berry (1990), where the per cent strengths of selection have been estimated.
|
Table 1: Stabilising Selection for Intermediate Phenotypes |
|
|
Selection for |
% Selection Strength |
|
Coiling in snails (Clausilia laminata) |
8 |
|
Variation in over-wintering wasps (Vespa vulgaris) |
10 |
|
Size and hatchability in duck eggs |
10 |
|
Birth weight and survival in human babies |
2.7 |
|
Tooth size in Rattus rattus |
4 |
|
Non-metrical variation in skeletons of over-wintered Mus musculus |
up to 27 |
| Shell variability in Dog-whelks (Nucella lapillus) on shores exposed to different strengths of wave action |
up to 91 |
|
Colour types in Sphaeroma rugicauda |
over 50 |
Artificial selection is often applied for extreme expressions of isolated characters, and it then comes into collision with natural selection for balance and fitness. The extraordinary power of natural selection is well shown in a classical experiment of Mather and Harrison (1949), on the number of bristles on the abdomen of the fruit fly Drosophila melanogaster. This bristle number must be controlled by a polygenic system which produces ‘phenotypically balanced individuals’, and thus maintains fitness (Lerner, 1954; cf. Berry, 1990).
When Mather and Harrison ‘selected for a high number of bristles, progress was at first rapid, the number steadily increasing. But this was accompanied by a serious drop in reproductive capacity, and after about 20 generations artificial selection had to be suspended. The number of bristles then fell for a while, finally stabilising at a new level. Clearly heterozygosity had been restored, and new balanced systems of polygenes established. If selection was begun again … the number rose for some time, but finally reached a plateau, where it remained for many generations despite the persistence of artificial selection. … Artificial and natural selection were here in deadlock. But suddenly there was another rise, followed by another long plateau, and then another rise. A means had been found for increasing the bristle number without losing heterozygosity. This is explained on the basis of fairly strong linkage between the polygenes concerned; it is supposed that each small rise to a new plateau was made possible by crossing-over, which permitted new combinations of polygenes to be formed, some of them giving rise to higher bristle number even in heterozygous form.’ (Russell, 1959) In 1959, Scossiroli repeated the experiment (Wallace, 1968). But his stock must have had less genetic variability, for he only got a plateau.
Since the experimenters were only breeding from the small percentage of the population having the highest bristle number, natural selection could only cancel out the selection differential during a plateau by favouring, of this breeding group, the individuals with the lowest bristle number. Precisely this mechanism was found in artificial selection for increased shank length in chickens: the shorter the shanks out of the breeding group, the more offspring they produced (Lerner, 1954). The achievement of actual plateaux remains astonishing, and striking evidence of the power of natural selection. The original bristle experiment is shown in Figure 1, simplified from Mather and Harrison.
The power of Natural selection is thus well shown by the Mather-Harrison experiment; its phenomenal speed is equally well shown by events at Tring in Hertfordshire. (Sheppard, 1954; Ford, 1975) Kettlewell bred the scarlet tiger moth, Panaxia dominula, for ten generations (one a year in this species), selecting for striking changes in the colour pattern of the wings, probably mainly controlled by polygenes. In 1948, intending to leave England for some time, he released larvae from this altered stock in the grounds of Tring Museum, where the colony could not be contaminated by migration, for there are no populations of this species nearby, and in fact these moths rarely wander more than a few hundred yards from their habitat. In 1949 the larvae produced very few adult moths – clearly fitness had been reduced. But by 1951 the numbers had increased, and by 1953 the moths had almost totally reverted to their original colour pattern. Natural selection had thus reversed in five generations the alterations that had required ten generations of artificial selection. This is a case of restabilisation.
This moth is distasteful and uses warning colouration, which Kettlewell’s alterations had actually enhanced. It does not use its wing colour pattern in courtship. So evidently the pattern, like the fruit fly bristle number, is one effect of a polygenic system controlling variables important for balance and fitness, interfered with by Kettlewell’s alterations, and restored by natural selection between 1948 and 1953.