FORTY YEARS OF QENECOLOQY
207
of 24 h, after which the eight high and the eight low chaeta number
females were separated. From their progeny, the eight low and eight
high flies of each sex were selected and mated a t random, and the process was repeated in each generation. By the fourth generation of selection, almost all the high flies selected came from the progeny of high
females and almost all the low flies from the progeny of low females. By
the twelfth generation, the distribution curves did not even overlap. In
this experiment, mating was not enforced in any particular pattern, and
initially it must have been entirely at random. As the experiment progressed, reproductive isolation developed, limiting the formation of
hybrids between the high and low modes. In an earlier experiment
(Thoday and Boam, 1959) mating was enforced between low and high
lines so as to ensure the maximum amount of gene flow; significant
divergence ’ still occurred in consequence of the disruptive selection.
Thoday and Boam remark that the two sub-populations in this experiment are “in a formal sense, in the same relative situations as would be
two parts of a population in a mosaic environment consisting of two
distinct habitats arranged, for example, as a chequerboard. In such a
situation a population could in principle develop two different forms,
one adapted to each of the component environments even if there were
forced (50%) gene-flow between the two forms (ecotypes)”.
It will be recognized that with higher plants there could be several
forms of the basic disruptive-selection situation. For example, the dissimilar habitats could intermingle in the mosaic fashion envisaged by
Thoday, or they could interdigitate, or they could occupy distinct
provinces within the potential breeding area of the population. Again,
the selective pressures in the habitats could differ in intensity, so that a
wide range of phenotypes might survive in one and a very narrow range
in the other.
Taking an extreme case, it might be that the dispersal capacity of a
species was such that throughout an area where two -habitats, A and B,
occurred, seeding was essentially random. Selection in each generation
would then establish that phenotypes appropriate to the conditions of
habitat A survived there and were eliminated from B, and that those
adapted to B similarly persisted in B and were eliminated from A. The
breeding population would then be effectively dimorphic, although still
panmictic. The parallel with Thoday’s (1 962) experimental situation is
close though not exact, since the dimorphic Drosophila breeding population was derived by selecting the extremes from the total population,
whilst in the plant case the morphs are selected from the sample of seeds
happening to reach each of the two habitats. Nevertheless, it is evident
enough that if Thoday’s results can be extrapolated at all they mean
that the occupants of each habitat in successive generations should pro-
207
of 24 h, after which the eight high and the eight low chaeta number
females were separated. From their progeny, the eight low and eight
high flies of each sex were selected and mated a t random, and the process was repeated in each generation. By the fourth generation of selection, almost all the high flies selected came from the progeny of high
females and almost all the low flies from the progeny of low females. By
the twelfth generation, the distribution curves did not even overlap. In
this experiment, mating was not enforced in any particular pattern, and
initially it must have been entirely at random. As the experiment progressed, reproductive isolation developed, limiting the formation of
hybrids between the high and low modes. In an earlier experiment
(Thoday and Boam, 1959) mating was enforced between low and high
lines so as to ensure the maximum amount of gene flow; significant
divergence ’ still occurred in consequence of the disruptive selection.
Thoday and Boam remark that the two sub-populations in this experiment are “in a formal sense, in the same relative situations as would be
two parts of a population in a mosaic environment consisting of two
distinct habitats arranged, for example, as a chequerboard. In such a
situation a population could in principle develop two different forms,
one adapted to each of the component environments even if there were
forced (50%) gene-flow between the two forms (ecotypes)”.
It will be recognized that with higher plants there could be several
forms of the basic disruptive-selection situation. For example, the dissimilar habitats could intermingle in the mosaic fashion envisaged by
Thoday, or they could interdigitate, or they could occupy distinct
provinces within the potential breeding area of the population. Again,
the selective pressures in the habitats could differ in intensity, so that a
wide range of phenotypes might survive in one and a very narrow range
in the other.
Taking an extreme case, it might be that the dispersal capacity of a
species was such that throughout an area where two -habitats, A and B,
occurred, seeding was essentially random. Selection in each generation
would then establish that phenotypes appropriate to the conditions of
habitat A survived there and were eliminated from B, and that those
adapted to B similarly persisted in B and were eliminated from A. The
breeding population would then be effectively dimorphic, although still
panmictic. The parallel with Thoday’s (1 962) experimental situation is
close though not exact, since the dimorphic Drosophila breeding population was derived by selecting the extremes from the total population,
whilst in the plant case the morphs are selected from the sample of seeds
happening to reach each of the two habitats. Nevertheless, it is evident
enough that if Thoday’s results can be extrapolated at all they mean
that the occupants of each habitat in successive generations should pro-