206
J. HESLOP-HARRISON
other hand, the effectiveness of wind-pollination and seed dispersal together may be such as to reduce substantially the isolating effect of local
discontinuities, even when the distribution pattern is as fragmented, as
in a coastal species like Plantago maritima (Gregor, 1946a).
Any form of spatial isolation which is effective in restricting gene flow
between two populations of a species will act to facilitate the independent response of the populations to local selective influences. Isolation
must therefore always favour the establishment of discontinuities in the
variational range of adaptive characters where the separate populations
occupy dissimilar habitats. This will be true whatever the intrinsic
properties of the genetic system, provided only that there is some
release of genetic variation for selection to act upon.
By establishing that the distributional pattern will be one of numerous more or less isolated populations, the ecological predilections of
species with the “colonial” type of distribution mentioned above themselves create the conditions for ready fractionation into discrete ecotypes adapted for the minor modulations of the characteristic habitat.
Even in the case of outbreeding species with tendencies t o form continuous stable communities and to show clinal variation, the occurrence
of a topographical barrier imposing some degree of reproductive isolation will favour the appearance of variational discontinuity, producing
regional ecological races, or “stepped” ecoclines (Gregor, 1944).
These propositions concerning the role of spatial isolation as a determinant of discontinuity can, of course, be traced back to Darwin and
Wagner, and they form a well established part of genecological lore.
There is, however, the important question as to whether or not spatial
isolation is a sine qua non for the development of variational discontinuity. The view of many contemporary evolutionists is that it is, and
that divergence in geographical isolation must necessarily precede the
evolution of the genetically determined bars to crossing that mark the
final step of speciation. This opinion has been especially canvassed by
animal systematists (e.g. Mayr, 1942, 1947, 1959; Cain, 1954). There is
now, however, excellent evidence from the experimental work of Thoday
(1959), Thoday and Boam (1959) and Thoday and Gibson (1962) to
show that disruptive selection as defined by Mather (1953, 1955) can
produce what is effectively a racial divergence in the face of a very high
level of intercrossing.
Thoday’s experiments were carried out with Drosophila melanogaster,
but the results have immediate interest for the plant situation. One
experiment (Thoday and Gibson, 1962) began with eighty flies of each
sex from a wild strain. From each sex the eight flies with the highest
sternopleural chaeta number and the eight with the lowest were selected.
The thirty-two flies were permitted to mate at random during a period
J. HESLOP-HARRISON
other hand, the effectiveness of wind-pollination and seed dispersal together may be such as to reduce substantially the isolating effect of local
discontinuities, even when the distribution pattern is as fragmented, as
in a coastal species like Plantago maritima (Gregor, 1946a).
Any form of spatial isolation which is effective in restricting gene flow
between two populations of a species will act to facilitate the independent response of the populations to local selective influences. Isolation
must therefore always favour the establishment of discontinuities in the
variational range of adaptive characters where the separate populations
occupy dissimilar habitats. This will be true whatever the intrinsic
properties of the genetic system, provided only that there is some
release of genetic variation for selection to act upon.
By establishing that the distributional pattern will be one of numerous more or less isolated populations, the ecological predilections of
species with the “colonial” type of distribution mentioned above themselves create the conditions for ready fractionation into discrete ecotypes adapted for the minor modulations of the characteristic habitat.
Even in the case of outbreeding species with tendencies t o form continuous stable communities and to show clinal variation, the occurrence
of a topographical barrier imposing some degree of reproductive isolation will favour the appearance of variational discontinuity, producing
regional ecological races, or “stepped” ecoclines (Gregor, 1944).
These propositions concerning the role of spatial isolation as a determinant of discontinuity can, of course, be traced back to Darwin and
Wagner, and they form a well established part of genecological lore.
There is, however, the important question as to whether or not spatial
isolation is a sine qua non for the development of variational discontinuity. The view of many contemporary evolutionists is that it is, and
that divergence in geographical isolation must necessarily precede the
evolution of the genetically determined bars to crossing that mark the
final step of speciation. This opinion has been especially canvassed by
animal systematists (e.g. Mayr, 1942, 1947, 1959; Cain, 1954). There is
now, however, excellent evidence from the experimental work of Thoday
(1959), Thoday and Boam (1959) and Thoday and Gibson (1962) to
show that disruptive selection as defined by Mather (1953, 1955) can
produce what is effectively a racial divergence in the face of a very high
level of intercrossing.
Thoday’s experiments were carried out with Drosophila melanogaster,
but the results have immediate interest for the plant situation. One
experiment (Thoday and Gibson, 1962) began with eighty flies of each
sex from a wild strain. From each sex the eight flies with the highest
sternopleural chaeta number and the eight with the lowest were selected.
The thirty-two flies were permitted to mate at random during a period