126
Μ. J. D. WHITE
vents the breakdown of the heterotic combination through crossingover.
The conditions we have outlined above are rather special ones. But
it seems likely that they have existed several dozen times in the phylogeny of the several hundred species of acridid grasshoppers that have
been studied, at least 9 times in the Australian eumastacid grasshopper
subfamily Morabinae and, from Table 28 in Patterson and Stone
(1952) perhaps 30 to 40 times in the history of the genus
Drosophifo.
Even when the chiasma distribution is most favorable, heterozygosity
for such fusions is likely to be mechanically inefficient and polymorphisms of this kind are hence especially likely to be short-lived, a consequence explaining their rarity at any particular time in evolutionary
history. Centric fusions involving the sex chromosomes seem to be in
a rather special category and will be discussed later (p. 133).
The processes of chromosomal evolution in the Lepidoptera seem to
differ in certain general respects from those which operate in Drosophifo
and the grasshoppers. The status of centromeres in lepidopteran
chromosomes is still rather uncertain, but it seems probable that diffuse
or multiple centromeres are generally present, as in the Homoptera and
Heteroptera (Suomalainen, 1953).
The great majority of lepidopteran species have haploid numbers between 28 and 31, the order as a whole showing considerable stability
in this respect. However, there are a number of genera in which fusions
or "fragmentations" (or both) seem to have occurred on a grand scale.
Some of the most extreme instances of fragmented karyotypes have
been found in the European butterfly genus Lysandra (Lorkovic, 1941;
de Lesse, 1953b, 1954). This genus belongs to the family Lycaenidae,
the overwhelming majority of whose members have η = 23 or η = 24;
it has in fact only recently been split off by systematists from the large
genus Lycaena, in which all the 13 species that have been investigated
show one or other of these numbers. Within the genus Lysandra, however, only one species shows η = 23. The others have haploid numbers of 45, 82, 84-85, 88-90, 124r-125, 131-150, while L. nivescens from
the Sierra Nevada of Spain has η = 190-191 (the latter being the highest chromosome number hitherto recorded in any animal species). The
chromosome numbers of some of these species may be variable within
or between local populations. There can be no doubt in this case that
the ancestral chromosome number was 23 or 24 and that the higher
numbers are derivative. In the ancestry of L. nivescens we must assume
that at least 166 structural rearrangements, each involving a gain of 2
telomeres, have occurred. And it seems likely that at least an equal
number of adaptive polymorphisms (each based on a structural re-
Μ. J. D. WHITE
vents the breakdown of the heterotic combination through crossingover.
The conditions we have outlined above are rather special ones. But
it seems likely that they have existed several dozen times in the phylogeny of the several hundred species of acridid grasshoppers that have
been studied, at least 9 times in the Australian eumastacid grasshopper
subfamily Morabinae and, from Table 28 in Patterson and Stone
(1952) perhaps 30 to 40 times in the history of the genus
Drosophifo.
Even when the chiasma distribution is most favorable, heterozygosity
for such fusions is likely to be mechanically inefficient and polymorphisms of this kind are hence especially likely to be short-lived, a consequence explaining their rarity at any particular time in evolutionary
history. Centric fusions involving the sex chromosomes seem to be in
a rather special category and will be discussed later (p. 133).
The processes of chromosomal evolution in the Lepidoptera seem to
differ in certain general respects from those which operate in Drosophifo
and the grasshoppers. The status of centromeres in lepidopteran
chromosomes is still rather uncertain, but it seems probable that diffuse
or multiple centromeres are generally present, as in the Homoptera and
Heteroptera (Suomalainen, 1953).
The great majority of lepidopteran species have haploid numbers between 28 and 31, the order as a whole showing considerable stability
in this respect. However, there are a number of genera in which fusions
or "fragmentations" (or both) seem to have occurred on a grand scale.
Some of the most extreme instances of fragmented karyotypes have
been found in the European butterfly genus Lysandra (Lorkovic, 1941;
de Lesse, 1953b, 1954). This genus belongs to the family Lycaenidae,
the overwhelming majority of whose members have η = 23 or η = 24;
it has in fact only recently been split off by systematists from the large
genus Lycaena, in which all the 13 species that have been investigated
show one or other of these numbers. Within the genus Lysandra, however, only one species shows η = 23. The others have haploid numbers of 45, 82, 84-85, 88-90, 124r-125, 131-150, while L. nivescens from
the Sierra Nevada of Spain has η = 190-191 (the latter being the highest chromosome number hitherto recorded in any animal species). The
chromosome numbers of some of these species may be variable within
or between local populations. There can be no doubt in this case that
the ancestral chromosome number was 23 or 24 and that the higher
numbers are derivative. In the ancestry of L. nivescens we must assume
that at least 166 structural rearrangements, each involving a gain of 2
telomeres, have occurred. And it seems likely that at least an equal
number of adaptive polymorphisms (each based on a structural re-