74
DAVID D. KECK
umbellifer complex. Particular interest pertains to S. crassicaulis itself,
for this is a widespread and polymorphic species composed of plants
belonging to a series of three polyploid levels, tetraploid, hexaploid, and
octoploid. Variation directly ascribable to polyploidy is of much less
significance than genie or even environmental variation, and so polyploidy of itself is not of taxonomic importance here. It appears that the
tetraploid S. crassicaulis arose through amphiploidy from two of the
diploid species or their ancestral counterparts. Hexaploid S. crassicaulis
presumably arose through the fusion of a reduced and an unreduced
gamete of the tetraploid, hence it is a sort of autoploid. The octoploid
forms of the species are thought to have arisen in three ways: two sorts
are thought to be autoploids out of geographically and somewhat morphologically different strains of the tetraploid, hence autoalloploids; the
third sort, which is morphologically somewhat different from the others,
presumably arose through hybridization between hexaploid S. crassicaulis
and a related diploid species.
Occasionally a polyploid species is found with characteristics that suggest either an autoploid or amphiploid origin. The common timothy,
Phleum pratense L., η = 21, is often considered to be an amphiploid derived from P. nodosum L., η = 7, χ P. alpinum L., η = 14, for it has
been artificially synthesized in this manner (Gregor and Sansome, 1930;
Nordenskiöld, 1937). But Nordenskiöld (1949) has also synthesized a
plant indistinguishable from P. pratense through autoploidy—by producing a hexaploid form of P. nodosum from the diploid. Chromosomal behavior can be utilized to defend either hypothesis as to the origin of
wild type P. pratense.
It also seems plausible that in this case we are
dealing with an autoalloploid, a hexaploid species that is autoploid with
respect to one genome, but alloploid in that it contains a different genome
(Stebbins, 1947).
A similar case of synthesis of a polyploid grass species through both
autoploidy and amphiploidy is treated by Nygren (1946). He doubled
the chromosomes of Calamagrostis
canescens (Weber) Roth, 2n = 28,
and obtained a 56-chromosome plant identical with the closely related
species, C. purpurea (Trin.) Trin. Nygren feels that this demonstrates
that some C. purpurea types are autoploid C. canescens forms. He also
synthesized C. purpurea, 2n = 56, /rom the cross C. epigeios (L.) Roth,
2n = 42, χ C. canescens
(Weber) Roth, 2n = 28, through an unreduced gamete of C. epigeios fusing with a reduced one from C. canescens. Incidentally, C. canescens
reproduces sexually (an amphimict
species), whereas C. purpurea reproduces by apomixis (an apomict
species) and has chromosome numbers ranging from 56 to 91. The importance of apomixis in taxonomy is discussed further below.
DAVID D. KECK
umbellifer complex. Particular interest pertains to S. crassicaulis itself,
for this is a widespread and polymorphic species composed of plants
belonging to a series of three polyploid levels, tetraploid, hexaploid, and
octoploid. Variation directly ascribable to polyploidy is of much less
significance than genie or even environmental variation, and so polyploidy of itself is not of taxonomic importance here. It appears that the
tetraploid S. crassicaulis arose through amphiploidy from two of the
diploid species or their ancestral counterparts. Hexaploid S. crassicaulis
presumably arose through the fusion of a reduced and an unreduced
gamete of the tetraploid, hence it is a sort of autoploid. The octoploid
forms of the species are thought to have arisen in three ways: two sorts
are thought to be autoploids out of geographically and somewhat morphologically different strains of the tetraploid, hence autoalloploids; the
third sort, which is morphologically somewhat different from the others,
presumably arose through hybridization between hexaploid S. crassicaulis
and a related diploid species.
Occasionally a polyploid species is found with characteristics that suggest either an autoploid or amphiploid origin. The common timothy,
Phleum pratense L., η = 21, is often considered to be an amphiploid derived from P. nodosum L., η = 7, χ P. alpinum L., η = 14, for it has
been artificially synthesized in this manner (Gregor and Sansome, 1930;
Nordenskiöld, 1937). But Nordenskiöld (1949) has also synthesized a
plant indistinguishable from P. pratense through autoploidy—by producing a hexaploid form of P. nodosum from the diploid. Chromosomal behavior can be utilized to defend either hypothesis as to the origin of
wild type P. pratense.
It also seems plausible that in this case we are
dealing with an autoalloploid, a hexaploid species that is autoploid with
respect to one genome, but alloploid in that it contains a different genome
(Stebbins, 1947).
A similar case of synthesis of a polyploid grass species through both
autoploidy and amphiploidy is treated by Nygren (1946). He doubled
the chromosomes of Calamagrostis
canescens (Weber) Roth, 2n = 28,
and obtained a 56-chromosome plant identical with the closely related
species, C. purpurea (Trin.) Trin. Nygren feels that this demonstrates
that some C. purpurea types are autoploid C. canescens forms. He also
synthesized C. purpurea, 2n = 56, /rom the cross C. epigeios (L.) Roth,
2n = 42, χ C. canescens
(Weber) Roth, 2n = 28, through an unreduced gamete of C. epigeios fusing with a reduced one from C. canescens. Incidentally, C. canescens
reproduces sexually (an amphimict
species), whereas C. purpurea reproduces by apomixis (an apomict
species) and has chromosome numbers ranging from 56 to 91. The importance of apomixis in taxonomy is discussed further below.