Tải bản đầy đủ - 0trang
XI. Hybridization and Chromosome Relationships
FIG. 15. Scanning electron micrographs of a pollen grain of pearl millet. (a) A spheroidal grain;
note a single germ pore (potus) with cmfa around it. (b) A portion of pollen grain magnified to show
the porus with costa (C)around i t . [(a) X2600; (b) x5400l
PREM P. JAUHAR
Natural hybridization between related species followed by natural polyploidy has
in fact given rise to our most important food, fiber, oilseed, and fodder crops.
The catalytic effect of hybridization on evolution lies primarily in enlarging the
size of the gene pool that would facilitate a favorable response of a population to
a changing environment (Grant, 1963; Stebbins, 1969, 1974).
The study of chromosome pairing in hybrids has facilitated genome analyses
and, thus, helped in the elucidation of phylogenetic relationships between different taxa. Meiotic pairing in hybrids, of course, depends upon the nature and
extent of differentiation among the parental genomes. Moreover, different types
of genetic control of chromosome pairing can complicate the pairing patterns in
intergeneric, interspecific, and even intervarietal hybrids between geographically
diverse ecotypes, especially of polyploid species (see Jauhar, 1975c, 1977b). In
view of these restrictions, de Wet and Harlan (1972) have questioned the value of
chromosome pairing data in interpreting phylogenetic affinities. However, I do
feel that, in spite of the aforesaid limitations, chiasmate pairing has provided and
will continue to provide useful information on the nature of ploidy of a species
and on phylogenetic affinities between species. Chiasmate pairing is a specific
process generally confined to chromosomes (or segments) of corresponding
genetic similarities. Homologous or homoeologous segments are probably able
to recognize each other based on congruence of pairing sites (recognition is
probably based on similarity in nucleotide sequences). Information on chromosome pairing in hybrids coupled with that in their amphiploids should therefore
give a realistic picture of genomic relationships of the parental species.
In the genus Pennisetum, some interspecific, intergeneric, and even intertribal
(Penniserum X Oryza) hybrids have been reported. The interspecific hybrids
involved mostly pearl millet as one of the parents, and have helped in the study of
chromosome relationships between pearl millet and several other species of
A . INTRASPECIFIC HYBRIDS
There are several semiwild, annual, diploid races in the section Penicillaria
(Table I) with which the cultivated pearl millet is interfertile and essentially
forms a single, composite reproductive unit. So the hybridization between pearl
millet and other annual penicillarias will, in effect, be intraspecific hybridization.
Pennisetum typhoides is a polymorphic species. Its polymorphism could be
due primarily to its hybridization with the taxa in the section Penicillaria. The
wild, annual relatives of pearl millet have been treated as separate species by
Stapf and Hubbard (1933, 1934) and Clayton (1972), although the latter
suggested their merger into single species with pearl millet. They are interfertile
with pearl millet and have contributed to its genetic enrichment. Such a process
CYTOGENETICS OF PEARL MILLET
of hybridization and intefflow of genes is facilitated by the protogynous nature of
Chromosome studies by Krishnaswamy (1951), Thevenin (1952), and Veyret
(1957) have shown that the annual relatives of pearl millet-viz., P . ancylochaete, P . cinereum, P . echinurus, P . gambiense, P . leonis, P . maiwa, P .
nigritarum, and P . pycnostachyum'-are diploids with 2n = 14 chromosomes
like the cultivated pearl millet (see Table I). Moreover, the fact that the chromosome morphology of several of these taxa was similar to one another and also to
that of pearl millet (Veyret, 1957) further indicated their affinity with pearl
P . violaceurn and P . fallax are two of the important wild, annual relatives of
pearl millet. Genetic studies by Bilquez and Lecomte (1969) and by Brunken
(1977) have shown that these taxa form fully fertile hybrids with pearl millet and,
hence, are not reproductively isolated from it. They must obviously have 2 n =
14 chromosomes in order to form fertile hybrids with pearl millet.
The chromosomal, hybridization and genetic studies showed conclusively that
the wild or semiwild annual relatives of pearl millet are not sufficiently isolated
from it to deserve specific ranks. Brunken (1977) has therefore merged all annual
penicillarias with pearl millet (Pennisefum typhoides), which he calls P .
americanum. For the sake of convenience, he has subdivided the species into
three subspecies: (1) subsp. americanum, which includes all the cultivated races
of pearl millet; (2) subsp. monodii, which encompasses all the wild and semiwild
races of pearl millet; and (3) subsp. stenostachyum, which is morphologically
intermediate between subspecies americanum and monodii, and includes all
mimetic weeds associated with the cultivation of pearl millet.
Several hybrids between pearl millet and other species of Pennisetum have
been made by different workers. In some cases amphidiploids have also been
produced by colchicine-induced chromosome doubling of the interspecific hybrids or by natural meiotic nonreduction.
1 . P . typhoides x P . purpureum Hybrids
Interspecific hybrids between P. typhoides ( 2 n = 14) and P . purpureum (2n
= 4x = 28) are the most widely studied in the genus Pennisetum and probably
also in the entire tribe Paniceae. These species of the section Penicillaria are
reported to cross in nature to produce spontaneous hybrids (Stapf and Hubbard,
'These are all infraspecific categories within the species P . ryphoides
PREM P. JAUHAR
1934). They have also been hybridized by numerous workers using either one of
them as the female parent. However, using pearl millet as the female parent has
several advantages: (1) its protogynous nature generally eliminates the need for
emasculation; (2) seed shattering is absent or minimal; and (3) it is easier to
identify hybrids at seedling stage.
Burton (1944), working in the United States, was the first to make interspecific hybrids, which were later produced in India (Krishnaswamy and Raman,
1949, 1953a,b; Krishnaswamy, 1951), South Africa (Gildenhuys, 1950; Gildenhuys and Brix, 1958, 1964), Pakistan (Khan and Rahman, 1963), Australia
(Pritchard, 1971; Muldoon and Pearson, 1977), Sri Lanka (Dhanapala et a l . ,
1972), Nigeria (Aken’Ova and Chheda, 1973), and other countries. The main
objective of hybridizing these species was to produce a high-quality, highyielding, perennial fodder plant that would inherit pearl millet’s forage quality,
nonshattering nature, and capacity to establish readily, and also have the perennial, aggressive nature and rust resistance of napier grass. In South Africa, the
main objective was to breed a large-seeded perennial for use in ley farming
The hybrids produced in different countries are generally high-yielding and
more acceptable as fodder plants than napier grass. They exhibit considerable
heterosis for both fodder yield and quality (Burton, 1944; Krishnaswamy and
Raman, 1949; Patil, 1963; Khan and Rahman, 1963; Hussain et al., 1968;
Pritchard, 1971; Aken’Ova et a f . , 1974; Gupta, 1974; Muldoon and Pearson,
1977). Powell and Burton (1966b) described a commercial method of producing
interspecific hybrids by using a male-sterile line of pearl millet (Tift 23A) as the
female parent; the hybrid thus produced was described as the highest-yielding
forage millet grown in the United States. However, the hybrids’ complete seed
sterility (they can be propagated only by vegetative means) has restricted their
adoption on a large scale. While these hybrids are widely grown in some countries, particularly in the Indian subcontinent, they are being grown in trials in
several other countries (see Muldoon and Pearson, 1979; see also Section
XI ,B,2 ,e) .
a. Gross Morphology. With respect to several vegetative characters such as
panicle morphology, internode length, leaf and ligule size, the hybrid is either
intermediate between the parents or more often approaches the purpureum parent. This is to be expected in view of the greater contribution of genetic material
from the purpureum parent. However, depending on the genotype of the parental
species used, there is a considerable variation in expression of heterosis for
different vegetative characters like height, stem thickness, tillering, and leaf size.
b. Chromosome Pairing. The hybrid is a triploid with 2n = 3x = 21
chromosomes (Fig. 16a), 7 contributed by diploid (2x) typhoides and 14 by
tetraploid (4x) purpureum. As discussed in Section I1,1, pearl millet has very
large chromosomes, larger than those known for any other species of Pen-
CYTOGENETICS OF PEARL MILLET
nisetum, except P . ramosum. Burton (1944) could identify the 7 large chromosomes of pearl millet ir. the interspecific hybrids, but he did not study chromosome pairing. Krishnaswamy (1951) and Krishnaswamy and Raman (1953a,
1954) studied chromosome pairing in the triploid hybrids. On the basis of the
formation of 711 + 7, in most of the cells and the absence of trivalents, these
authors concluded that the genome of typhoides was homologous to one of the
genomes of purpureum. Khan and Rahman (1963), Ramulu (1968), Sethi et al.
(1970), and Rangaswamy (1972) made essentially similar observations on
chromosome pairing in the hybrids and reached the same conclusions regarding
the genomic makeup of the parental species.
Raman (196% reviewing the earlier work of himself, Krishnaswamy, and
co-workers, designated the genomic constitution of P . ryphoides as AA, and that
of P . purpureum as A'A'BB. The formation of 711in the triploid hybrid (AA'B)
was attributed to synapsis between the A genome of typhoides and A' of purpureum. It was evident, however, that A and A' were not completely homologous.
c . Analysis of Inter- and Intragenomal Chromosome Pairing. In the hybrids
studied by Jauhar (1968), the easily recognizable size differences of the parental
chromosomes (Fig. 16a-c) permitted a detailed analysis of inter- and intragenomal chromosome pairing. A range of 0-9 bivalents was observed, the
mean per cell being 5.3. Whereas the majority of bivalents were formed between
A and A' genomes and, thus, were clearly heteromorphic (Fig. 16b,c), some
bivalents resulted from intragenomal pairing. Up to five heteromorphic AA'
bivalents were observed per cell, which suggested that the two genomes are
evolutionarily related, and that they probably arose from a common progenitor
with x = 5 chromosomes.
All three genomes-A, A', and B-showed
chromosome pairing. Such bivalents were almost homomorphic and hence easily
distinguishable from the heteromorphic, intergenomal (A-A') bivalents described above. The bivalents formed by the A genome chromosomes were the
largest (Fig. 16b), whereas those of A' and B genomes were, respectively,
intermediate and the smallest in size. The intragenomal pairing appeared to be
limited to a maximum of two bivalents, further suggesting x = 5 as the phyletically basic number from which, probably, the genomes with x = 7 were subsequently derived. Thus, it was inferred that x = 5 is the original basic number of
the genus Pennisetum and that P . typhoides is a secondarily balanced species
(Jauhar, 1968). Later studies by Jauhar (1970b), Minocha and M . Singh (1971b),
and Minocha and A. Singh (1971a) provided corroborative evidence favoring
these conclusions. It may be pointed out that as early as 1951, Krishnaswamy
had suggested that intragenomal pairing occurred within the B genome.
On the basis of pachytene pairing in the hybrid, Pantulu (1967b) reported that
the chromosomes 1-5 of typhoides were homologous with chromosomes 1-5 of
PREM P. JAUHAR
Flc. 16. Meiotic stages in the Penniserurn ryphoides X P . purpureum ( 2 n = 3 x = 21). Note
clear size differences of parental chromosomes. [Small arrow, wphoides univalents; medium arrows,
intragenomal bivalents within ryphoides complement; thick arrows, intergenomal bivalents formed
by A genome of typhoides and A ' of purpureurn.] (a) Metaphase I showing 21 chromosomes, 7 large
ryphoides chromosomes ( A genome) and 14 small purpureum chromosomes (A'B genomes). (b)
7, that comprise 1 heteromorphic intergenomal bivalent (thick m o w ) , 2 large
Metaphase I with 7,,
CYTOGENETICS OF PEARL MILLET
purpureum, and that chromosomes 6 and 7 of typhoides were homologous with
chromosomes 8 and 14 of purpureum, respectively. However, in view of the
markedly larger chromosomes of typhoides, a part of its chromosome should
remain unpaired with the corresponding purpureum chromosome. Consequently,
terminal forks, terminal unpaired regions, or intercalary loops should be seen at
pachytene. Pantulu (1967b) did not, however, report any such structures, as his
drawings of different bivalents show almost perfect pairing.
2 . P . typhoides x P . purpureum Amphidiploids
a . Chromosome Pairing and Fertility. Krishnaswamy and Raman (1949)
produced amphidiploids by treating the P. typhoides X P . purpureum hybrid
seedlings with 0.4% colchicine. Chromosome doubling largely restored the
fertility in the synthetic amphidiploids (AAA'A'BB; 2n = 6x = 42). During
meiosis they generally formed 2 I,,, multivalents being either absent or infrequent
(Krishnaswamy, 1951; Krishnaswamy and Raman, 1954; Ramulu, 1968, 1971;
Jauhar and Singh, 1969b; and Jauhar, unpublished results). Some univalents
observed at metaphase resulted from precocious separation of bivalents and
caused disjunctional abnormalities.
In view of the formation of 711 + 71 in the triploid hybrid (AA'B), some
quadrivalents or trivalents should be expected in the derived amphidiploids, but
they are observed rarely, if at all. This is probably due to preferential pairing
between the A-A, A'-A', and B-B genome chromosomes, resulting in 21,1.
Although the A and A' genomes are somewhat differentiated, the corresponding
chromosomes of these genomes can pair in the absence of their own homologous
partners. On chromosome doubling, however, bivalent formation is probably
brought about by strong preferential pairing. However, the possibility of some
sort of genetic control of pairing cannot be ruled out. The diploidizing genes in a
double dose probably bring about diploid-like (genetically enforced preferential)
pairing in the allohexapioid. [Such diploidizing genes that are effective only in a
double dose are known in polyploid species of Festuca (Jauhar, 1975a,c).] Thus,
the synthetic amphidiploid behaves meiotically like a typical allohexaploid.
b. Chromosomal Instability. Gildenhuys and Brix (1961) also produced an
amphidiploid by colchicine treatment of the cuttings of triploid hybrid. Although
largely fertile, the amphidiploid showed marked instability in somatic chromointrahaploid bivalents within typhoides complement (medium m o w ) , and 4 intragenomal bivalents
within A ' and B genomes; 2 large univalents belong to A genome and the remaining 5 univalents
belong to A' and B genomes. (c) Metaphase I with 6,, + 9,. Note clearly heteromorphic bivalents
(thick arrows) and distinct size differences among univalents. (d) A cell at anaphase with 22 chromosomes showing 2 typhoides and 5 purpureum chromosomes going to one pole, 3 fyphoides and 9
purpureum chromosomes going to other pole, and 3 Wphoides chromosomes lagging. The ryphoides
chromosomes are marked with m o w s . [(a-c) x ca. 1800; (d) x ca. 13501
PREM P. JAUHAR
some number within the plant; its number was in the range 2 n = 36-49, with 2 n
= 42 occurring most frequently (Gildenhuys and Brix, 1964). The backcross
progeny obtained from the cross 2x P. typhoides X 6x amphidiploid also
showed intraplant variation in chromosome number. Gildenhuys and Brix ( 1 964)
concluded that this intraplant numerical mosaicism was under genetic control and
that the genetic determinants expressed themselves only when present in a double
or higher dose. Thus, it appears that these genes were hemizygous ineffective.
Conversely, Krishnaswamy and Raman (1956) and Ramulu (1968, 1971) did
not report any intraplant or even interplant numerical mosaicism in the amphidiploids or their derivatives. Because of the formation of univalents, however, some
variation in chromosome numbers in the progenies of amphidiploids is inevitable
and may be detected if a large population is scored.
c . Phenotypic Manifestation of Different Genomes. The amphidiploids are
vigorous in growth and show gigantic features typical of several polyploids.
Generally, they have thicker stems, broader leaves, and larger panicles compared
to the parental triploids. They show greater morphological resemblance to P.
purpureum than to P . typhoides. This is expected in view of the greater genomic
contribution from the tetraploid purpureum. Contrarily, if it is considered that A
genome of typhoides and A‘ of purpureum are similar in genic content, then the
amphidiploids have, in effect, four A genomes and should resemble the
typhoides parent more closely. However, their greater resemblance to purpureum would show that either the A and A’ genomes are sufficiently differentiated and have different phenotypic expressions, or else it is more likely
that the B genome dominates the A as well as A‘ genomes and masks their
That the B genome is indeed “dominant” is borne out by the studies of
Krishnaswamy and Raman (1949, 1953a, 1954, 1956), Raman and Krishnaswami (1961), Raman et al. (1963), and Raman and Nair (1964). All these
workers have demonstrated that even when the ratio of A to B genomes is altered
from 2 : 1 to 5 : 1, the phenotypic manifestation of the B genome is noticeably
greater than that of A genomes combined. In an amphiploid derivative with
AAAAA’B constitution (Raman and Krishnaswami, 1961), for example, there
was only one dose of B genome, but the characters of the wild parent, P.
purpureum, were still expressed. This indicated that one dose of B genome was
dominant (or perhaps epistatic) over five doses of the A genome. Studying the
morphology of selfed progenies of the amphidiploids by metroglyph analysis,
Ramulu and Ponnaiya (1967) and Ramulu (1971) also found a distinct skewedness towards P . purpureum in the expression of several morphological features.
d . Amphidiploid Derivatives. Synthetic amphidiploids have been successfully backcrossed to pearl millet to produce derivatives with different genomic
constitutions. Thus, Krishnaswamy and Raman (1956) found that hybrids could
be easily secured whichever way the cross was attempted, e.g., 2x x 4x, 2x x
CYTOGENETICS OF PEARL MILLET
6x, 4x x 2x,or 6x x 2x. All nine tetraploids produced from the cross 2n P.
typhoides x 6x amphidiploid had 2n = 28 chromosomes, showing that the
amphidiploids generally formed 2 1 -chromosome gametes or else such gametes
were at a competitive advantage over the aneuploid gametes. This is in sharp
contrast to the observations of Gildenhuys and Brix (1964), who reported that the
compatibility of the cross improved when subhaploid pollen of the hexaploid
fused with the haploid egg cell of P . typhoides.
Gildenhuys and Brix ( 1 964) also found a high degree of incompatibility when
amphidiploid was used as pollen parent in backcrosses to P . typhoides. Of the 31
offsprings thus obtained, only 4 had the expected number of 2n = 28 or less. The
remainder had 2n = 35 or thereabout, and hence arose from the unreduced egg
cells of P . typhoides. It appeared that the pollen from the hexaploid was more
compatible with diploid rather than haploid eggs of typhoides. Conversely, the
compatibility also seemed to be enhanced when the pollen from hexaploid contained less than the haploid set of chromosomes (Gildenhuys and Brix, 1964).
From the standpoint of plant breeding, however, this is not a welcome situation.
It was further inferred from these results as well as from the cross [2x P.
typhoides x 6x (P.typhoides x P . purpureum)] x 2x P . typhoides, that
incompatibility probably resided in the hybrid embryo itself and that its normal
development (e.g., the success of the cross) was not dependent on the normal
endosperm acting as a nurse tissue (Gildenhuys and Brix, 1964, 1965).
Chromosome pairing was studied in several other amphiploid derivatives, with
various genomic constitutions, obtained after a series of backcrossings of 6x
amphidiploid to 2x and 4x P. typhoides, followed by selfing and further crossing
of the derivatives (Raman and Krishnaswami, 1960, 1961; Raman et af., 1963;
Nair et al., 1964; Raman and Nair, 1964). As expected, these derivatives formed
different frequencies of multivalents, bivalents, and univalents, depending upon
their genomic composition. They were sterile by varying degrees.
From the studies on triploid hybrid (AA'B), the amphidiploids (AAA'A'BB),
autoallotetraploids (AAA'B) and numerous other amphiploid derivatives with
different genomic constitutions, the following conclusions can be made:
I . When present in duplicate, the genomes A, A', and B maintain their meiotic
integrity, forming mostly bivalents between homologous partners; this is probably due to genetically enforced preferential pairing.
2. When present in a single dose or in more than two doses, there is intergenomal pairing between A and A'; however, A-B and A'-B synapsis is
seemingly absent, showing thereby that A and A' are largely homologous to each
other, whereas they both are nonhomologous to the B genome.
3. There is some amount of intragenomal pairing probably limited to a
maximum of two bivalents within each of the three genomes.
These studies have confirmed the allotetraploid nature of P . purpureum. It is a
PREM P. JAUHAR
genomic allotetraploid (A’A’BB) with one genome largely homologous to the
typhoides (AA) genome at least in terms of its pairing behavior. The donor of B
genome may be one of the diploid members of the section Penicillaria, which may
exist somewhere in Africa or is probably extinct.
e . Selection of Superior Forage Hybrids. As described above, several
superior triploid hybrids between pearl millet and napier grass have been produced in different countries. However, they have not been adopted on a wide
scale primarily because they are completely sterile and can be propagated only by
vegetative means. From the point of view of easy distribution to farmers,
superior, fertile amphidiploid hybrids or derivatives would need to be developed,
so that seed can be supplied to farmers. Since the colchicine-induced amphidiploids are largely regular meiotically and their progenies show a wide range of
pollen and seed fertility, it may be possible to select fertile, superior forage
amphiploids which produce large amounts of good seed.
f. Possible Synthesis of Perennial Pearl Millet. From the point of view of
incorporating the desirable features of P . purpureum into P . vphoides, the
triploid plant obtained by Raman and Krishnaswami (1960) is interesting. This
plant was derived by crossing 2x typhoides (AA) with an autoallotetraploid
(AAA’B). The derived triploid had four nucleolar chromosomes, and the pattern
of pairing showed that it had mostly chromosomes of A and A’ genomes and
probably none of the B genome. It is interesting that most, if not all, B genome
chromosomes were selectively eliminated. Raman and Krishnaswami (1960)
suggested that from such triploid plants it might be possible to secure fertile
allodiploids with A and A’ genomes that would probably combine the desirable
features of ryphoides and purpureum.
It was observed earlier at the University of Hawaii Agricultural Experiment
Station (Anonymous, 1947) that the 34 F , hybrids, obtained by crossing a selfed
line of pearl millet (2n = 14) with an East African strain of napier grass (2n =
28), included plants with 2 n = 28,21, 18, 16, and 14. The occurrence of plants
with 2 n = 14 suggests the possibility of selecting the desired “allodiploid”
plants. In this way perhaps a perennial pearl millet can be produced. The selfed
pearl millet line used in the above cross was probably desynaptic and produced
imbalanced gametes and, hence, aneuploid hybrids. If a desynaptic purpureum
line can be used as a male parent and crossed with a diploid typhoides with
desirable features, different plants with a whole array of chromosome numbers
can be obtained to exercise selection for the desired 2 n = 14, meiotically regular, fertile plants.
In successive backcrosses of the 6x amphiploids to P . typhoides, Gildenhuys
and Brix (1969) observed a selective elimination of imbalanced gametes and
zygotes, particularly those with chromosomes of the B genome, and to a lesser
extent, those of the A‘ genome (of purpureum). After three generations of
backcrossing, all plants expressed the annual pearl millet habit. However, Gildenhuys and Brix (1969) could not combine the desirable features of typhoides
CYTOGENETICS OF PEARL MILLET
(viz., fertility, date of flowering, and grain size) with the perennial habit of
purpureurn. If further attempts are made using gamma-ray treatments also, it
may be possible to incorporate the desirable segments from purpureum into the
If a perennial pearl millet is indeed produced through the techniques described
above, it would be useful in the drought-stricken, semiarid regions of Africa,
tropical India, and other tropical and subtropical regions.
3 . P . typhoides x P . squamulatum
These species have been hybridized with a view to evolving a grass combining
the forage quality of pearl millet with the frost resistance and perenniality of P .
squamulatum (2n = 6x = 54). Patil et al. (1961) and the present author (Jauhar,
unpublished results) successfully obtained the hybrids. Taking advantage of the
protogynous nature of pearl millet, we dusted profusely the pollen of
squamulatum on a number of typhoides ears that had freshly emerged, receptive
stigmas. From the progeny, Patil et al. (1961) picked up a single hybrid that had
2n = 41 chromosomes. The hybrid obviously arose from the union of an unreduced 14-chromosome egg of pearl millet with a haploid 27-chromosome male
gamete of squamulatum. Morphologically, the hybrid resembled the squamulatum
parent very closely although it had the leafiness of pearl millet.
Patil et al. (1961) observed mostly 1611 9, in the hybrid. It appeared to the
authors that seven bivalents (which must have been much larger than the rest)
were formed by the typhoides complement, whereas the remaining nine bivalents
(the small ones) and nine univalents resulted from pairing within the
squamulatum complement. The small bivalents in the hybrid could have formed
as a result of intergenomal pairing between the three genomes of squamulatum
and some by intragenomal pairing.
4 . TrispeciJic Hybrids:
( P . typhoides x P . purpureum) x P . squamulatum
Several hybrids involving three distinct species have been made in grasses,
e.g., Lolium-Festuca complex (Jauhar, 1975b, 1976). In the genus Pennisetum,
there is one report of a trispecific hybrid obtained by crossing ( P . typhoides X P .
purpureum) hybrid with P . squamulatum (Rangaswamy and Ponnaiya, 1963).
The two hybrid plants thus obtained resembled the squamulutum parent in several features, including the panicle and spikelet characters. They also had penicillate anther tips, a typical character of typhoides (see Fig. 14) and purpureum.
Rangaswamy and Ponnaiya (1963) found that the hybrid had 2n = 48 chromosomes, which showed that it had arisen from an unreduced 21-chromosome egg
of the female parent (typhoides X purpureum hybrid) fertilized by the normal
27-chromosome male gamete from squamulatum. The formation of quadrivalents and pentavalents at meiosis suggested relationship of the genomes of the