Tải bản đầy đủ - 0trang
V. Triticale Improvement at CIMMYT
THE DEVELOPMENT OF TRITICALE
by utilizing three entirely different climatic zones at which the populations
were grown and selected. In the winter, the triticale nursery is grown at
Centro de Investigaciones Agricolas del Noroeste (CIANO), State of
Sonora, at 28O N latitude, 35 meters elevation. The summer nursery in
Mexico is grown in the Toluca Valley at 1 8 S 0 N latitude and 2600 meters
elevation. The University of Manitoba summer nursery is grown in the
Winnipeg area at 50° N and 230 meters elevation. Thus not only did the
program expand in area and number of workers involved, but the number
of generations grown per year doubled, the number of environments
tripled, and for the first time triticale breeding was introduced to tropical
regions requiring daylength insensitivity similar to that of the wheats which
are adapted to regions between 30° N and 30° S latitudes. The alternation
of generations between CIANO and Toluca permitted screening for strains
capable of performing well at two widely different environmental conditions. The disease infestations differed greatly at the two locations. This
has greatIy enhanced the possibilities of obtaining selections having wider
adaptation. A further advantage was gained by having access to large, diversified material in aggressive durum and bread wheat programs to which
triticale crosses could be made and from which new primary triticales could
Preliminary work on triticale was started in 1963, when some triticale
strains were included among wheat populations obtained from Dr. J. A.
Rupert in Chile. Ingenieros Ricardo Rodriguez and Marco Quiiiones because of scientific curiosity made a number of crosses between these triticales and several Mexican dwarf wheats (Quiiiones, 1967). The triticale
lines originated from the University of Manitoba. The hybrids and selected
plants were brought to the Toluca nursery (State of Mexico) for observation. The need to overcome daylength sensitivity, reduce plant height, and
improve resistance to stripe rust became apparent immediately. Some degree of daylength insensitivity was recovered from crosses between triticales, but this was insufficient for crop production in Mexico. Improved
daylength insensitivity and disease resistance, were recovered in later generations from triticale X bread wheat crosses made at this time.
By 1967, a number of strains with enough resistance to disease, and
insensitivity to daylength had been developed to be included in replicated
yield tests. Results from tests at CIANO and Toluca indicated that the
triticale strains produced about one half as much grain as the best wheat
cultivars under similar conditions. The triticale strains were tall, late maturing, and at least as vigorous in the production of total plant material as
the best wheat cultivars. The depressed grain yields were attributed to the
high incidence of sterility and severe endosperm shriveling.
F. J. ZILLINSKY
1. Improving Fertility
In 1968 a very intensive selection effort was devoted to finding plants
having better fertility. A few plants with improved fertility were found in
an F, population of a cross between two hexaploid triticales. The average
percentage of seed set of two of the original lines was about 6 % below
that of adapted bread wheat strains, and 15% above the best original hexaploid triticales. These few plants eventually provided a major contribution
to triticale improvement. Among the characters associated with these selections, which were later identified as Armadillo strains (Zillinsky and Borlaug, 1971), were high fertility, improved test weight, better grain yield,
insensitivity to daylength, one gene for dwarfness, early maturity, and good
nutritional quality. Each of the factors were found to be heritable and could
be easily transmitted to its progeny. Furthermore the Armadillo strains
were generally more cross compatible with bread wheat, durum wheat,
and rye than were the normal hexaploid strains in the program. This improved compatibility might be attributed to a higher proportion of viable
pollen when used as the pollen parent in crosses. A similar improvement
in the production of F, hybrids was observed when the strains were used
as the female parent.
Investigations on the origin of the unusual characteristics of the Armadillo strains (Fig. 4) revealed that the majority of the characteristics, such
as dwarfing, disease resistance, earliness, erect juvenile growth habit, short
spike, and smaller plumper kernels, must have been introduced from a
Mexican bread wheat having a NORIN 10 dwarfing gene. The bread wheat
is believed to have been introduced via spontaneous outcrossing on the
F, hybrid of cross X308, since the original cross, X308, from which the
Armadillo strains were selected, combined secondary hexaploid triticale
parents having no bread wheat in the progenitors. The outcrossed hybrid
was subsequently pollinated with hexaploid triticale pollen from neighboring plants in the triticale nursery. A verification that a bread wheat progenitor was involved in the origin of Armadillo was obtained in 1973 when
a D chromosome was found to be substituted for one of the rye chromosomes (Gustafson and Zillinsky, 1973; Gregory, 1973; Merker, 1973b).
Although the rye genome in Armadillo appears to have lost one pair
of chromosomes by a substitution, the total genotype was considerably improved by the modification. Whether this was due to the deletion of that
particular rye chromosome or to the favorable effects of the D genome chromosome has not yet been determined.
The Armadillo strains were used frequently as parents in crosses to other
hexaploid triticales both primary and secondary forms, to bread and durum
wheats, and to primary octoploid triticales during the following generations.
THE DEVELOPMENT OF TRITICALE
FIG.4. The plant type of the fertile selection Armadillo.
Selections with fertility approaching that of the Armadillo parent were obtained among the segregating populations. By 1970 practically all the material in the CIMMYT triticale program originated from crosses having Armadillo as a progenitor.
Susceptibility to lodging was a common problem encountered by many
of the early investigators, including Muntzing, Kiss, and Sfinchez-Monge.
The problem was intensified under Mexican conditions owing to the tendency of long day-sensitive material to grow taller under short-day conditions. Even the single factor for dwarfing possessed by the Armadillo
strains was not sufficient to prevent lodging since the increase in fertility
and grain density increased the weight of the mature spike (Fig. 5 ) .
Attempts to improve lodging resistance included increasing straw thickness and incorporating more dwarfing genes from wheat. New primary
amphiploids, both hexaploid and octoploid, were produced using dwarf
durum and bread wheats. These were used as parents in crosses to
F. J. ZILLINSKY
FIG.5. Dwarfing in triticale: differences in plant height.
hexaploid triticale. Early attempts to incorporate more dwarfing genes
from bread and durum wheats having NORIN 10 dwarfing were discouraging. It was very difficult to maintain fertility among the dwarf selections.
The grain quality tended to deteriorate conspicuously. Similar problems
were encountered in the early stage of the wheat breeding program in
Mexico when NORIN 10 was used as a dwarfing source. Borlaug 1968;
Zillinsky and Borlaug, 1971 ) pointed out that only semidwarfs (singlegene dwarfs) with good fertility and acceptable grain type could be
isolated from crosses between NORIN 10 X tall wheats. All double dwarf
segregates were highly sterile and possessed very shriveled grain. Subsequent
recrossing and selection for fertility and grain plumpness resulted in the development of excellent double dwarfs, such as SONOM 64, INIA 66, possessing complete fertility and excellent grain type. Kiss (1968) reported
similar difficulty with sterility and grain shriveling when using the NORIN
wheat dwarfing sources. He subsequently used TOM THUMB with more
Crosses between Armadillo strains and stiff-strawed, normal-height triticales resulted in only moderate improvements in lodging resistance. It
became obvious that if grain yields competitive with the Mexican dwarf
wheat were to be achieved, the straw length of triticale had to be reduced.
Since all the triticale germplasm possessed genomes of tall ryes, a major
obstacle to expression of the dwarfing characteristic in triticale was the
THE DEVELOPMENT OF TRITICALE
tall genotype of rye. An obvious solution was to replace these with genes
from dwarf ryes. A search for dwarf ryes among collections of spring ryes
resulted in the discovery of a single heterozygous dwarf plant in a rye population received from Dr. Darrell Morey of The Coastal Plains Experiment
Station, Tifton, Georgia (Zillinsky and Borlaug 1971 ) . The dwarf segregates among the progeny of this plant were identified as “Snoopy” selections. Unfortunately the original plant was susceptible to several diseases
(stripe rust, bacterial stripe, and scab) and had some other unfavorable
agronomic characteristics. It was necessary to improve the phenotype by
crossing to selected tall strains before crossing to triticale. Dwarf segregates
from crosses between Armadillo x Snoopy rye had sterility and seed
shriveling problems similar to those from the Armadillo X dwarf wheat.
Two-gene dwarf hexaploids were eventually obtained in 1972 which
were equal in fertility to Armadillo (Zillinsky and Lopez, 1973). These
originated from two sources: (a) hexaploid triticale x bread wheat, and
(b) octoploid triticale x hexaploid triticale. The F, hybrids from both
sources were equal to the triticale parent in height. The hybrids having
a bread wheat parent were much more sterile than those from octoploid x hexaploid crosses. It was necessary to overcome the sterility by
growing the F, plants in rows alternating with normal fertile hexaploid
triticale as a source of viable pollen for two generations.
Triticale was first released for commercial production in Hungary in
1968. Even today only a few countries are growing limited acreages commercially. Information on diseases is rather scarce. Wherever the crop is
grown, disease symptoms appear, apparently caused by plant pathogens
which parasitize wheat and rye species. They have not been reported as a
serious limiting factor in triticale development. Fuentes ( 1973) summarized
the literature on diseases of triticale. Larter el al. (1968) reported that in
higher latitudes ergot caused by Claviceps purpurea is a serious problem.
Grain contaminated with sclerotia of ergot causes toxicity problems in
animal feed. There is very little genetic resistance to the disease, although
considerable protection from infection can be obtained among highly selffertile strains. This form of protection is present among cultivars of wheat
and other cereals.
European investigators have reported that triticale generally is more resistant to diseases than wheat (Pissarev, 1963; Shulyndin, 1972; Kiss, 1973).
Leaf rust (Puccinia recondita) and stem rust ( P . graminis) attack triticale
and are considered the most serious diseases at many of the international
triticale yield nurseries.
Leaf rust and stripe rust ( P . glumarum) are serious pathogens of triticale
F. J. ZILLINSKY
in Mexico, and natural infestations occur regularly in the summer nurseries
in the State of Mexico. Rajaram et ul. (1972) observed that the many of
the triticale strains in the CIMMYT program were susceptible in the seedling stage to 4 races of leaf rust which attack INIA, and SIETE CERROS and
other cultivars or bread wheat. However, some of these (19 out of 75)
were resistant in the adult plant stage. During investigation of patterns of
leaf rust development in triticale, it was observed that some strains which
are susceptible to leaf rust in the seedling and early adult stage abruptly
produce the telial stage prior to maturation (Zillinsky, 1973). This would
tend to restrict the production of inoculum and thus provide some degree of
protection. Quiiiones et ul. (1972) reported that each of the strains 6A-190,
ROSNER, Armadillo, BRONCO, and TOLUCA 16a have a single dominant gene
for resistance to leaf rust which was derived from the wheat parent, and
that resistance carried by the rye parent was not expressed in the amphiploid. This is probably true for seedling resistance, but adult plant resistance
is carried by octoploid triticale strains derived from cross between INIA
wheat and several ryes to races that attack INIA in both seedling and adult
stages (Rajaram et ul., 1972).
Resistance to stripe rust was essential to maintain a nursery in Toluca,
where stripe rust infestations can be devastating. Quiiiones and Rodriguez
(1973) observed almost 100% of the triticale strains were destroyed by
stripe rust in the first season the triticale nursery was grown at Toluca.
Resistance was obtained from intercrosses among resistant plants and backcrosses to resistant wheats. The continued use of resistant strains as parents
and heavy selection pressure for resistance has resulted in a degree of
resistance superior to that found in most durum and bread wheats to races
currently prevalent in Mexico.
Bacterial diseases attack triticale strains in the Mexico nurseries and
other areas of North America. Dr. Bradbury of the Commonwealth Mycological Institute, Kew, isolated Pseudomonus striufuciens from a bacterial
leaf stripe lesions on triticale from the Toluca nursery in 1972. During
the next growth cycle at CIANO in the Yaqui Valley, he isolated the bacterium Xanthornonus trunslucens from bacterial lesions on infected leaves
(J. M. Waller, private communication W-1549 and W-1557).
A very serious outbreak of bacterial stripe, probably due to Xanthornonus trunslucens, occurred on triticale in the nursery at Navojoa, Sonora,
in February and March 1970. Many of the strains were susceptible, and
they were almost completely defoliated. Resistant plants were selected and
used in crosses. The spread of the disease is highly dependent upon a favorable environment, which occurs occasionally in the nursery areas, and continuous dependable screening for resistance to the disease has not been
possible. In Mexico, rye strains have generally been more severely damaged
by bacterial diseases than the wheats.
THE DEVELOPMENT OF TRITICALE
Leaf blight caused by Fusarium nivale occurs regularly on triticale and
wheat in the Toluca nursery throughout the growing season. This fungus
is of little importance to spring crops in other regions of the world (Richardson et al., 1972), but it is devastating on susceptible strains of triticale
and wheat in the Toluca Valley. There does not appear to be clear cut
resistance among the triticale strains, although some strains are killed and
others damaged only slightly. The more tolerant strains are infected much
later in the growth cycle. This disease has not been observed at elevations
below 7000 feet in Mexico. It is possible that the disease is indigenous
on grass species in regions at high elevations in Mexico and other countries.
Fuentes ( 1973) has investigated leaf blotch on Septoria tritici wheat
and triticales in the CIMMYT breeding program. He has found that triticales are generally more resistant than wheats to the strains of this pathogen found in Mexico. He assumes that other areas, such as North Africa,
the Middle East, and South America, may have strains that are more
virulent, since some reports on the reaction of triticales to Septoria tritici
have indicated high susceptibility. Lesions of infected triticale leaves from
the nurseries in Mexico have been examined regularly since 1971. The
pycnidia of Septoria tritici have been isolated only rarely although pycnidia
resembling Septoria nordorum and Septoria avenue f. sp. triticea occur regularly. More intensive investigations need to be carried out on diseases causing leaf blotching, particularly in the cooler and more humid areas of the
tropics. It may be possible to replace wheat with triticale, which is more
resistant to these diseases.
Triticale appears to be more resistant to powdery mildew (Erisiphe
graminis) and the smuts (Ustilago spp.) than wheat. However D. D.
Morey observed powdery mildew on triticale in the winter nursery at
Tifton, Georgia. A few spikes infected with loose smut have been found
in the Toluca nursery. Occasional plants infected with downy mildew
(Sclerophthora macrospora) are found in the CIANO nursery each year.
Head blights, foot rots, and seed infections occur regularly in the summer
M. J. Richardson of East Craigs, Scotland, and J. M. Waller of the
C.M.I. Kew investigated diseases of triticale in the Mexico nurseries during
the fall of 1973. They observed fruiting structures of Ophiobolus graminis,
Cochliobolus sativus, and Fusarium graminearum on triticale plants. They
also isolated several seed-borne pathogens on seed produced in Toluca and
El Batan (Richardson and Waller, 1973).
Several virus diseases have appeared in triticale nurseries. The aphidtransmitted barley yellow dwarf virus infects triticale strains in the Mexico
nurseries. The proportion of plants infected is generally higher than among
wheats in the same area, but much less than either oats or barley. Symptoms on infected triticale plants are similar to those developed on infected
F. J. ZILLINSKY
bread wheats. Triticale plants infected with African cereal streak virus were
observed in nurseries at Njoro, Kenya, and Debre Zeit, Ethiopia, during
February and March 1973. Viruslike symptoms on the leaves of juvenile
triticale plants were observed in the nursery at Ankara, Turkey, in 1972,
but positive identification was not made.
Diseases have not generally appeared as a serious problem in the
CIMMYT triticale breeding program. However, as commercial production
increases, diseases that parasitize triticale will increase. A close watch on
disease development will have to be maintained as production spreads.
Genetic resistance appears to be available for most of the diseases observed
to date. It is extremely important that breeding programs maintain a broad
germplasm base to provide protection against present and future disease
Dr. Alejandro Ortega, an entomologist in the CIMMYT corn program,
observed that triticale are generally attacked by the same insects as other
cereal crops. Infestations of corn leaf, English grain, and cereal root aphids
are common. Occasionally heavy infestations of shoot fly, frit fly, and stink
bugs have been observed on triticale in localized areas in Mexico. Triticale
plants infected with Hessian fly and root knot nematodes have been found
in North Africa. Care must be exercised in using insecticides on triticale.
Some strains of triticale are quite sensitive to pesticides applied as foliar
The introduction of the Armadillo strains into replicated tests resulted
in a significant increase in grain yield. The degree of improvement was
influenced by soil fertility, diseases, and other environmental factors. The
Armadillo strains approached the Mexican bread wheats in grain yields
at low to moderate levels of nitrogen, but dropped off sharply with increases in levels of nitrogen. The Armadillo strains were less responsive
to nitrogen fertilization and more susceptible to lodging than the Mexican
dwarf bread wheats.
An estimate of the rate of improvement in grain yield of triticales compared to that of bread wheats in Mexico during the past 6 years can be
obtained from yield data from replicated tests at CIANO Experiment Station, Sonora (Fig. 6 ) . The rapid increase in yield improvement of triticale
between 1967 and 1969 was due to the introduction of Armadillo strains.
Improvement in yield during the next 2 years (1969-1970, 1971-1972)
occurred at more or less the same rate in both wheat and triticale. The
introduction of fertile two-gene dwarf triticales into the 1972-1 973 yield
tests resulted in a second significant increase in the rate of yield improvement. It is expected that as more dwarf triticale strains from the breeding
THE DEVELOPMENT OF TRITICALE
FIG.6. Yields of triticale at the Sonora, Mexico, winter nurseries. Comparison of
yields of top wheat vs the average of the top triticale strains.
program are advanced to the replicated trials, triticale yields will equal
or surpass those of the best bread wheats in the Yaqui Valley of Sonora.
Grain yields of triticale are already competitive with wheat in some of the
high mountain valleys and on some sandy soils areas of Mexico.
Further increments in grain yield in triticales are expected with the introduction of more dwarfing genes and improvements in tillering capacity,
grain density, plant structure. An immediate increase of 10-1 5 % could
be achieved if triticales could produce grain of equal density to wheat.
Increasing spike length may also result in yield increases perhaps compensating for the present deficiencies in tillering capacity.
5. Grain Quality
The most important unsolved problem in triticale breeding is abnormal
endosperm formation resulting in seed shriveling, low test weight, and low
germination rate. As the spikes approach maturity, abnormalities appear
and seed development becomes progressively more abnormal as it ripens.
The ripe seeds have a wrinkled seed coat, lack luster, and have a deep
crease. The endosperm is chalky in contrast to the hard vitreous seed of
durum and bread wheat. The test weights range from 58 to 72 kg/ha,
while the best bread wheats have test weights in excess of 80 kg/ha.
F. J. ZILLINSKY
Lebedeff (1934) and others have suggested that abnormal seed development may be due to the deleterious effect of inbreeding on the rye genome.
Sinchez-Monge (1969) showed that improvement in fertility and grain
quality could be achieved by using self-fertile ryes as parents in triticale
breeding. Inbred ryes are being used as parents in the CIMMYT program
although no naturally self-pollinating ryes are available.
More intensive research has recently been undertaken at the University
of Manitoba in an effort to identify the causes of seed shriveling and the
means to overcome the problem. Larter (1973) reported that research was
being conducted on cytogenetics and cytology (by Darvey and Kaltsikes) ,
histology (by Shealy) , and biochemistry (by Hill). This research has provided information on the physical and developmental aspects of seed
Bennett (1973) suggested that differences in the duration of the meiotic
cycle might influence endosperm development. Distinctive chromatin formation in the chromosomes of the rye genome may require more time for
replication than for the wheat chromosomes. Thus, disturbances in cell
reproduction in endosperm tissue result from segments of late-replicating
heterochromatin at the telomeres of rye chromosomes.
Improvements in grain quality have been achieved by breeding, although
progress is slow. Triticale is more sensitive to environmental influence than
wheat. A higher protein content tends to be associated with increased endosperm shriveling. Thus, the higher protein content of the grain observed
in the earlier triticale strains is deceiving and may be the result of abnormal
development at the expense of other nutrients.
Villegas (1973) has shown that a marked decrease in protein content
has occurred with improvement in yield capacity and kernel plumpness.
The increase in yield has more than compensated for the loss in protein
so that the production of protein per unit area has increased (Figs. 7 and
Visual screening for plumper seed had to be applied with considerable caution. There is a strong tendency to eliminate all selections from
wide crosses possessing dwarfing genes, since these forms produce shriveled seeds. Improvement in seed type is obtained at the expense of
desired plant types or those having wide genetic diversity unless care is
exercised to avoid discarding those selected for characters other than plump
Dr. Ake Gustaffson of Lund, Sweden, initiated mutation research to improve seed quality in triticale in 1969, using mutagenic chemicals and radiation at several concentrations. Some improvement in seed type was obtained from this material among selections made in Mexico on the third
and fourth generations after treatment.
THE DEVELOPMENT OF TRITICALE
FIG.7. Comparison of triticale seed produced at CIANO, Sonora, in 1967, 1970,
- I 2
FIG.8. Grain yield (0-0) vs percent protein ( O - - - O )in triticales at