Tải bản đầy đủ - 0trang
XIV. Breeding and Genetics of Tropical Grasses
E. M. HUTTON
cies of Digitaria. Setaria is almost entirely cross-pollinated (Gildenhuys,
1960), which facilitates hybridization, and although S. almum is selfcompatible, hot water can be used to emasculate whole inflorescences
(Pritchard, 1965a). A number of the important tropical grasses are apomictic. These include buffel (Snyder et al., 1955), guinea (Warmke, 1954),
green panic, molasses, paspalums (Bashaw and Holt, 1958), and species
in the genera Brachiaria (Pritchard, 1967), and Urochloa. In obligate
apomicts, the only variation available for selection is that which exists
between the accessions collected from different ecological niches. Most
produce some functional pollen, so crosses are possible if sexual forms of
the apomicts can be found. Burton and Forbes (1960) overcame the
apomictic barrier in P . notatum by crossing the common apomictic tetraploid and fertile induced tetraploids from the sexual diploid Pensacola
bahiagrass. The search for a sexual type in other important apomicts has
been successful only in buffelgrass (Bashaw, 1962), and the resultant
crosses have released considerable variation and have shown apomixis
to be recessive to sexuality.
The main objectives of the breeding work with setaria at the Cunningham Laboratory are to produce cultivars with frost resistance, high feeding value, low oxalic acid content, and an extended growing season. The
diploid Nandi (2n = 18) and tetraploid Kazungula ( 2 n = 36) belong to the
S . sphacelata complex, which also contains pentaploid, hexaploid, octoploid, and decaploid races (Hacker, 1966). Crosses have been obtained by
Hacker ( 1967) between all proximate ploidy levels except diploid and
tetraploid, and also between high and low levels. Thus for seed production, lines of setaria should be isolated from each other. Hacker (1968a)
has cast doubt on the validity of the separation of species in the S.sphacelata complex as he has been able to hybridize diploid forms of S . anceps and S . trinervia, and S.anceps and S . splendida, and hexaploid lines
of all three species. From Hacker's work (1 968b) it appears that the S .
sphacelata complex forms an autopolyploid series.
B. Sorghum almum
The aim is to breed lines of S . almum with higher yield and persistence
than the Australian cultivar Crooble and possessing juicy stems, distinctive brown glumes, late flowering, and tolerance to leaf diseases. Pritchard ( 1965a) crossed S . almum and perennial sweet sudangrass (Hoveland, 1960) and found that juicy stem and brown glume and plant color of
the latter were linked and mainly tetrasomically inherited. Selection was
facilitated by an association between translucent midrib, juicy stem, and
high soluble carbohydrate, and some of the advanced lines have a 20%
higher soluble carbohydrate in the stem than Crooble.
Using a tetrasomically inherited albino seedling character in S . almum,
Pritchard ( 1965b) studied natural crossing in S . almum and between S .
almum and the weed S . halepense and concluded that there was a degree
of genetic isolation between these species. From cytological examination
of aneuploid plants of S . almum and S . halepense I’ritchard (196%) suggested that these species were autotetraploids. Pritchard ( 1965d) crossed
diploid Sorghum with S . almum and obtained tetraploids and triploids.
Segregation in tetraploid progeny resulted in tramfer of certain characters from the diploid to tetraploid level. Slight fertility of the triploids also
enabled transfer of characters from the diploid to tetraploid (or near tetraploid) by backcrossing and selfing. The triploids could also be used to
transfer such characters as perenniality from tetraploid to annual diploid
Burton’s notable development ( 1 947, 1954) of (coastal bermudagrass
has been followed by further work aimed at improving its agronomic
characters and feeding value. Dry matter digestibility of the many genotypes was determined by Burton er al. ( 1 967), who found that quality of
a number of clones decreased as age of forage increased and that genotype
X age interactions were not significant. A coastal >: Kenya 56 F1 hybrid
averaged 12.3% more digestible dry matter than coastal over a four-year
period. Several Midland X Kenya 6 1 hybrids had higher yields and better
digestibility than either parent. This work has indicated that the quality
of C . ductylon may be improved by breeding.
At the Cunningham Laboratory, Pritchard (196’7) crossed the sexual
buffelgrass from Texas (Bashaw, 1962) with the main apomictic cultivars
and obtained more variation than has been assembled in over thirty years
of introduction. A number of promising leafy lines which flower later than
the parents and have greater cold tolerance have been selected. The in
vitro digestibility at maturity of some of these is siiperior to that of the
parents, so it may be possible to improve the feeding value of buffelgrass.
Pritchard (1 967) found that the original sexual plant has a chromosome
number of 2n =36 and that selfed progeny numbers range from 2n = 35 to
2n = 38. Chromosome numbers of the apomictic cultivars are 2n = 36 for
Molopo and Lawes and 2n = 43 for Tarewinnabar, Nunbank, and Biloela.
E. M. HUTTON
Progeny of crosses between these apomicts and the sexual type have
chromosome numbers from 2n = 34 to 2n = 45.
The examples given indicate that agronomic characters and feeding
value of tropical grasses can be improved by breeding. Of the grasses not
mentioned in this context, pangola and kikuyu merit attention although
they pose special difficulties for the breeder. There is a need to increase
their adaptability and it is possible that even their relatively high feeding
value and response to nitrogen could be improved. In future it is hoped,
as more information becomes available on the physiological and biochemical characters which control growth and adaptation of grasses, that
more precise selection in breeding populations will be possible.
Xv. Beef Production from Legume-Based Tropical Pastures
The degree to which tropical pasture research and development is
successful can be measured only in terms of animal production and its
profitability. Productivity of nitrogen-fertilized grass systems has already
been dealt with, and in this section animal production from legume-based
pastures is discussed. Because of their flexibility and cheap production of
N for pasture and grazing animal, legume-based pastures will continue to
predominate in development of tropical areas. Tropical legumes are not
able to produce enough N for the associate grasses to attain their potential dry matter production, but does this matter? Feeding value of the
pasture is the main determinant of animal production, so a higher proportion of legume and smaller bulk of less digestible grass is an advantage.
In Norman and Stewart’s experiments (1 964) (see Table 2) at Katherine
in the Northern Territory, liveweight gain of cattle was directly related
to the proportion of Townsville stylo in the pasture.
Dry Season Performance of Cattle on Sown Pastures with
Varying Proportions of Grass and Legume
Composition of pasture
of pasture at
start of grazing
In general, progress in tropical beef production depends on persistent
and adapted legumes, regular application of superph,osphate (giving both
P and S), and use of adapted tropical cattle with tick. resistance and heat
tolerance (Schleger and Turner, 1965). Significant advances have been
made despite Whyte’s pessimistic conclusions ( 1962) concerning improvement of tropical grasslands. The following examples from areas in
the main tropical climates will make this clear.
A. WET TROPICS
Younge et al. (1964) estimated that about one-quarter of the Hawaiian
rangelands are unproductive low wetlands but capable of trebling the
current annual beef production. On the island of h4aui, pangolagrassD . intortum pastures fertilized once per acre with lime at about 3000 Ib
and a starting fertilizer comprising 44 Ib N, 84 P, 104 K, 3.5 B, and 2.5
Mo gave a mean over two years of 764 Ib liveweight gain per acre per
annum at a stocking rate of about two beasts an acre, which was highly
profitable. Also on Maui, pangola, dallis, kikuyu, and native grasses each
mixed with kaimi clover and given one dressing per acre of 6 tons of
calcium carbonate and starting fertilizer as in the previous experiment
produced over four years annual liveweight gains per acre of 720, 630,
575, and 524 Ib, respectively. At the Kauai Branch !station on an aluminous-ferruginous latosol, Younge and Plucknett (1966) with a pangolagrass-D. intortum pasture given an application of basic fertilizers and
four rates of P produced as high as I164 Ib liveweight gain per acre per
annum with yearling steers. Mean stocking rate varied from 1.18 beasts
an acre at the lowest P level to 2.38 an acre at the highest. There was a
curvilinear response to P, but the highest liveweighi: gain was obtained
from the heaviest P application. Unimproved pastures produce about 30
Ib liveweight gain per acre per annum.
Over two dry seasons at Katherine in the Northern Territory, liveweight of steers at a beast to 2 acres of Cenchrus-Townsville stylo pasture increased and that of steers at a beast to 17 acres on native pasture
declined substantially (Norman and Arndt, 1959). Shiiw (196 I ) at Rodd’s
Bay, Queensland, showed that year-round productivity of native speargrass pasture could be increased markedly by oversowing with Townsville stylo and topdressing annually with 1 cwt molybdenized superphosphate an acre. Carrying capacity of native pasture was trebled from a
steer to 9 acres to a steer to 3 acres, annual liveweight gain per acre was