Docstoc

Drought Avoidance Screening

Document Sample
Drought Avoidance Screening Powered By Docstoc
					Drought Avoidance Screening
Materials and methods

A full description of the experimental procedure is given in Price et al. 2002; Plant
Molecular Biology 48 683-695.

Plant material and drought screening experiments

F6 seeds were collected from a total of 205 F5 plants produced by single-seed descent
from an original set of 310 F2 plants derived from a cross between varieties Bala and
Azucena (see Price et al. 2000 TAG 100: 49-56 for full details). A summary of the
drought conditions is given in Table 1. Details for each experiment are given below.

IRRI 1996 (I96)
A total of 176 lines plus one entry of both parental varieties were sown on 23 January
1996 in the Upland site of the International Rice Research Institute (IRRI). The layout
was a factorial design replicated twice with fully irrigated and stressed treatments
(two replicates of each). Each plot consisted of 3.0  3.0-m-long rows of plants
spaced 0.1 m apart, with 0.25 m between rows. Three seeds were sown per hill which
were thinned to one plant after emergence. After every 20 plots, four check-variety
plots were sown (only data for check varieties IR50, Moroberekan, and IRAT 104 are
presented here). Fertilization followed the standard procedures used at IRRI. For
weed control, a spraying just after sowing of the pre-emergence herbicide Oxadiazon
at 1.0 kg ha-1 was followed by hand weeding. For insect control, Carbofuran was
applied in the furrow at 1.0 kg ha -1 at sowing and then broadcast every four weeks.
Sprinkler irrigation was given every other morning for 2 h, enough to wet the soil
profile. Watering was withheld from two replicates on 2 March (40 d after sowing).
On 19 March (after 18 d of stress), 18 mm of rain fell. After 12, 14, 17, 21 and 24 d of
stress, the degree of leaf rolling was assessed visually on a scale of 1-5 (1 = unrolled,
5 = fully rolled) essentially according to the standard evaluation system of IRRI
(1996). These assessments were averaged to give a mean leaf rolling score over the
drought period (I96LR). After 26 and 28 d of stress, the degree of leaf drying was
assessed visually, again using a scale of 1-5 (1 = no evidence of drying, 5 = all leaves
apparently dead), essentially according to the standard evaluation system of IRRI
(1996) and these two assessments were averaged to give a mean leaf drying score.

IRRI 1998 (I98)
A total of 118 lines (randomly chosen from the previous experiments) plus the two
parental lines together with six other check varieties were sown in 0.75  2.0-m plots
according to an -lattice design with two replications on 12 December 1997 using the
same plant spacing and chemical treatments as above. The plots were irrigated three
times weekly by sprinkler until 27 January 1998. Thereafter, water was withheld
(starting at 45 d after sowing) and no rain fell during the stress period. Leaf rolling
(after 14, 16, 18, and 22 d of drought), leaf drying (after 20, 22, and 23 d of drought),
and relative water content (after 18 and 23 d of drought) were assessed and averaged
as above.

WARDA 1997 (W97)
A total of 142 lines plus the parental varieties were sown in three replicated plots in a
plot size, layout, and randomization similar to the experiment at IRRI in 1996. After
every eight experimental plots, one plot of a check variety (Moroberekan, OS6, or
SP14) was included. Seeds were sown on 11 January 1997. Fertilization followed the
standard procedures used at WARDA (basal NPK (10-18-18) at rate 200 kg ha-1 + 40
kg N (urea) at 35 and 60 day after sowing). For weed control, a spraying just after
sowing of the pre-emergence herbicide Oxadiazon at 4.0 l ha -1 was followed by hand
weeding. The plants were watered by boom irrigation every other day. On 5 February
(25 d after sowing), water was withheld from two of the replicates for 30 d. After 10,
12, 14, 16, 18, 20, 22, 24, 26, 28, and 30 d of stress, the degree of leaf rolling was
assessed visually, while the leaf drying was assessed after 14, 16, 18, 20, 22, 24, 26,
28, and 30 d of stress. As with the IRRI data, an average leaf rolling and drying score
was calculated. Relative water content was assessed after 21 days of stress.

WARDA 1998 (W98)
A total of 110 lines plus five replicates of each parental varieties were sown in four
replicated plots of 1m x 1m in a factorial design using a plant spacing as used
previously. After every eight experimental plots, one plot of a check variety
(Moroberekan, OS6, or SP14) was included. Seeds were sown on 7 January 1998 and
watered by boom irrigation every other day. On 1 February (24 d after sowing), water
was withheld from two of the replicates for 30 d. After 10, 12, 14, 16, 18, 20, 22, 24,
26, 28, and 30 d of stress, the degree of leaf rolling was assessed visually, while leaf
drying was assessed after 14, 16, 18, 20, 22, 24, 26, 28, and 30 d of stress. As with the
IRRI data, an average leaf rolling and drying score was calculated. Relative water
content was assessed after 26 days of stress.

Statistical analysis

Initially, the data contained variation because of position as well as variation caused
by genotype and other environmental factors. All data except IRRI 1998 data were
corrected for spatial variation by using linear regression to fit equations of the form

   z = a + bX + cX2 + dY + eY2

where z is the measured data value, X and Y are the coordinates of a plant within the
experimental area, and a, b, c, d and e are parameters to be estimated. Separate
regressions were fitted for each measured variable at each site. In each case, the fitted
values were subtracted from the measured values. This removes spatial variation
accounted for by the regression. The overall mean for that variable was then added to
these values to give estimates of the measured values, corrected for positional effects
in the plot. The IRRI 1998 data were not treated in this way because the -lattice
design accounts for spatial heterogeneity.
Broad-sense heritability was computed from two-way analysis of variance (factors;
genotype and replicate) of the F6 from the estimates of genetic (2G ) and residual
(2e) variance derived from the expected mean squares as h2 = 2G /( 2G + 2e/k)
where k was the number of replications. For each trait, two-way analysis of variance
(factors; genotype and experiment) was conducted using all genotypes which were
common to all experiments (74-77 genotypes).
In many cases, results from QTLs are indicated for the average of the two sites and
for the overall average. For calculating these average values data were normalised by
dividing the value for each individual by the population mean for that trait and means
were then calculated using only lines which were common to experiments. These
calculations were done in order to identify regions with a broad environmental
stability from those that are environment specific and to increase the power of
detection of QTLs common to specific sites or both sites by reducing the error in
analysis (i.e. increasing replication).