The mean of all plants of every F4 and F5 families in each
character was calculated and considered as a plot-mean, and subjected
to statistical analysis. The analysis of variance was performed, for
each F4 and F5 families of every cross separately (including Giza 9
and Sinai 1) as randomized complete block design (Gomez and
Gomez, 1984). The least significant difference (L.S.D) was estimated
to compare among means of the families, Giza 9 and Sinai 1. The
analysis of variance was made for each F4 and F5 families of each
cross again, excluding Giza 9 and Sinai 1.
The genotypic mean squares of the analysis of variance were
tested against error mean squares and the expectation of mean squares
were derived from plot data (Table, 2), then equated to the observed
values to estimate the component of variance. In addition sqares the
variance of F4 and F5 families in every cross was estimated separately
using MSTAT computer program.
Table (2): Form of variance analysis and mean squares expectations.
Source of Degrees of Mean squares Expectation of mean
variance Freedom squares
Family f–1 M2 Ơ2 e + r Ơ 2 g
Error (r-1) (f-1) M1 Ơ2 e
The variance components were estimated by the following
formulae (Miller et al., 1959):
Genotypic variance (Ơ2 g) = (M2 – M1)/r
Error (Ơ2 e) = M1
Phenotypic variance (Ơ2P) = (Ơ2 g) + (Ơ2 e /r)
The broad sense heritability (h2 b.s.) was calculated according to
Allard (1960) as follows:
Broad sense heritability (h2 b.s.) = (Ơ2 g / Ơ2P) x 100
The expected genetic advance from selections was estimated by
using the method suggested by Allard (1960) as:
GA=K. (Ơ2P)1/2. h2 b.s.
GA%= (GA/mean)× 100
Where: GA represents the expectation of genetic advance under
selection, (k) is a selection differential, which is considered 2.o6 in
The genotypic (G.C.V) and phenotypic (P.C.V) and environmental
(E.C.V) coefficient of variation were calculated as described by
Burton (1952) as follows:
G.C.V. = [(Ơ2 g)1/2 /mean] x 100
P.C.V. = [(Ơ2P)1/2 /mean] x 100
E.C.V. = [(Ơ2e)1/2 /mean] x 100
Simple correlation coefficients among all studied characters were
calculated for F4-families, and F5–families (Gomez and Gomez,
RESULTS AND DISCUSSION
1- Performance of the F3 base population
The base population in the present study was the F3-bulk
populations of five crosses. The mean values of all studied
characters for the five crosses with the check varieties Giza 9 and
Sinai 1 are presented in Table (3). Plant height ranged from 24.25
cm for cross 1 to 32.32 cm for cross 3. In comparison Giza 9 and
Sinai 1 showed plant heights of 36.20 and 25.39 cm, respectively.
Cross 3 had higher mean values for all characters than other
crosses, but it was late in flowering and maturity.
The average of seed yield/plant was very low for all crosses,
except cross 3, and ranged from 0.16g for cross 2 to 0.97g for
cross 5. Cross 3 gave a reasonable average seed yield of 1.47
g/plant. Low seed yield was mainly due to broomrape’s
(Orobanche) infection, which spread over the field. However,
many plants were not infected and gave reasonable seed yield.
Thus useful pedigree selection was practiced in all the five
crosses. Many selected plants gave seed yield above 1.4 g/plant
with average seed yields of the five crosses ranged from 1.0 to 1.8
g/plant (Appendix 1).
2. Mean values, variability and selection response
2.1. F4 generation
2.1.1. Performance of F4 families
The analyses of variance were made for each cross separately.
Results presented in Table (4) reveal significant differences among F4