| dc.description.abstract | This study was carried out with the objective of identifying the
important quantitative predictors of yield and to assess the utility of
such predictors in improving seed yield of two Kenyan bean accessions,
PAS/DOl and PAS/002.
To assess the components of variation in the two populations, 45
plants from each population were randomly chosen. The seeds from these
plants were used to develop 45 lines from which data were taken. The
lines were planted out on a randomized complete block design with two
replications during the short rains of 1986 and long rains of 1987 at the
Field Station of the Department of Crop Science, University of Nairobi.
Variation analysis was conducted for the total number of effective primary
branches, the total number of effective pods per plant, the average number
effective pods per branch, the number of seeds per pod, the weight of
twenty seeds, the total seed yield per plant, the average pod length, the
number of effective podding nodes and the number of days to flowering.
In the population PAS/DOl significant variation among the lines was
observed for the number of effective pods per plant, the number of effective
pods per branch, the weight of twenty seeds, the average pod length and the
number of effective podding nodes. The contribution of seasonal differences
on phenotypic variation was significant for all the traits studied except the
average pod length. The variation due to line-season interaction was significant
for all characters except the number of seeds per pod and the weight
of twenty seeds. Similarly, the variance component analysis for the PAS/002
population revealed significant variation among the lines for the number of
effective pods per plant, the number of effective pods per branch, the weight
of twenty seeds, the total seed yield per plant, the average pod length, the
number of effective podding nodes, and the number of days to flowering.
Variation due to seasonal differences was important for the number of
effective branches per plant, the number of effective pods per plant, the
number of seeds per pod, the number of effective podding nodes per plant
and the number of days to flowering. The variation due to the interaction
between the lines and seasons was significant for the weight of twenty
seeds only.
Phenotypic correlations analyses revealed the number of effective pods
per plant and the average pod length as the morphological characteristics
with the highest correlation to yield in PAS/O01. Similarly, the number of
effective pods per plant and the number of effective podding nodes were
shown to be traits most closely associated with the total seed yield in
PAS/DOZ. The multiple linear regression analysis confirmed the number of
effective pods per plant to be an important determinant of yield in the two
populations. The contribution of the average pod length to the total seed
yield per plant was significant for PAS/DOl only.
The populations arising from the various cycles of selection were subjected
to a comparative performance trial in a three-replicate completely
randomized block design at the Department of Crop Science Field Station
and at the National Dryland Farming Research Station, Katumani during the
long rains of 1987. Selection for increased number of effective pods, the
total seed yield per plant and the average pod length in PAS/DOl lead to no
significant direct response. However, the use of these traits as selection
criteria lead to significant indirect responses in the number of effective
podding nodes. Improving seed yield in PAS/DOl by using the number of
effective pods per plant as the selection criterion was 42 per cent more
efficient in improving seed yield during the second cycle of selection than
than was selection for yield per see In PAS/002, the use of the number of
effective pods per plant and the number of effective podding nodes per
plant also lead to significant responses in seed yield. The number of
effective pods per plant was 9 per cent more efficient in improving seed
yield of PAS/002 during the first cycle of selection and 4 per cent more
efficient during the second cycle of selection. The number of effective
podding nodes per plant was the most efficient approach for yield improvement in this population. It was inferior to selection for yield per se during
the first cycle of selection. However, its efficiency improved tremendously
during the second cycle of selection in which it was observed to be 16 per
cent more efficient in improving yield than was selection for yield per. | en |
