Published online by Cambridge University Press: 27 March 2009
Progenies of randomly chosen plants from spaced plant populations of four American and one African variety of Gossypium hirsutum L. were grown in replicated single-plant plots at Kimberley Research Station in 1967–68 (all varieties) and 1968–9 (two varieties only) to estimate intra-varietal parameters of genetic variability.
Estimates of broad-sense heritability were usually similar when derived from components of variance on either a one- or two-season analysis basis or by parentprogeny regression Heritabilities were in general lowest for lint yield and boll number and highest for lint percentage. Mean value of heritability for lint yield and boll number, considered consistent enough on a varietal and seasonal basis to allow averaging, was about 0·15. Components of variance estimates of heritability on a single environment basis for other attributes ranged as follows: boll weight from 0·05 to 0·56, lint percentage from 0·29 to 0·80, mature height from 0·00 to 0·51, micronaire value from 0·20 to 0·63 and Pressley strength index from 0·05 to 0·56.
Genetic coefficients of variation showed an opposite trend to that for heritability: lint yield and boll-number values were large and lint percentage, micronaire value, and Pressley strength indices small. Expected responses to selection for the top 2·5 % of the population for lint yield and boll number at around 10% were usually higher than for other attributes.
Genotypic and phenotypic correlations were mostly similar but correlations often deviated widely between varieties. Positive lint yield correlations with micronaire value and negative lint yield correlations with Pressley strength index were recorded; these are often undesirable associations for breeding.
The association between the period of maintenance under Ord conditions and variability shown phenotypically in the original populations was not evident from the quantitative inheritance studies. Residual variability following breeding was considered as accounting for the larger portion of the genetic variability.