|Authors: ||E.S. Redjeki, S. Mayes, S. Azam-Ali|
|Keywords: ||linear regression, genotypic and environmental interaction, regression coefficient, deviation from regression, well adapted landrace|
Bambara groundnut has been planted in Indonesia for hundreds of years.
Researchers have evaluated where the Indonesian Bambara groundnut landraces were introduced from but no-one has evaluated the stability and adaptability of Bambara groundnut in Indonesia.
Thirty-six landraces were planted in Indonesia, together with putative Indonesian × African hybrids and their offspring.
These were assessed for their stability and adaptability by the methods of Finlay and Wilkinson (1963) and Eberhart and Russel (1985). Results from seven landraces are presented.
The seven landraces were: ‘LunT’ from Sierra Leone; ‘AHM753’; ‘SB165A’ and ‘S19-3’ from Namibia; ‘DODR’ from Tanzania; ‘Uniswa Red’ from Swaziland; ‘DIPC’ from Botswana; and the Indonesian landrace ‘Gresik’ as control.
Thirty plants of each landrace were planted in a randomised block design with three replicates at Gresik, Bojonegoro and Jatikerto in Indonesia in November 2009. Each location had a different altitude, soil type and rainfall.
Gresik is the main Bambara groundnut growing region in the East of Java, Indonesia.
Prior to this experiment, farmers in Bojonogoro and Jatikerto were not familiar with this crop.
Many traits were assessed based on the list of descriptors of Bambara groundnut issued by IPGRI, but in this report we present only the results of stability and adaptability analysis for 50% flowering, days to maturity, pod number per plant and the
100 seeds weight traits.
Analysis of variance showed highly significant differences in all three locations and combined analysis of variance over sites (Gomez and Gomez, 1983) indicated that location, landraces and location × landraces interaction are substantially different (1%). Stability and adaptability parameters were obtained as the linear regression coefficient (bi) of the mean of all data observed and deviation from the regression analysis (S2di) with the hypothesis that bi=1 and S2di=0. The results indicated that almost all landraces observed were stable, but only three landraces revealed good adaptability in all three locations, namely ‘SB165A’, ‘Uniswa Red’ and ‘DIPC’. Meanwhile ‘LunT’ and ‘S19-3’ are considered promising landraces because they are well adapted in two of the four variables used.
This information could prove useful for breeding programmes.
Download Adobe Acrobat Reader (free software to read PDF files)