Fibre length and strength (negative) and maturity percentage (positive) showed consistent and highly significant associations with lint yield across these two data sets.
Micronaire and fineness did not have consistent associations with yield.
We therefore conclude that there were definite negative associations between high yield and desirable quality levels in cotton breeding populations in Australia and US.
The literature is not clear on reasons for these associations and precise study of those reasons is required to develop strategies for breaking the associations.
Larger population sizes, eliminating families with high negative association between yield and quality and recurrent selection with the best combination of yield and quality are suggested as strategies to utilise at present.
Breeding for genuine length and strength improvements at modern yield levels have not had 100 years of effort.
The CSIRO program in Australia has only had about 12 years of specific effort. We cannot claim success yet but the identification of targeted strategies is a positive step forward for breeding.