Over 100 wheat varieties and breeding lines from India and Australia were screened in alkaline and waterlogged soils in 10 environments over two years at one drained location and two naturally waterlogged locations in India. Mean trial grain yield was reduced up to 70% in the environments where genotypes were waterlogged for up to 15 days at the vegetative stage in alkaline soil relative to plants in drained soils. Agronomic traits (plant height, tiller number, 1000-grain weight) of genotypes were also reduced under waterlogging. At one waterlogged site, up to 68% of the genetic diversity for predicted grain yields under waterlogging could be accounted for by number of tillers (r² = 0.41–0.68 in 2011 and 2010, respectively) and positive correlations also occurred at the second site (r² = 0.19–0.35). However, there was no correlation between grain yields across varieties under waterlogging in any trials at the two waterlogged locations. This may have occurred because waterlogged sites differed up to 4-fold in soil salinity. When salinity was accounted for, there was a good correlation across all environments (r² = 0.73). A physiological basis for the relationship between tillering and waterlogging tolerance is proposed, associated with crown root development. Results are compared with findings in Australia in acidic soils, and they highlight major opportunities for wheat improvement by selection for numbers of tillers when crops are waterlogged during vegetative growth.