The degree of photoperiod sensitivity was studied in the F5 and Fs of three crosses between sensitive and insensitive rice cultivars under field conditions. Photoperiod sensitive lines could be identified as nonflowering lines in a February-seeded nursery. These lines had a short critical daylength (i.e. late flowering date). The strongly sensitive lines also had a high degree of photoperiod sensitivity, measured as the ratio of the difference in growth duration for two seeding dates to the difference in seeding dates in the main wet season. Some lines, however, with a relatively high degree of sensitivity and a longer critical daylength, were able to flower in the February-seeded nursery. In the cross KDML 105 x IR42, photoperiod sensitivity showed a bimodal distribution, suggesting major gene control. In Latisail x IR28 and KDML 105 x IR50, the inheritance was not clear. In KDML 105 x IR42, 37 out of 97 Fs lines were strongly sensitive and in KDML 105 x IR50,7 out of 95 Fs lines were strongly sensitive. Thus the choice of the insensitive parent greatly influenced the number of strongly sensitive lines recovered. Degree of photoperiod sensitivity was positively correlated with plant height in the wet season. In the dry season this was not observed, because the sensitive lines flowered earlier.