Sunflower (Helianthus annuus L.) is cultivated over a wide range of photoperiod and temperature conditions. The objective of this research was to determine the response of 16 diverse sunflower genotypes to varying photoperiod and temperature. Field research sites were at five locations including Argentina, Alaska, and Hawaii. The photoperiod (daylength + twilight hours) at vegetative emergence (VE) for most sites, except Hawaii and Alaska, ranged from 14.5 to 16.2 h. Within this range there was no evidence that photoperiod affected the time from VEto bud visible stage (R1). Instead, differences in temperature at various sites were responsible for differing rates of development. However, the short photoperiods at VE at Hawaii (11.2 h) decreased development rate and long photoperiods at Alaska (24 h) increased development rate dramatically. A nonparametric method was applied to rank the genotypes into temperature response groups. Four distinct groups resulted for the VE to R1 growth stage intervals, which were classed as very quick, quick, medium, and slow with respect to the time taken to reach RI from VE. Four groups also resulted for the R1 to anthesis (R5.1) time period. The groups for R1 to R5.1 differed from those for VE to Rl. Genotypes within these groups were dissimilar to those for VE to Rl. It was concluded that models used to predict phenological development in sunflower could be based on temperature alone provided photoperiod was within the 14.5 to 16.2 h range. To predict development outside this photoperiod range would require photoperiod be included in the model.

Synchronization of anthesis of male and female sunflower (Helianthus annuus L.) inbreds to be crossed is important for successful hybrid seed production. The use of weather‐related models, which predict anthesis dates in scheduling planting times could assist in this synchronization. The field plantings of 16 sunflower genotypes in locations from 32°51' S to 64°50' N lat. during 1982 and 1983 were used to develop an anthesis predictive model. This model which related mean daily screen temperature to daily rate of development accounted for genotypes classified into developmental response groups as very quick, quick, medium and slow. The model had satisfactory predictive ability when validated on independent data. The model requires refinement to account for temperature by photoperiod interactions when additional data are available.