Global warming has created increased nighttime temperatures both in field and greenhouse production of chrysanthemums during flower bud initiation (FBI) and development, causing heat delay or complete cessation of flowering. Integration of breeding and selection for heat delay insensitivity (HDI) has become imperative for greenhouse (cut, potted types) and must be accomplished on a genotypic basis, similar to winterhardiness. This is a breeding objective in the joint garden chrysanthemum breeding project between the Chungnam Provincial Agricultural Research and Extension Services and the University of Minnesota. The objectives of this research were to test 10 genotypes (cultivars, seedlings) from both breeding programs when grown in low-temperature (LT) and high-temperature (HT) short-day (SD) and long-day (LD) conditions (four environments: LTSD, LTLD, HTSD, and HTLD); determine the extent of heat delay and HDI for visible bud date (VBD), flowering, and other phenotypic traits; evaluate relative injury (RI) and cell membrane thermostability (CMT), and to select future parents with lowered RI values, higher CMT, shorter heat-induced flowering delay, and/or HDI. ‘Magic Ball’ and ‘Minnwhite’ had the shortest plant height in HTLD and HTSD, whereas ‘Geumbangul’ had stability for height in all treatments. Lowest long day leaf numbers (LDLN) occurred under LTSD in seven genotypes. However, both ‘Geumbangul’ and ‘Magic Ball’ had complete stability for LDLN across all environments. Sigmoid curves for RI% and temperature were found for all genotypes and environments with R2 = 0.79–0.89. Only ‘Mellow Moon’ had stability or equal VBDs in HTSD, LTSD, and LTLD conditions. This is the first-ever report of stability for VBD across inductive and noninductive HT/LT treatments. Only ‘Centerpiece’ flowered in all environments and also had 0 day of heat for VBD in LT and 1 day of heat delay in HT, as well as three others (Mn. Sel’n. 01-210-43, ‘Autumn Fire’, and ‘Geumbangul’). Few had linear regressions with positive slopes for heat-induced VBD or flowering delay regressed with RI%; most had no slope (R2 ≈ 0.0) for all treatments (‘Centerpiece’, Mn. Sel’n. 01-210-43), whereas others were negative (‘Mammoth™ Dark Bronze Daisy’, Flw LTLD–LTSD). Surprisingly, one linear regression had a slope of R2 = 1.0 (‘Geumbangul’, Flw LTLD–LTSD). These responses are all novel in chrysanthemums. Selecting the best parents in both breeding programs to maximize stability of all traits across these four environments with minimal crossing and selection across generations could be accomplished by stacking parental traits. A crossing scheme involving just three parents is proposed to incorporate stability for all traits in just a few generations.
