Temperature change is of potential to trigger the formation of unreduced gametes. In this study, we showed that short periods of high temperature treatment can induce the production of 2n pollen in Populus pseudo-simonii Kitag. The meiotic stage, duration of treatment, and temperature have significant effects on the induction of 2n pollen. Heat stress resulted in meiotic abnormalities, including failure of chromosome separation, chromosome stickiness, laggards and micronuclei. Spindle disorientations in the second meiotic division, such as parallel, fused, and tripolar spindles, either increased in frequency or were induced de novo by high temperature treatment. We found that the high temperature treatment induced depolymerisation of meiotic microtubular cytoskeleton, resulting in the failure of chromosome segregation. New microtubular cytoskeletons were able to repolymerise in some heat-treated cells after transferring them to normal conditions. However, aberrant cytokinesis occurred owing to defects of new radial microtubule systems, leading to production of monads, dyads, triads, and polyads. This suggested that depolymerisation and incomplete restoration of microtubules may be important for high temperature-induction of unreduced gametes. These findings might help us understand how polyploidisation is induced by temperature-related stress and support the potential effects of global climate change on reproductive development of plants.

Polyploid breeding has the potential to increase the economic secondary metabolites of Eucommia ulmoides. However, pollination with induced ploidy-mixed pollen has failed to produce polyploids (GAO, 2006). In this investigation, the morphological characterization and in vitro germination of heat-induced ploidy-mixed pollen of E. ulmoides were analysed to determine why there is no polyploid production. Heat-treated pollen grains were easily distinguished as large and small according to their length. The large pollen grains were significantly longer than both untreated and heat-treated small samples, suggesting that they were probably 2n pollen. Rather than the three germinal pores in small pollen, the large grains typically had four pores and, in some cases, shallow furrows, which might affect their germination. Although the maximal germination rates of the treated small and large pollen were not significantly different, the large pollen germinated tardily during the early stages of incubation. The small pollen maintained its growth during the incubation, but the tube growth of large pollen almost stopped after 24 h incubation. Both vegetative and generative nuclei in the large pollen moved into tubes later than in small pollen and the frequency of mitosis in generative nuclei of large pollen was low. Therefore, the tardy germination, poor tube growth, and weak activity of both vegetative and generative nuclei probably caused the poor competition of large pollen in certation. Finally, techniques to increase the competition of highploidy pollen and the prospect of 2n female gamete induction in the polyploid breeding program of E. ulmoides are discussed.