The Effect of Temperature on Reproduction in FivePrimulaSpecies

Primula vulgaris Huds., P. veris L., P. frondosa Janka, and three populations of P. farinosa L. were legitimately and illegitimately pollinated, and the self-fertile P. scotica selfed and cross-pollinated and then subjected to uniform temperature conditions of 6, 15 or 26 °C for 4 d before gynoecia were examined for pollen germination and pollen tube growth, or plants progressed to seed set at 15 °C, after which seeds were weighed, germinated, and seedlings grown on. The temperature responses of pollen germination and pollen tube growth were not always congruent, and varied between species, populations, and often between morphs (pin and thrum) in the distylous species. Nevertheless, optimal temperature responses tended to be lower for vernal species ( P. vulgaris and P. veris ) and for subarctic P. scotica than for later flowering montane species. However, no relationship was found between pollen temperature response, and fertility. The greatest seed set occurred after legitimate pollination at 15 °C in most cases; a flowering temperature of 26 °C tended to impede seed set, except for P. scotica and the low altitude population of P. farinosa . In P. veris, P. frondosa and the high altitude population of P. farinosa, some illegitimate pollen germination and pollen tube growth occurred at 26 °C, but this did not lead to increased within-morph seed set in these self-incompatible species at this relatively high temperature. Temperature at flowering frequently affected average seed weight, and in P. veris and two populations of P. farinosa this attribute may have been influenced by seed number, the average seed weight of few-seeded capsules tending to be greater than for many-seeded capsules. A high seed weight might mitigate the disadvantageous effects of low fecundity resulting from interactions with flowering temperature. However, in P. vulgaris and P. scotica interactions between flowering temperature and seed weight may have other, undetermined, causes. The seed of four species germinated least well in standard conditions when set following a flowering temperature of 6 °C, which tends to support the hypothesis that temperature at flowering can affect seed physiology; in contrast the seed of the two upland populations of P. farinosa germinated least well after flowering at 26 °C. We conclude that much more work is needed on interactions between temperature and reproductive efficiency, but that preliminary indications suggest that a global increase in temperature at flowering might adversely affect the quantity and quality of seed set in some species. Copyright 1998 Annals of Botany Company