Seed germination of the wetland emergent perennial Phragmites australis is stimulated by diurnally fluctuating temperatures. A germination experiment in darkness and light at different temperature regimes showed that P. australis germinated as well in darkness as in light over most of the temperature regimes tested. The germination requirements could partly explain why this species, despite a large annual production of small seeds, does not accumulate a persistent soil seed bank. A second experiment examined the effect of one to four diurnal temperature cycles with amplitude of fluctuations ranging from 0 to 30°C; diurnal mean temperature of 15°C. Germination in the absence of fluctuations was low, and logistic model estimates showed a positive effect of number of temperature cycles which was strongly influenced by amplitude size. For seeds that have fallen on moist ground during the winter, these laboratory results predict that a few large diurnal temperature fluctuations could be sufficient for onset of germination in the subsequently warmer spring period. Seeds that are located under water, however, are predicted to remain ungerminated until the water level falls. Hence, we suggest that the requirement for fluctuating temperature, in addition to being a sensor for ‘exposed’ seed sites, is also effectively acting as a germination timing mechanism.

Germination responses to 45 combinations of diurnal mean temperature and amplitude were examined in freshly collected seeds of two wetland perennials: Typha latifolia L. and Phragmites australis (Cav.) Trin. ex Steudel. Nearly all seeds (>95%) germinated in favourable temperature regimes. Mean temperature (range 10–30°C) and amplitude (range 0–20°C) affected final germination of both species. P. australis required a high amplitude (> 10°C) for germination over the entire range of mean temperatures. Final germination of T. latifolia was more sensitive to mean temperature than P. australis. The germinated proportion of T. latifolia had a maximum around 20°C, above which it decreased, and amplitudes were more stimulating at low than at high levels of mean temperature. The germination rate was rapid and increased with mean temperature for both species. More than 50% germination was achieved within 1–3 d at favourable temperatures. It is proposed that the thermal requirements provide the non-dormant seeds with a season-sensing mechanism which postpones germination of seeds dispersed during autumn, winter or early spring, until the soil surface is heated by the sun in the spring and sufficiently large diurnal fluctuations of temperature occur. Furthermore, the amplitude requirement implies a strong avoidance mechanism for germination of P. australis in sites with small temperature fluctuations (e.g. below water tables), whereas seeds of T. latifolia appear to be less exacting in the requirements when the soil or water becomes warmer.