Positive and Negative Signals Regulate Germination in the Post-fire Annual, Nicotiana Attenuata

The litter of many plant species is known to inhibit germination, and this phenomenon is commonly interpreted as allelopathic inhibition of one species by another. However, an alternative interpretation is that seeds may be using environmental signals to inform the timing of their germination and thereby use dormancy as a mechanism of habitat choice. Nicotiana attenuata Torr. ex Wats. (Solanaceae) is typically found for less than three years after fire in the sagebrush, blackbrush, and pinyon–juniper forests of the Great Basin desert of North America. A dormant seed bank is established during this ephemeral post-fire period, and pyrolysis products of a-cellulose (containing only C, H, and O) in wood smoke are known to initiate germination in dormant seeds of this species. We demonstrated in a glasshouse experiment that germination into burned soils (as compared to unburned) results in a 12-fold increase in lifetime seed production, which reflects a minimum estimate of the fitness benefit of making accurate germination decisions. With seed bioassays, we examined the distribution of this smoke signal in the A1 soil layer at several burned areas in southwestern Utah, United States to determine whether the presence of this smoke-derived germination cue predicts the spatial and temporal occurrence of N. attenuata populations after fires. Although we found no evidence for the germination signal in areas that had not been burned for 30 yr, the occurrence of the germination signal did not perfectly coincide with the distribution of populations. We found evidence for its transport by wind and water into adjacent unburned areas (from 40 m to 1 km away from a burned site) and its persistence over time (for $7 yr), making this signal an unreliable indicator of the plant’s habitat. To resolve this discrepancy, we examined the effect of unburned A0 soil horizon on smoke-induced germination. The litter-containing A0 soil horizon (and aqueous extracts thereof), collected from underneath seven dominant species from later stages of post-fire succession, completely inhibited germination of both dormant and nondormant seeds, even in the presence of a smoke cue in excess of that required to elicit germination. The inhibitory effect was limited to the early stages of germination (48 h after exposure to smoke), and we confirmed these results with natural seed banks. We demonstrated that the A0 soil horizons and their aqueous extracts are not toxic to N. attenuata seeds or growing plants, and they have no effect on lifetime seed production. Moreover, they do not inhibit the germination of the nondormant, conspecific native tobacco, N. trigonophylla, which grows in the same area but is not associated with fire. Hence, these negative factors do not function in allelopathically mediated competitive interactions between N. attenuata and later successional species. We propose that the occurrence of N. attenuata populations after fires can be explained by the combined stimulatory effect of smoke-derived signals on the dormant seed bank and the inhibitory effect of signals from unburned litter, and that both signals are required for N. attenuata to identify its germination niche.