Evolutionary history of fire-stimulated resprouting, flowering, seed release and germination

Fire has shaped the evolution of many plant traits in fire-prone environments: fire-resistant tissues with heat-insulated meristems, post-fire resprouting or fire-killed but regenerating from stored seeds, fire-stimulated flowering, release of on-plant-stored seeds, and germination of soil-stored seeds. Flowering, seed release and germination fit into three categories of response to intensifying fire: fire not required, weakly fire-adapted or strongly fire-adapted. Resprouting also has three categories but survival is always reduced by increasing fire intensity. We collated 286 records for 20 angiosperm and two gymnosperm families and 50 trait assignments to dated phylogenies. We placed these into three fire-adapted trait types: those associated with the origin of their clade and the onset of fire-proneness [primary diversification, contributing 20% of speciation events over the last 120 million years (My)], those originating much later coincident with a change in the fire regime (secondary diversification, 30%), and those conserved in the daughter lineage as already adapted to the fire regime (stabilisation, 50%). All four fire-response types could be traced to >100 My ago (Mya) with pyrogenic flowering slightly younger because of its dependence on resprouting. There was no evidence that resprouting was always an older trait than either seed storage or non-sprouting throughout this period, with either/both ancestral or derived in different clades and times. Fire-adapted traits evolved slowly in the Cretaceous, 120-65 Mya, and rapidly but fitfully in the Cenozoic, 65-0 Mya, peaking over the last 20 My. The four trait-types climaxed at different times, with the peak in resprouter speciation over the last 5 My attributable to fluctuating growing conditions and increasing savanna grasslands unsuitable for non-sprouters. All experienced a trough in the 40-30-Mya period following a reduction in world temperatures and oxygen levels and expected reduced fire activity. Thick bark and serotiny arose in the Mid-Cretaceous among extant Pinaceae. Heat-stimulated germination of hard seeds is ancestral in the 103-My-old Fabales. Smoke-(karrikin)-stimulated germination of non-hard seeds is even older, and includes the 101-My-old Restionaceae-Anarthriaceae. A smoke/karrikin response is detectable in some fire-free lineages that prove to have a fire-prone ancestry. Among clades that are predominantly fire-prone, absence of fire-related traits is the advanced condition, associated either with increased fire frequency (loss of serotiny and soil storage), or migration to fire-free habitats (loss of thick bark, pyrogenic flowering, serotiny or soil storage). Protea (Africa) and Hakea (Australia) illustrate the importance of stabilisation processes between resprouting/non-sprouting in accounting for speciation events over the last 20 My and highlight the frequent interchange possible between these two traits. Apart from Pinus, most ancestral trait reconstruction relative to fire has been conducted on predominantly Southern Hemisphere clades and this needs to be redressed. Despite these limitations, it is clear that fire has had a profound effect on fire-related trait evolution worldwide, and set the platform for subsequent evolution of many non-fire-related traits. Genetics of the triggering mechanisms remain poorly understood, except the karrikin system for smoke-stimulated germination. We exhort biologists to include fire-proneness and fire-related traits in their thinking on possible factors controlling the evolution of plants.