Fire-mediated germination syndromes in Leucadendron (Proteaceae) and their functional correlates

Abstract

A mechanistic understanding of fire-driven seedling recruitment is essential for effective conservation management of fire-prone vegetation, such as South African fynbos, especially with rare and threatened taxa. The genus Leucadendron (Proteaceae) is an ideal candidate for comparative germination studies, comprising 85 species with a mixture of contrasting life-history traits (killed by fire vs able to resprout; serotinous vs geosporous) and seed morphologies (nutlets vs winged achenes). Individual and combined effects of heat and smoke on seed germination of 40 species were quantified in the laboratory, and Bayesian inference applied to distinguish biologically meaningful treatment effects from non-zero, but biologically trivial, effects. Three germination syndromes were identified based on whether germination was dependent on, enhanced by, or independent of direct fire cues (heat and smoke). Seed storage location was the most reliable predictor of germination syndromes, with soil-stored seeds c. 80% more likely to respond to direct fire cues (primarily smoke) than canopy-stored seeds. Notable exceptions were L. linifolium, with an absolute requirement for smoke to germinate (the third serotinous species so reported), and two other serotinous species with smoke-enhanced germination. Nutlet-bearing species, whether serotinous or geosporous, were c. 70% more likely to respond to fire cues than winged seeds, but there was no evidence for an effect of phylogeny or persistence strategy on germination. This comprehensive account of seed germination characteristics and identification of germination syndromes and their predictors, supports propagation, conservation and restoration initiatives in this iconic fynbos genus and other fire-prone shrubs with canopy or soil-stored seeds.

Fire Seasonality, Seasonal Temperature Cues, Dormancy Cycling, and Moisture Availability Mediate Post-fire Germination of Species With Physiological Dormancy

“…This is despite a growing number of reports of positive responses to a heat pulse in a diverse range of families with PD [e.g., Apiaceae ( Baker et al, 2005c ), Ericaceae ( Moreira et al, 2010 ), Lamiaceae ( Kazancı and Tavşanoǧlu, 2019 ); Myrtaceae ( Auld and Ooi, 2009 ), and Rutaceae ( Mackenzie et al, 2016 )]. Where both cues have been investigated in combination, responses are highly variable and species-specific, ranging from neutral to additive ( Keith, 1997 ; Kenny, 2000 ; Newton et al, 2021 ), unitive (both cues required ( Thomas et al, 2007 ; Mackenzie et al, 2016 ; Collette and Ooi, 2020 ); synergistic ( Gilmour et al, 2000 ; Baker et al, 2005a ); and negative (a heat pulse inhibits the smoke response without loss of viability; Keeley and Fotheringham, 1998 ; Collette and Ooi, 2017 ). Potential interactions between fire cues and environmental cues such as light ( Bell, 1994 ; Gilmour et al, 2000 ; Collette and Ooi, 2017 ) and moisture ( Thomas et al, 2010 ) create additional complexity.…”

Seed biologists beware: Estimates of initial viability based on ungerminated seeds at the end of an experiment may be error‐prone

“…During a study of the germination requirements of Leucadendron species in relation to alternating temperatures, smoke and heat (Newton et al . 2021), I noticed that estimated initial viability using this standard procedure varied greatly between some treatments in a number of species when they should have been at least non‐significantly different. Estimated viability declined the lower the level of germination, i.e .…”

“…A hypothetical scenario where the viability of the seeds that remain ungerminated under various treatments, especially the control, lost viability during the trial (based on unpublished data of Newton et al . 2021). The circles correspond to (idealized) data points that fit on the best‐fit curve.…”

Dealing with ‘the spectre of "spurious" correlations': hazards in comparing ratios and other derived variables with a randomization test to determine if a biological interpretation is justified

“…The relationship is invariably shown to be negative implying that small-seeded species have a significantly higher relative growth rate than large-seeded. Switching to Bayesian inference using uncorrelated data derived from the empirical data for comparisons can provide unique insights into the basis of any relationships (Newton et al 2021). Starting with 100 uncorrelated X and Y values drawn from five possible distribution types in nature and calculating lnX and ln[(X + Y)/X], we produced variable R values but they were all in the same (negative) direction and significant at p < 0.0001 (Table 1).…”

“…One can go further and take into account a standardized effect size on ΔR by dividing it by the standard deviation of R random (we only give the ranges here, Table 1, 2) (Ulrich andGotelli 2013, Perry et al 2020). This gives a better idea of the 'strength' of any biological relationship for comparison with other data sets and gives emphasis to the effect size rather than statistical significance as currently advocated (Newton et al 2021). Heed must also be paid into which tail of the randomization distribution the Table 2.…”