Decadal survival of tropical pioneer seeds in the soil seedbank is accompanied by fungal infection and dormancy release

Pioneer trees require high-light environments for successful seedling establishment. Consequently, seeds of these species often persist in the soil seed bank (SSB) for periods ranging from several weeks to decades. How they survive despite extensive pressure from seed predators and soilborne pathogens remains an intriguing question.

This study aims to test the hypotheses that decades-old seeds collected from the SSB in a lowland tropical forest remain viable by i) escaping infection by fungi, which are major drivers of seed mortality in tropical soils, and/or ii) maintaining high levels of seed dormancy and seed coat integrity when compared to inviable seeds.

We collected seeds of Trema micrantha and Zanthoxylum ekmanii at Barro Colorado Island, Panama, from sites where adult trees previously occurred in the past 30 years. We used carbon dating to measure seed age and characterized seed coat integrity, seed dormancy, and fungal communities.

Viable seeds from the SSB ranged in age from 9 to 30 years for T. micrantha, and 5 to 33 years for Z. ekmanii. We found no evidence that decades-old seeds maintain high levels of seed dormancy or seed coat integrity. Fungi were rarely detected in fresh seeds (no soil contact), but phylogenetically diverse fungi were detected often in seeds from the SSB. Although fungal infections were more commonly detected in inviable seeds than in viable seeds, a lack of differences in fungal diversity and community composition between viable and inviable seeds suggested that viable seeds are not simply excluding fungal species to survive long periods in the SSB.

Synthesis: Our findings reveal the importance of a previously understudied aspect of seed survival, where the impact of seed-microbial interactions may be critical to understand long-term persistence in the SSB.

Methods
Seeds were extracted from soil cores by rinsing the soil with tap water under a series of sieves, with a minimum pore diameter of 2 mm for Z. ekmanii and 1 mm for T. micrantha. Seeds were identified to species and then partitioned for tests of i) seed viability measured using the tetrazolium test and seed-endophyte isolation frequency (692 and 593 seeds of T. micrantha and Z. ekmanii, respectively), ii) germination (% of germinated and dormant seeds) and carbon dating (a total of 239 and 196 seeds of T. micrantha and Z. ekmanii, respectively, where 9 viable seeds of each species were carbon dated), and iii) average seed fracture resistance as a measure of seed coat integrity (224 and 142 seeds of T. micrantha and Z. ekmanii, respectively).