Global variation in seed covering structure hardness of woody species with orthodox seeds.

Abstract

Background and aims: Seed covering structure hardness could play a role in defence/predation, physical dormancy and in situ longevity/persistence. However, research to date has been limited regarding quantification methods, plant diversity and geographic distribution. In this study, we determined global variation in seed covering structure hardness of woody species with desiccation tolerant seeds and analysed its relationships with relevant climatic variables, seed traits and ecological processes.

Methods: We measured seed covering structure hardness of 476 species from 459 genera and 113 families using puncture force. We used phylogenetic-informed regressions to test covering structure hardness against potential quantitative predictors (19 climate variables [n=405], 10 seed morphological traits [n=413], elevation [n=405], genus age [n=375]) and response variables (ex situ seed longevity [n=67], germination rate [n=82], species distribution/range size [n=403]). Categorical predictors (geographical region [n=444], plant lifeform [n=428], seed dormancy type [n=146], seed physical dormancy in the Fabaceae family [n=76], dispersal unit or mechanism [n=484], fruit type [n=427]) were tested using pairwise comparisons.

Key results: Seed covering structure hardness ranged from 0.13 to 366.38 N and seed and fruit (seed/fruit) size, seed/fruit roundness, seed/fruit colour (lightness) and precipitation of the driest quarter were significantly associated with hardness. In addition, dormancy types (vs non-dormancy), dispersal as fruit (vs seed), or certain fruit types (fleshy vs dry, drupes vs other types), as well as animal dispersal (vs other mechanisms) showed higher levels of hardness. Furthermore, covering structure roundness was higher in animal dispersed seeds/fruits (vs other dispersal strategies). Finally, covering structure hardness was shown to predict germination rate but not ex situ seed longevity or species range size.

Conclusions: Our results suggest roles for morphology, dormancy, dispersal and precipitation in explaining part of the global variation in seed covering structure hardness of woody species with orthodox seeds. However, we showed that the presence of physical dormancy does not always imply having a harder covering structure than non-dormant seeds and therefore terms like "hardseeded" or "hard coat" should no longer be used as synonyms for this trait.