Desiccation tolerant yet short-lived seeds: A conundrum for post-harvest handling of a high restoration value bunchgrass?

Abstract

The successful preservation of seeds for future use depends on factors that maintain or limit seed viability. Yet, seed biology knowledge that would facilitate seed storage of most wild species used in ecosystem restoration is absent. This study characterized changes in seed water content, germination, desiccation tolerance, and relative storage longevity of Aristida beyrichiana (wiregrass), a focal restoration species, with respect to habitat of collection. We collected mature seeds from mesic and xeric habitats over two years then exposed these to desiccation stress sufficient for germplasm storage and aging stress (60% relative humidity, 45°C). We followed each method with germination assays at simulated seasonal temperatures. We analyzed germination responses along with production of normal and abnormal seedlings. We then modeled potential seed longevity and compared this against longevity of reference species. We found that wiregrass seeds display sufficient desiccation tolerance for ex situ storage and germinate preferentially under spring or fall and summer temperature conditions. The negligible to small effects of ecotype on these responses do not support the hypothesis that habitat of occurrence represents an adequate predictor of desiccation tolerance or germination response. However, seeds from both xeric and mesic habitats are estimated to be short-lived in storage. The contradiction between high desiccation tolerance yet short-lived nature of seeds implies that proper post-harvest seed handling, particularly regarding seed moisture management, is critical for maintaining seed viability. Further implications of this interesting seed physiology are discussed in the context of restoration.

Fig 1. Water content of wiregrass seeds collected in: (A) 2021 from two mesic (M01-21, M02-21) and one xeric (X-21) populations and (B) 2022 from one mesic (M-22) and one xeric (X-22) population.

For (A) all seeds were stored at room temperature (ca. 22°C, 50% RH) in sealed brown paper bags for six months until experiments commenced for the control (dark blue bars) and LiCl-dried (light blue bars) treatments. Seeds in the desiccation treatment were dried over a saturated LiCl solution (12% RH). For (B) all seeds were stored at room temperature until measurements were taken 2 months after harvest. Control seeds represented by dark blue bars. Seeds in the drying treatment (light blue bars) were dried over a saturated LiCl solution (11% RH).

Fig 2. Kaplan-Meier estimates of survivor functions for control and LiCl-dried wiregrass seeds germinating under 35/25°C.

All seeds were stored at room temperature (ca. 22°C, 50% RH) in sealed brown paper bags until experiments commenced after 2 and 6 months of laboratory storage for the control and dried treatments, respectively. Seeds in the desiccation treatment were dried over a LiCl saturated salt solution (12% RH) prior to germination. Dotted lines indicate 95% confidence limits. Circles represent censored observations.

Fig 3. Forest plots depicting linear orthogonal contrasts of wiregrass germination for seeds collected in (A) 2021 and (B) 2022.

Seeds were dried over saturated LiCl solutions (LiCl) or remained undried (Ctrl) prior to germination. In (A) X-21, M01-21, and M02-21 represent seed lots collected from xeric (X) or mesic (M) habitats during 2021. In (B) 28/15, 35/25, and 21/8°C represent simulated temperatures during the spring or fall, summer, and winter, respectively, Squares and horizontal lines represent values for hazard ratios and confidence limits, respectively. The vertical black line represents the reference line equal to 1.00 for determination of statistical significance. Confidence limits bracketing the reference line denote a non-significant comparison. Yellow and black squares also denote significant and non-significant hazard ratios, respectively. Squares to the right of the reference line denote that the first factor in a comparison displayed a greater likelihood of germination than the second factor. Squares to the left of the reference line denote that the second factor in the comparison displayed a higher likelihood of germination. Taking the reciprocal 1/HR of values less than 1 provides a comparable hazard ratio.

Fig 4. Kaplan-Meier estimates of survivor functions for control and LiCl desiccated wiregrass seeds from two sources (M-JUL-22 and X-JUL-22) germinated under simulated seasonal temperatures representing (A–B) spring or fall (28/15°C), (C–D) summer (35/25°C), and (E–F) winter (21/8°C) in north-central Florida.

Seeds in the desiccation treatment were dried to 11% RH via LiCl saturated salt solution prior to germination. Circles represent censored observations. Ninety-five percent confidence limits omitted for clarity.

Fig 5. Deterioration curves for wiregrass seeds harvested in (A) 2021 from mesic (M01-21, M02-21) and xeric (X-21) sites then incubated at simulated summer (35/25°C) temperatures after controlled deterioration treatments or (B) 2022 from mesic (M-22) or xeric (X-22) sites then incubated at simulated summer (35/25°C), spring or fall (28/15°C), and winter (21/8°C) temperatures following controlled deterioration treatments.

The p₅₀ values denote the time in days for germination to decline to 50%.