The effect of seed-dispersal timing on seedling recruitment is modulated by environmental conditions that vary across altitude in a threatened palm

Abstract

Background and aims: The timing of seed dispersal determines the environmental conditions that plants face during early life stages. In seasonal environments, selection is expected to favour dispersal timing that is matched to environmental conditions suitable for successful recruitment. Our aim here was to test whether the timing of seed dispersal influences seedling establishment success in two populations of Euterpe edulis that are located at contrasting altitudes, have different seed-dispersal phenologies and are subjected to distinct climatic conditions.

Methods: We sowed E. edulis seeds in contrasting altitudes on different dates, and monitored seed germination, emergence and seedling establishment at each altitude over 4 years. At the high-altitude site, five seed-dispersal cohorts were established during the natural dispersal period. At the low-altitude site, three seed-dispersal cohorts were established during natural dispersal, and two were established either before or after natural dispersal.

Key results: At the high-altitude site, seed-dispersal timing did not affect seed germination, seedling emergence or seedling establishment success. In contrast, at the low-altitude site, late seed dispersal near the end of the wet season resulted in a lower probability of seedling establishment, possibly due to the exposure of seeds, germinants and seedlings to unfavourable drought conditions. In addition, at the low-altitude site, the natural seed-dispersal period was poorly matched to favourable environmental conditions for seedling establishment.

Conclusions: The greater effect of seed-dispersal timing on seedling establishment at the low-altitude site is probably related to a more seasonal and drought-prone environment that favours a restricted period of seed dispersal. The magnitude of the effect of dispersal timing on seedling establishment success was modulated by environmental conditions that vary across altitude. Furthermore, reproductive phenology appears to be subject to more intense selection at the lower limit of the altitudinal range, due to a more restrictive window of opportunity for successful seedling establishment.

Keywords: Euterpe edulis; Dispersal timing; altitudinal gradient; divergent selection; establishment; germination; palm; phenology.

Fig. 1.

Ombrothermic diagram for each altitude. Bars represent the mean monthly precipitation (left axis), and lines represent the mean monthly temperature (right axis) during the study period, from 2012 to 2015. The horizontal lines represent the natural seed dispersal period at each altitude during the study period.

Fig. 2.

Proportion of germinated seeds of E. edulis (mean ± s.e.) for each seed-dispersal cohort (sown at different dates) at both altitudes. Germination proportion was calculated by dividing the number of germinated seeds by the total number of seeds planted (N = 200 per seed-dispersal cohort per site). The x-axes represent the date of each census at each altitude. Dashed arrows indicate the date of sowing.

Fig. 3.

Proportion of dead germinated seeds of E. edulis (mean ± s.e.) for each seed-dispersal cohort at both altitudes. The proportion of mortality was calculated by dividing the number of dead germinants by the initial number of germinated seeds over each period. The x-axes represent the date of each census at each altitude. Dashed arrows indicate the date of sowing. For high altitude: 12 May 2012, N = 99; 12 July 2012, N = 125; 31 October 2012, N = 118; 6 December 2012, N = 121; 7 March 2013, N = 97. For low altitude: 5 June 2012, N = 117; 24 August 2012, N = 158; 25 October 2012, N = 122; 7 December 2012, N = 132; 21 February 2013, N = 74.

Fig. 4.

Proportion of seedling emergence of E. edulis (mean ± s.e.) for each seed-dispersal cohort in both altitudes. Seedling emergence proportion was calculated by dividing the number of emerged seedlings by the total number of seeds planted (N = 200 per seed-dispersal cohort per site). The x-axes represent the date of each census at each altitude. Dashed arrows indicate the date of sowing.

Fig. 5.

Proportion of seedling mortality of E. edulis (mean ± s.e.) for each seed-dispersal cohort in both altitudes. The proportion of mortality was calculated by dividing the number of dead seedlings by the initial number of emerged seedlings over each period. The x-axes represent the date of each census at each altitude. Dashed arrows indicate the date of sowing. For high altitude: 12 May 2012, N = 62; 12 July 2012, N = 68; 31 October 2012, N = 74; 6 December 2012, N = 68; 7 March 2013, N = 50. For low altitude: 5 June 2012, N = 62; 24 August 2012, N = 99; 25 October 2012, N = 68; 7 December 2012, N = 38; 21 February 2013, N = 12.

Fig. 6.

Proportion of seedling death caused by different agents at each altitude. The results are shown for each seed-dispersal cohort. At the high-altitude site: 1 = May 2012; 2 = July 2012; 3 = October 2012; 4 = December 2012; 5 = March 2013. At the low-altitude site: 1 = June 2012; 2 = August 2012; 3 = October 2012; 4 = December 2012; 5 = February 2013.