Can Fertilizers Increase the Seed Yield of Two Native Herb Species in the Subarctic? Implications for Wild Seed Collection

Can Fertilizers Increase the Seed Yield of Two Native Herb Species in the Subarctic? Implications for Wild Seed Collection

Brittany Rantala-Sykes (corresponding author: Department of Biology, Vale Living with Lakes Centre, Laurentian University, Sudbury, Ontario, P3E 2C6, Canada, bgrantala sykes{at}gmail.com) and Daniel Campbell (School of the Environment, Vale Living with Lakes Centre, Laurentian University, Sudbury, Ontario, P3E 2C6, Canada).

Local native seed is collected from the wild to restore disturbed land when local commercial seed sources are unavailable. What if wild local populations have low seed output? Low seed yields can result from: 1) pollen limitation and pollinator scarcities, especially for self-incompatible species (Burd 1994, Ashman et al. 2004); and 2) abiotic limitations, including resources such as light, water, and nutrients (Stephenson 1981). Pollen limitation and climate are difficult to circumvent. Managers may more easily manipulate certain resources, such as nutrients. In boreal and arctic biomes, N and P are the most commonly limiting nutrients (Shaver and Chapin 1995, Weintraub 2011), but fertilization studies have found variable and species-specific responses on flowering or seed output (Shaver and Chapin 1995, Grainger and Turkington 2013, Petraglia et al. 2013). In this study, we posed the question whether fertilizer addition in a natural subarctic ecosystem over two growing seasons could increase the seed yields for desirable restoration species.

We conducted our study in north-central Canada along a 1.5-km south-facing section of the Attawapiskat River floodplain (52.88° N, 83.91° W, 83 m elevation). The soils were calcareous, alkaline and poorly-developed, with low N and P within primary rooting zones (Garrah 2013). The vegetation was open as a result of annual river ice damage, dominated by herbs and scattered shrubs. We selected two perennial herbs, Potentilla anserina (silverweed), a spreading ground cover, and Vicia americana (American vetch), a N-fixing legume. Both species are self-incompatible and insect-pollinated (Gunn and Kluve 1976, Eriksson 1987, Miyanishi et al. 1991) and are regionally common but have low seed output. We hypothesized that fertilization would increase flowering, seed set, and total seed yield after two seasons and that V. americana would not need N amendment because it is a N-fixing species.

For each species, we set up ten blocks of 1-m² plots, containing one plant of either P. anserina or V. americana. In June 2015 and again in June 2016, we applied two treatments (control and NPK fertilization) to P. anserina and three treatments (control, PK, and NPK fertilization) to V. americana. We used urea [CO(NH₂)₂], superphosphate [Ca(H₂PO₄)₂] and potash [KCl (95%) NaCl (5%)] at a rate that contributed 4 g m^-2 each of N, P, and K. We watered plots after fertilization. In August 2016, we assessed: (i) the number of ramets; (ii) the number of flowers per ramet; and (iii) the number of flowers per plot. We tested for significant differences among treatments for each dependent variable using a permutation test with 10,000 random shuffles, using Resampling Stats^® (Excel add-in v. 4.0, Resampling Stats, Arlington, Virginia). For P. anserina, we resampled the difference between the two treatment means, and for V. americana, we resampled the F statistic and then the pairwise difference in the three means.

For P anserina, we found no significant differences between the two treatments in the total number of ramets per plot (mean difference = 3.07; n_unfert = 9, n_fert = 8, p = 0.28; Figure 1A), but a borderline difference in the total number of flowers per plot (difference = 4.04, p = 0.08; Figure 1B). None of the P. anserina ramets set seed. For V. americana, we found significantly more ramets in the NPK treatment compared to the control (F_{2, 25} = 3.00, p = 0.025), and the PK treatment was intermediate (Figure 1C). Only three of all the V. americana plots produced flowers (Figure 1D), and only one plot produced seed. Although we found some differences, we were unable to increase seed output for either species, despite two years of spring fertilization.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Box plots of the total numbers of: A) ramets per plot; B) flowers per plot for Potentilla anserina; C) ramets per plot; and D) flowers per plot for Vicia americana. Treatments are control with no fertilization; PK with phosphorus and potassium fertilizers at a rate of 4 g m^-2; and NPK with nitrogen, phosphorus and potassium fertilizers at the same rate. Lowercase letters indicate significantly different means based on pairwise permutation test on mean differences (p ≤ 0.05; 10,000 random shuffles).

Eriksson (1987) found pollen, rather than nutrients, limited seed set in a northern Sweden population of P. anserina, but that population had higher plant and flowering densities than those tested here. Relatively low P. anserina flower and plant density in our plots may have reduced pollinator attraction and the amount of compatible pollen sources (Ashman et al. 2004,Waites and Ågren 2004), potentially leading to poor seed set in our study.

For V. americana, which had significantly more ramets in NPK fertilized than control plots, the newly-introduced nutrients appear to have been allocated to vegetative growth before sexual reproduction; a common trade-off allocation strategy among perennial plants (Wilson and Thompson 1989, Obeso 2002). The single V. americana plot that we found producing seed was growing among N-fixing Alnus incana ssp. rugosa (speckled alder) and was observably larger than other plants within our plots. Outside of our experiment, we observed that other V. americana plants growing with Alnus produced seed. Besides N-fixation, Alnus shrubs may allow vetch to climb and reach a larger size and offer root protection during annual ice break up.

The short growing season and the relatively short timeframe of our study (two years) may have limited the impact of fertilization. Furthermore, the alkaline soil at our study site may also have adsorbed added nutrients, especially phosphorus (Weintraub 2011), limiting nutrient availability. However, these climatic or soil factors are difficult or impossible to manipulate. We also may not have fertilized sufficiently. Other fertilization studies reported mixed results, sometimes with increased vegetative biomass instead of sexual reproduction, more competition from surrounding plants, or even a shift to more competitive species (Daws et al. 2013, Petraglia et al. 2013). In northern climates, adding nutrients, even over the long term, may not shift a species reproductive behaviour to increase seed yield, perhaps due to the climate restrictions or the specific reproductive strategy of that species (Grainger and Turkington 2013). These responses make it difficult to predict whether increasing the length of our study or increasing fertilizer rates would have eventually increased our seed yield for these species.

For restoration projects requiring local wild seed sources from plants with low seed outputs, using fertilizers to increase seed yield in a natural population may not be effective. We know both of these species produce seeds in other subarctic environments. This suggests that cultivating local plants under optimal conditions to produce seed is a better strategy to increase seed yield in a reasonable time frame and maintain local genetic provenance.