Ecological Archives E092-055-A4

Bitty A. Roy, Tim Coulson, Wilma Blaser, Tobias Policha, Julie L. Stewart, G. Kai Blaisdell, and Sabine Güsewell. 2011. Population regulation by enemies of the grass Brachypodium sylvaticum: demography in native and invaded ranges. Ecology 92:665–675.

Appendix D. Seed germination tests.

To determine germination rates, we performed three different experiments. The first experiment was designed to determine whether or not the field sites were seed limited, and also to determine whether there were soil or seed-borne pathogens that reduced germination in the field (seed collected Fall 2005 and planted Spring 2006). For this experiment, drought reduced seed availability in Switzerland, so we used commercial seeds from Switzerland and seeds from a single study site in the US (Pisgah). The second experiment was similar to the first, but was done in pots outside and included seeds collected from all US sites (seeds collected Fall 2005 and planted Spring 2006; same seed sources as Expt. 1). In the third experiment, we used seeds collected from all sites in both ranges (seed collected Fall 2007 and planted in Spring 2008). This was a pot experiment in a single common garden setting, with seeds from each field site planted into soil from invaded sites, and was designed to determine whether differences in germination among sites and ranges were due to differences in seed quality.

Methods

Experiment #1: We tested for seed or soil-borne pathogens by excluding fungi from the seeds with the fungicide Captan. We used seeds from a single source in each country because no 2005 seeds were available from the Swiss field sites due to a severe drought, so we purchased seeds to represent the native range (from Semences UFA Samen, Fenaco, Switzerland). To represent the invaded range, we used seeds that we collected in 2005 from the Mt. Pisgah site. Single seeds were glued with radicle facing downwards about 1/3 of the way up wooden toothpicks with water-soluble glue (school glue, RiteAid). Half the seeds were treated by dipping the toothpick+seed in Captan powder, and half were untreated controls. In each plot (both treated and controls) at each site, eight fungicide treated and eight control seeds were planted for a total of 96/site (48/seed treatment). The seeds were placed inside the main demography plots, near the outer edge, in a grid with seeds spaced 1 cm apart. Seedling emergence was monitored every two weeks when pesticide treatments were applied. We performed this experiment at all sites in Switzerland, but at only one of the sites in the US (Jasper), in both a typical shady habitat as well as a sunny one. Based on densiometer readings, light levels at the sunny site were most similar to Pisgah and Bald Hill, and those at the shady microsite were more like that at Sweethome (unpublished data). To determine whether there were effects of above ground fungicide treatment, below ground fungicide treatment or range, we used ANOVA on arcsine transformed percent germination with site as the replicates. An initial analysis indicated there was no above-ground fungicide effect, so the two fungicide treatments were merged to increase sample size.

Experiment #2: This experiment was similar to the first, but better replicated and in pots. We collected one Brachypodium sylvaticum plant from each of six plots at the eight field sites (Pisgah, Jasper, Bald Hill and Sweethome in Oregon, Albisgütli, Flaach, Galmiz and Hönggerberg in Switzerland). Plants were collected with their natural soil using a tulip bulber and soil and were placed in round pots. 12 untreated seeds were planted around the adult B. sylvaticum plant (three pots per treatment per site). We used the same seeds as in the Field Enemy Experiment [expt. 1]. In the native range, the seeds were planted at the end of March and the pots were randomly arranged and placed outside in Zürich, Switzerland. In the invaded range, seeds were planted in mid-May and the pots were randomly allocated and paced in partial shade in Eugene, Oregon, USA. All pots were watered daily. Seedling emergence was monitored for 25 days in the US and 79 days in Switzerland (Blaser 2008). To make the two experiments comparable, we truncated the data from Switzerland to make it a more similar number of days (25 d US and 27 d Switzerland; the small difference in number of days was due to not monitoring germination every day). To determine whether there were effects of treatment or range, we used ANOVA on arcsine transformed percent germination.

Experiment #3: The purpose of this experiment was to determine whether under a common garden environment the seeds from both ranges would germinate more similarly. Unlike the first two experiments, here we used seeds that were field collected from all sites in 2007. We planted 20 untreated seeds/site in soil composited from the four invaded sites in 2008. Germination was observed over a 67-day period. To determine whether there were range effects, we used ANOVA on arcsine transformed percent germination with site as the replicate.

Results and Discussion

Germination was variable (Table D1), but was higher in the invaded range than the native range across all experiments, and the difference was statistically significant in two out of three tests:

Expt. 1, invaded mean  ±  SE = 30  ±  4%, native mean = 14  ±  3%, F = 8.841,8, P = 0.0178;

Expt. 2: invaded mean  ±  SE = 36  ±  6%, native mean= 22  ±  6%, F = 3.831,7, P < 0.0986;

Expt. 3 invaded mean  ±  SE = 45  ±  16%, native mean = 8  ±  1%, F = 6.201,6, P = 0.0471.

The fungicide treatment in experiments 1 and 2 had no statistically significant effects on germination (respectively, F = 1.11 1,8, P = 0.3223; F = 0.55 1,7, P = 0.8173), nor were there significant interactions between treatment and range (respectively, F = 2.67 1,8, P = 0.1417; F = 3.89 1,7, P = 0.0720), indicating that the treatments had similar effects in both ranges. The lack of a treatment effect of the seed fungicide Captan suggests that germination was not significantly influenced either by pathogens in the soil or on the surface of the seeds. We do not know, however, whether the seed applied fungicide Captan affected the endophytes in the seeds, or whether endophyte infection itself may alter germination in some way.

In experiments one and two we examined the performance of seeds in soils from different sites, whereas in experiment three we examined the performance of seeds from different sites in the same soil. In all cases germination was better in the invaded range.

TABLE D1. % Germination ± SE for each experiment. Because there were no significant seed fungicide or foliar spray effects from Experiment 1 and 2, we present the means without separating them by treatment (to see the variation by plot and treatment within site, see Appendix C).

Range

Site

Expt. 1

In soil at field sites
2005 seeds in 2006

Expt. 2

In soil from all field sites
in pots outside in both ranges 2005 seeds in 2006

Expt. 3

In invaded soil in pots in a greenhouse
2007 seeds in 2008

Native

Albisguetli

13.5  ±  3.0

12.5  ±  2.8

 10  ±  8.2 a

Native

FlaachI

16.7  ±  2.6

4.2  ±  2.8

10  ±  8.2

Native

Höngerberg

11.5  ±  3.0

12.5  ±  5.2

10  ±  8.2

Native

Galmiz b

14.6  ±  2.5

5.6  ±  4.1

--

Native

Rafz c

--

--

05  ±  8.2

Invaded

Bald Hill

--

4.2  ±  1.9

10  ±  8.2

Invaded

Jasper

20.8  ±  4.0

27.8  ±  4.1

70  ±  8.2

Invaded

Jasper2, sun d

39.6  ±  6.0

--

--

Invaded

Pisgah

--

22.2  ±  8.2

75  ±  8.2

Invaded

Sweet Home

--

18.1  ±  7.3

25

a The standard errors are based on 4 replicate lots of 5 seeds each from each population.
b The Galmiz site was lost to logging in the winter of 2006.
c For the 2008 experiment we added in Rafz to make up for the loss of Galmiz to logging.
d The Jasper2 site was not in the main demography experiment, but was about 10 m away in an open area that was more similar in terms of light levels (unpublished data) to Bald Hill and Pisgah, and the main Jasper site was more similar to Sweet Home.

LITERATURE CITED

Blaser, W. 2008. Common gardens as a tool to address questions related to plant invasions. Masters Thesis. ETH (Swiss Federal Institute of Technology), Zürich, Switzerland.


[Back to E092-055]