Differences in Soil Seed Bank of Heracleum mantegazzianum and H sosnowskyi

Dynamics of the seed banks of both species were studied in the field by sampling three times a year and in a common garden burial experiment, where sampling was carried out repeatedly in the course of the year. Because the results for H. mantegazzianum have been published in detail elsewhere (Krinke et al., 2005), the following account focuses on H. sosnowskyi, and is based on primary data and a comparison of both species.

Seed bank dynamics and composition

To obtain data on seed bank composition and dynamics comparable to those that are available for H. mantegazzianum (Krinke et al., 2005), a similar study, using the same methods, was carried out at sites dominated by H. sosnowskyi in its invaded range (Table 10.1). Soil samples were taken in spring before seed germination (April), summer before seed release (July) and autumn after seed release (October). H. mantegazzianum was studied at seven sites in the Czech Republic (Krinke et al., 2005), H. sosnowskyi at three sites in Lithuania. The geographical location, altitude and characteristics of the H. sos-nowskyi populations are given in Table 10.1.

The vertical distribution of seeds in the soil seed bank is similar for both species. In the spring sample of H. sosnowskyi, 98.2% of the total seed, including dead seeds, are in the upper soil layer of 0-5 cm, with little in the deeper layers of 6-10 cm (1.5%) and 11-15 cm (0.3%) (Fig. 10.1). Nevertheless, no living seeds were found in the deepest soil layer (11-15 cm). The vertical distribution of living and dead seeds also varied significantly within

Table 10.1. Geographical location, altitude and characteristics of populations of H. sosnowskyi at three sites in Lithuania.

Density of

Mean

flowering

plant

Latitude

Longitude

Population

Year of

plants/m2

height

N

E

Altitude

size (m2)

invasion

(2003)

(m)

Santariskes

54°44'55.7"

25°16'39.9"

191

4560

1987

1.1

3.50

Bajorai

54°45'14.6"

25°15'25.0"

182

1452

1990

0.4

3.21

Visoriai

54°45'07.9"

25°16'06.8"

183

9640

1989

0.9

2.57

H. sosnowskyi H. mantegazzianum

H. sosnowskyi H. mantegazzianum

Fig. 10.1. Vertical distribution in the soil in spring of living (empty bars) and dead (black bars) seeds of H. sosnowskyi. Deletion tests (Crawley, 2002) on square-root + 0.5 numbers of seeds indicate that the seeds occur mainly in the upper soil layer (0-5 cm) (P < 0.001), and the numbers in the 6-10 cm and 11-15 cm soil layers do not differ (P < 0.001). Numbers of seeds per core sample (49.8 cm2) are shown; horizontal lines are standard errors of means. Corresponding figures for H. mantegazzianum are based on data from Krinke et al. (2005). Note different scales are used for the two species.

Fig. 10.1. Vertical distribution in the soil in spring of living (empty bars) and dead (black bars) seeds of H. sosnowskyi. Deletion tests (Crawley, 2002) on square-root + 0.5 numbers of seeds indicate that the seeds occur mainly in the upper soil layer (0-5 cm) (P < 0.001), and the numbers in the 6-10 cm and 11-15 cm soil layers do not differ (P < 0.001). Numbers of seeds per core sample (49.8 cm2) are shown; horizontal lines are standard errors of means. Corresponding figures for H. mantegazzianum are based on data from Krinke et al. (2005). Note different scales are used for the two species.

the individual sites (Table 10.2). The results for H. mantegazzianum show the same trend, with 95% of seeds in the upper soil layer (Fig. 10.1) and significant differences within sites (Krinke et al., 2005, Table 6).

As the vast majority of the seeds are located in the upper soil layer, the study of the seasonal dynamics of the seed bank was based only on samples taken from the 0-5 cm layer. The variation among study sites of H. sos-nowskyi was significant for all seed groups (dormant, living and total, i.e. their sum). For the total seed bank of H. sosnowskyi, averaged across spring, summer and autumn samples, 31.7% of the variation was linked to among sites and 68.3% to within sites (Table 10.3). If compared with results for H. mantegazzianum (77.9% of variation attributed among and 22.1% within sites; Krinke et al., 2005), these figures indicate that the Lithuanian sites in which H. sosnowskyi was studied were less heterogeneous than the H. man-tegazzianum sites sampled in the Slavkovsky les, Czech Republic.

The composition of the H. sosnowskyi seed bank in the course of the season, expressed as the numbers of non-dormant, living and total seeds,

Table 10.2. Nested anova of the vertical distribution of H. sosnowskyi seeds in the soil (layers: 0-5, 6-10, 11-15 cm) in spring. Data were transformed to square root numbers + 0.5 of living and dead seeds. Layer is evaluated as a fixed effect. *** P < 0.001, NS - not significant.

Source of variation

Living

Dead

df

MS

F

df

MS

F

Layer

2

323.12

51.402***

2

175.23

31.409***

Sites within layers

6

6.286

6.257***

6

5.579

5.788***

Replicates within sites

81

1.005

81

0.964

Table 10.3. anova of the soil seed bank of H. sosnowskyi among sites and within sites. Data are log transformed numbers plus 0.5 of the dormant, living and the total seeds, averaged for autumn, spring and summer samples. Sites are evaluated as random effects and variance is expressed in percentages. *** P < 0.001, * P < 0.05.

Dormant Living Total Source of _ _ _

variation df MS F Variance df MS F Variance df MS F Variance

Among sites 2 2.450 9.301*** 40.8 2 0.715 5.895* 30.4 2 0.783 6.263* 31.7 Within sites 27 0.263 59.2 27 0.121 69.6 27 0.125 68.3

showed significant differences among the spring, summer and autumn samples and varied significantly within individual sites (Table 10.4).

Average numbers of non-dormant, living and total seeds in the seed bank of H. sosnowskyi were highest in spring and lowest in summer (Fig. 10.2).

lOO BO

30 SO

Spring

Spring

C B Summer

SO 70

Living

30 20

Spring ao 30 30 lO

Qi cC

ao 30 30 lO

Fig. 10.2. Seasonal dynamics of the H. sosnowskyi seed bank (empty bars), inferred from autumn, spring and summer samples. Bars are mean numbers of dormant, non-dormant, living, dead and total seeds. Each value is pooled across three localities and ten replicates within each locality. Vertical lines are standard errors of the means. Bars with the same letters did not differ significantly (P < 0.05) in deletion tests (Crawley, 2002); capital letters refer to H. sosnowskyi and lower case letters to H. mantegazzianum. Corresponding values for H. mantegazzianum (black bars) are from Krinke et al. (2005). Germinated seeds were considered as non-dormant; non-germinated seeds were tested for viability by staining with tetrazolium; viable seeds were considered as dormant.

Fig. 10.2. Seasonal dynamics of the H. sosnowskyi seed bank (empty bars), inferred from autumn, spring and summer samples. Bars are mean numbers of dormant, non-dormant, living, dead and total seeds. Each value is pooled across three localities and ten replicates within each locality. Vertical lines are standard errors of the means. Bars with the same letters did not differ significantly (P < 0.05) in deletion tests (Crawley, 2002); capital letters refer to H. sosnowskyi and lower case letters to H. mantegazzianum. Corresponding values for H. mantegazzianum (black bars) are from Krinke et al. (2005). Germinated seeds were considered as non-dormant; non-germinated seeds were tested for viability by staining with tetrazolium; viable seeds were considered as dormant.

Table 10.4. Nested anovas of the variation of the soil seed bank of H. sosnowskyi among seasons. Data are square rooted numbers + 0.5 of the dormant, non-dormant, living, dead and the total of seeds. Season is evaluated as a fixed effect. *** P < 0.001, * P < 0.05.

Source of variation

Dormant

Non-dormant

Living

Dead

Total df MS

df MS

df MS

df MS

Season

Sites within season

Replicates within sites

2 1.163 0.269 n.s. 2 193.50 33.292*** 2 134.92 15.213*

81 0.554

81 1.360

81 1.691

81 1.792

6 1.133

81 0.265

Table 10.5. Number of seeds of H. sosnowskyi per m2 in the soil seed bank at the three localities studied. Each value is the mean ± sd of ten replicates. values per m2 were extrapolated from the original data, which were used in statistical analyses.

Spring

Locality Non-dormant Dormant

Dead

Total

Summer

Autumn

Non-dormant Dormant

Dead

Total

Non-dormant Dormant

Dead

Total

1 11,477 ±6,744 2,312 ± 1,663 9,407 ±5,937 23,195 ± 13,859 462 ± 403 1,286 ± 753 3,055 ± 2,175 4,804 ±2,835 4,502 ± 1,254 704 ±437 4,040 ± 2,241 9,246 ±3,501

2 6,512 ±3,532 503 ±977 4,201 ±3,543 11,216 ±7,032 161 ± 208 643 ± 388 1,809 ± 957 2,613 ± 1,180 3,437 ± 1,771 402 ±268 2,854 ± 2,348 6,693 ±4,120

3 5,970 ± 1,700 362 ±247 3,899 ± 1,283 10,231 ±2,702 141 ± 269 663 ±519 3,819 ± 2,127 4,623 ±2,720 6,613 ± 4,582 543 ±391 3,980 ± 1,805 11,135 ±6,363 Total 7,986 ±3,035 1,059 ± 1,087 5,836 ±3,097 14,881 ±7,217 255 ± 180 864 ± 366 2,894 ±1,014 4,013 ± 1,216 4,851 ± 1,616 550 ± 151 3,625 ± 668 9,025 ±2,229

This seemingly contradicts the pattern found for H. mantegazzianum, where most are found in autumn, and then the number of dormant, living, dead and total seeds decreases (Fig. 10.2). This discrepancy seems to be a result of differences in the sampling regime. The seed bank in Lithuania was sampled over a single season (from spring to autumn), while in the Czech Republic it was over two subsequent seasons (from autumn to summer the following year). Thus, the pattern in total numbers of H. sosnowskyi seeds reflects between-year fluctuations in the number and density of flowering plants. This explains why more seeds were found in spring than after seed set in autumn (Fig. 10.2); the seeds present in spring and summer were produced in the previous year.

To avoid the bias caused by the different sequence of sampling times, the percentages of non-dormant, dormant and dead seeds were compared (Fig. 10.3A). The percentage of living seeds of H. sosnowskyi in the total seed bank did not change from autumn (59.8%) to spring (60.8%), but decreased to 22.8% in summer. The percentage of non-dormant seeds among living seeds was similar in autumn (89.8%) and spring (88.3%), but decreased to 27.9% in summer (Fig. 10.3B).

The main difference between the species is in the autumn seed bank, when almost all the seeds of H. mantegazzianum were dormant and nearly 90% of H. sosnowskyi seeds were non-dormant (Fig. 10.3B). It needs to be noted that due to unusual climatic conditions in Lithuania in the year of sampling, ripe seeds were covered by early snow. Thus, the autumn sample was taken after the snow had melted. This short period of wet and cold conditions might have been enough to stratify the seeds and break their dormancy. An easy breaking of seed dormancy in autumn accords well with the laboratory finding that seeds of H. sosnowskyi require a shorter period of cold stratification for breaking dormancy than those of H. mantegazzianum (see below).

The average density of H. sosnowskyi seeds, expressed per m2 and pooled across localities, was 9025 ± 2229 (mean ± sd) in autumn, 14,881 ± 7217 in spring and 4013 ± 1216 in summer for total seeds, and 5400 ± 3281, 9045 ± 6411 and 1119 ± 889, respectively, for living seeds (Table 10.5). For H. mantegazzianum, it was 6719 ± 4119 total seeds in autumn, 4907 ± 2278 in spring and 1301 ± 1036 in summer (Krinke et al., 2005).

Seed bank depletion

Field data on changes in the seasonal dynamics of the seed bank provide important information on the strategy of alien species in terms of population regeneration and competition with native taxa (Van Clef and Stiles, 2001). The results, however, can be biased by factors beyond an investigator's control, such as the seasonal variation in the weather and the fact that the amount of seeds entering the soil is not known precisely. Burial of controlled numbers of seeds and the monitoring of their germination on a fine temporal scale can provide more reliable information on the temporal pattern of seed bank depletion. Moravcova et al. (2006) record the fate of buried seeds of

A

Was this article helpful?

0 0

Post a comment