Ecological Archives E096-143-A4

Hannah M. Griffiths, Julio Louzada, Richard D. Bardgett, Wallace Beiroz, Filipe Franҫa, Daniel Tregidgo, and Jos Barlow. 2015. Biodiversity and environmental context predict dung beetle-mediated seed dispersal in a tropical forest field experiment. Ecology 96:16071619. http://dx.doi.org/10.1890/14-1211.1

Appendix D. Unrecovered seed information and calculation of seed dispersion.

The assumption that seeds were buried outside of the 50 × 50 cm sampling area was based on observed burial depths of seeds within the plots. As only 6 seeds (0.38%) were found deeper than 40 cm it is unlikely that a significant proportion of the unrecovered seeds were within the sampling area but below 50 cm depth. Furthermore, seeds were generally buried progressively deeper, the further from the central dung pat they travelled (main text, Fig. 1). Seeds in the section three, therefore, were those most likely to be buried deeper than the 10cm limits of the plot edges.

Table D1. Number of seeds unrecovered from in each site and size class with values imputed to calculate seed dispersion.

Site

Seed

Number below 10cm

Number lost

Median depth of seeds below 10cm

Max depth in section 3

Imputed value

Clay

Medium

1

7

NA

20.5

15.25

Clay

Small

7

7

13

28

13

Clay

Very small

2

25

NA

28

19

Sand 1

Medium

25

17

17.5

37

17.5

Sand 1

Small

27

32

19

35.5

19

Sand 1

Very small

11

53

21

44

21

Sand 2

Medium

34

5

18

30

18

Sand 2

Small

44

57

22.75

46

22.75

Sand 2

Very small

32

81

21

41

21

 

Trials to assess seed removal by seed predators
Following completion of destructive sampling at site Sand 2, a trial was established to assess the possibility that unrecovered seeds were not found because of removal by seed predators (birds, other vertebrates, or invertebrates). Ten plastic pots containing 100g of dung and twenty beads from the four size classes were placed on the forest floor and left for 24 hours (the maximum time that experimental plots remained open during the study). Dung beetles could not the bury dung and beads because the base of the pots prevented excavation. After 24 hours dung was washed to locate the seeds. All 200 seeds from the ten pots were recovered, indicating that seed predators did not remove beads in experimental plots.

 

FigD1

Fig. D1. Associations between seed dispersion values calculated with and without imputed values for missing seeds. Bivariate plots (bottom left panel), distributions (top left and bottom right panels), and Pearson's ρ (top right panel).


 

 

Table D2. The influence of dung beetle community on seed dispersion when unrecovered seed depths were not imputed for the calculation of seed dispersion. Significant effects are highlighted in bold. Model results remain the same with and without the imputation of missing seed values with the exception of FEve, which changes from having no effect to significantly impacting on seeds dispersion.

 

Effect terms

Model tested

Beetle community

 

Site

 

Beetle community × site

LRT

df

P

 

LRT

df

p

 

LRT

df

P

gls(seed dispersion ~ effect terms

 

 

 

 

 

 

 

 

 

 

 

Species richness

0.00

1

0.9936

 

11.74

2

0.003

 

3.48

2

0.1757

Functional richness

0.61

1

0.4332

 

11.74

2

0.003

 

9.56

2

0.0084

Functional evenness

8.86

1

0.0029

 

11.74

2

0.003

 

1.04

2

0.5944

Functional divergence

2.86

1

0.0909

 

11.74

2

0.003

 

5.12

2

0.0773

Functional dispersion

1.50

1

0.2203

 

11.74

2

0.003

 

0.33

2

0.8493

Total biomass

3.58

1

0.0585

 

11.74

2

0.003

 

1.55

2

0.4617

 

FigD2

Fig. D2. The effect of FEve on seed dispersion in the three experimental sites when seed dispersion is calculated with (top panels) and without (bottom panels) imputation of missing seed values.


[Back to E096-143]