We regret that two errors were inadvertently introduced in the paper by Steudel et al. (2012); their correction required changes to Figs 2, S5 and S8 as well as to Tables 2 and 3:

Table 2. Empirical studies used for analyzing relationships between the relative stress-response buffering effect (rSRBE) and stress response intensity (SRI)
ReferenceTreatmentlog (aSRBE)log (aSRBE) without monoculturesrSRBErSRBE without monoculturesStress response intensity (SRI)
Allison 2004Low 1−0.416NA0.920NA0.73
Allison 2004Low 2−1.050NA−0.481NA0.48
Caldeira et al. 2005 −1.653NA1.375NA0.84
De Boeck et al. 2008A−0.916NA−0.222NA0.33
De Boeck et al. 2008B−1.050NA−0.291NA0.29
De Boeck et al. 2008C−0.223NA0.167NA0.40
De Boeck et al. 2008D−0.844NA−0.375NA0.25
De Boeck et al. 2008E−0.619NA−0.213NA0.20
De Boeck et al. 2008F0.890NA−0.663NA0.17
Downing & Leibold 2010Macrophyte−0.811NA−0.089NA0.37
Downing & Leibold 2010Herbivore−0.172NA0.470NA0.45
Dukes 2002New0.5131.3860.8180.0650.35
Dukes 2002Established0.0300.7150.109−0.0690.33
Goodsell & Underwood 2008Bungan Head0.932NA0.223NA0.21
Goodsell & Underwood 2008Narrabeen Head0.560NA0.129NA0.24
Hughes & Stachowicz 2004 1.887−0.2600.1960.1630.48
Ji et al. 2009Grasses 20040.194NA0.249NA0.34
Ji et al. 2009N-fixers 2004−1.444−1.641−1.327−0.2430.50
Ji et al. 2009N-fixers 2005−0.0070.7551.5891.3160.60
Ji et al. 2009Forbs 20050.087NA0.371NA0.22
Joshi et al. 2000 0.513−0.0652.9000.10.71
Langenheder et al. 2012 −0.074NA0.042−0.1560.14
Lanta et al. 20122003−0.239NA−0.0370.0530.12
Lanta et al. 201220040.125NA0.3710.4320.32
Lanta et al. 20122005−1NA−0.338−0.2560.13
Liiri et al. 2002 −0.083NA0.933NA0.47
Mulder et al. 2001 2.140NA0.553NA0.34
Nagase & Dunnett 2010Root−0.8271.0780.4872.0790.56
Nagase & Dunnett 2010Shoot0.0000.0470.0050.2120.29
Roger et al. 2012High temperatureNANA−0.725NA0.47
Roger et al. 2012Low salinityNANA−0.386NA0.53
Steudel et al. 2011Above-ground drought0.5730.5700.2700.2620.21
Steudel et al. 2011Above-ground salt0.6100.5250.2840.2250.20
Steudel et al. 2011Above-ground shade0.7720.7330.5430.4880.38
Steudel et al. 2011Below-ground drought0.5780.7880.7851.0940.49
Steudel et al. 2011Below-ground salt0.5010.5640.4950.4930.36
Steudel et al. 2011Below-ground shade−0.6750.9603.0188.0960.82
Table 3. Measurements of stress-response buffering effects (SRBE) and stress response intensity (SRI) of our algal experiment
Salinity or temperaturelog (aSRBE)log (aSRBE) without monoculturesrSRBErSRBE without monoculturesStress response intensity (SRI)
27.5 °C0.1510.2110.1780.1250.06
30.0 °C0.5380.5380.5360.6130.32
32.5 °C0.3110.7050.6311.0070.40
35.0 °C0.032−0.2800.2790.0860.45
37.5 °C−1.553−1.333−0.022−0.0690.65
40.0 °C−1.305−0.8890.0360.3070.72

Figure 2. Relationships between relative stress-response buffering effect (rSRBE) calculated without monoculture data and stress response intensity (SRI) derived from monoculture data. Black dots represent literature data, grey dots data from our algal experiment. Black triangles represent studies not included in the published paper. The black line is the regression line of a linear mixed model including published data only, with ‘study’ treated as random factor (intercept = −4.24, SE = 1.17, t = −3.63; slope = 10.36, SE = 1.98, t = 5.22; = 0.001, = 6). The grey line is the regression line of a linear mixed model including published data and those of the present experiment, with ‘study’ treated as random factor (intercept = −1.26, SE = 0.76, t = −1.66, slope = 4.45, SE = 1.52; t = 2.92, = 0.009, = 7). The dashed line is the regression line of a linear mixed model including all data (published, algae, and new studies), with ‘study’ treated as random factor (intercept = −1.02, SE = 0.58, t = −1.94, slope = 2.23, SE = 0.55, t = 4.07, = 0.0046, = 9).

Download figure to PowerPoint

  1. Our erroneous inclusion of stress response intensity (SRI) in the formula for calculating the relative stress-response buffering effect (rSRBE) positively biased the slope of the relationship between the relative biodiversity effect of the control treatment (rBEc) and the predicted slope at high stress intensities. The correct rSRBE is the difference between the relative biodiversity effect of the stress treatment (rBEs) and the rBEc (rBEs-rBEc). Reanalysis of the corrected data resulted in a positive relationship between rSRBE and SRI when the monoculture data are excluded in our mixed models analysis with literature and our own experimental data combined (Fig. 2, Table 2, Table 3) and with literature data only (Fig. 2, Table 2). However, reanalysis of only our experimental data revealed no significant relationship (Fig. S5A). With the monoculture data included, the relationship between rSRBE and SRI was positive for the literature data alone and in combination with our experimental data (Fig. S8). Again, reanalysis of our experimental data only revealed no significant relationship (Fig. S5B). Addition of three new case studies to the literature analysis with our experimental data included (Langenheder et al. 2012, Lanta et al. 2012, Roger et al. 2012) confirmed the positive relationship when monoculture data are excluded from the analysis (Fig. 2) or included in the analysis (Fig. S8). Thus, despite the miscalculation of the rSRBE in our original paper, the main conclusion of a positive relationship between rSRBE and SRI is essentially confirmed by the correct analysis. We thank Dr. T. Banitz for pointing out the mistake.
  2. In the figure of Box 1, the absolute stress-response buffering effect (aSRBE) was incorrectly defined as aBEc/aBEs instead of aBEs/aBEc, as was correctly stated in the main text. This typographical error had no consequences for the reported results or conclusions.


  1. Top of page
  2. References