Figure 2
Parts (a,b) Ammonium efflux (μmol m−2 h−1) versus assemblage type ▪ ▴ Data sources: Austrovenus35; Echinocardium46,52; Macomona53. Means and standard deviations of untransformed data for ammonium release from the sediments in the dark (part a) and log10 back transformed data for release in the light (part b). The dark efflux data had non-homogeneous variances and were non-normal, not corrected by transformation; so, the non-parametric Kruskal-Wallis rank sum test was used for statistical evaluation. A log10 transform corrected non-normality of the light flux data. Dark flux: Kruskal-Wallis test statistic = 6.23, df = 2, p < 0.05; light flux: ANOVA on log10 transformed values: F2,9 = 5.74, p < 0.05. Solid horizontal lines separated by a space and labelled with different letters indicate statistical significance by Holm-Šidák a posteriori tests, adjusted p < 0.05: dark efflux: Austrovenus:Macomona p = 0.5, Austrovenus:Echinocardium p < 0.05, Macomona:Echinocardium p < 0.05; light efflux: Austrovenus:Macomona p = 0.92, Austrovenus:Echinocardium p < 0.05, Macomona:Echinocardium p < 0.05.
