Sulfur has been proposed as a useful element to employ in addition to carbon and nitrogen in stable isotope studies of marine food webs, but variability in δ 34S of primary producers may prevent food web resolution. δ 34S values in green leaves of the seagrass, Zostera capricorni, showed considerable variability (12.7–17.6‰) in a survey in Moreton Bay, Australia. We demonstrated that δ 34S values were correlated with sediment organic matter (OM) content and height of seagrass on the tidal gradient, but these relationships were opposite to those expected from work elsewhere. In our survey, δ 34S values were relatively depleted at sites higher on the shore and with lower OM content. We did find the expected relationship of depleted δ 34S values where sediment porewater sulfide concentrations were higher. Any influence of OM content on δ 34S values would have been confounded in the survey by the relationship between height on shore and OM content itself. We separated the effects of height and OM content by creating the following treatments at one height on the shore: (1) OM added, (2) procedural control, and (3) untouched control. δ 34S values of seagrass in OM added plots were significantly depleted (5.6‰) relative to procedural (10.1‰) and untouched (11.0‰) controls 8 weeks after the manipulation. This demonstrated that OM content on its own does have the expected effect on δ 34S values of seagrass, so in the initial survey another factor, probably related to height on shore, must have overridden the influence of OM content. Seagrass roots are able to exude excess oxygen produced during photosynthesis, reoxidising sulfides in surrounding porewater. We demonstrated that the above and below-ground biomass of seagrass was higher low on the shore, and contend that higher seagrass productivity low on the shore results in greater reoxidation of sulfides and leads to more enriched δ 34S values of seagrass.