Appendix A. Herbivore density in enriched and ambient nutrient conditions at two sublittoral depths.
Determining herbivore abundance
We estimated invertebrate densities to evaluate the intensity of herbivory and the possible variation in grazer abundance and species composition with respect to nutrient availability and depth. We placed tiles, similar to those used for colonization substrates for algae, at a depth of one (n = 12) and three (n = 10) meters on 26th June. The tiles at these two depths had similar nutrient enrichment treatments (control n = 8, low n = 7 and high n = 7) as in the algal colonization cages (see above). While collecting the tiles by scuba diving, grazer escape was prevented by quickly enclosing each tile underwater in a mesh-net pouch. We collected about half of the tiles (n = 12) at the end of July and the rest (n = 10) in early September, and identified and counted the invertebrates on each tile. The design was a 2 (depth) × 3 (nutrient availability) × 2 (date) factorial; the total number of tiles was 22. Since these tiles were underwater for a shorter time than the substrates for measuring algal colonization, they had less algal biomass on them, and the invertebrate densities are likely to represent minimum estimates. We excluded the bivalves from further analysis because they are not macroalgal consumers.
We analyzed the abundances of different groups of invertebrates using generalized linear models. We used SAS GENMOD procedure in the data analysis (SAS Institute 1999). Our data fitted well to the negative binomial distribution and we used logarithmic link function that models the log of the mean. Depth, nutrient enrichment and date were treated as grouping factors. Because separate tiles were used to sample animals at both dates, each tile formed an independent replicate.
Patterns of herbivore abundance and species composition
The pooled density over all grazer species varied temporally, being higher in early September than in late July (χ21 = 5.41, P < 0.05; Table A1). Furthermore, separate taxa showed variations in abundance with respect to both the depth and nutrient availability. Gastropods were the dominant grazer group at our study site (Table A1). The abundance of hydrobid species (H. ulvae, H. ventrosa, and P. antipodarum) was an order of magnitude greater than any other group and differed between the two depth levels; throughout the summer these snails were clearly more abundant at the three-meter depth (means [-SE, +SE]: 104 [-13, +15] and 595 [-81, +93] ind. 200/cm2 at the 1- and 3-m depth respectively; Hydrobidae: χ21 = 33.14, P < 0.0001; Table A1). They also responded to nutrients, although differently at the two depths; the density was the lowest in the shallow-water ambient concentration (37 [-7, +11] ind. 200/cm2) and the highest in the deep-water low enrichment treatment (790 [-187, +243] ind. 200/cm2; depth × nutrient pooled over sampling dates: χ22 = 6.65, P < 0.05; Table A1). Likewise the abundance of the gastropod T. fluviatilis depended on both depth and nutrient treatment: in deep water it was most abundant in the low enrichment (5 [-1, +2], 22 [-5, +7], and 8 [-2, +3] ind. 200/cm2 in ambient concentration, low, and high enrichments respectively), while in shallow water it was most abundant in the ambient concentration (32 [-6, +8], 13 [-3, +4], and 5 [-1, +2] ind. 200/cm2 in ambient concentration, low, and high enrichments respectively; depth × nutrient, pooled over sampling dates: χ2 2 = 19.88, P < 0.0001; Table A1).
The densities of the crustaceans, dominated by Gammarus spp., I. baltica, and I. chelipes, did not differ among the nutrient enrichments (P > 0.1 for each taxon) but varied with the depth. Idotea spp. were more abundant in shallow water (2.7 [-2.1, +3.45] ind. 200/cm2) than in deep water (0.7 [-0.5, +1.5] ind. 200/cm2; χ21 = 10.45, P = 0.001). In Gammarus spp., depth and time interacted (χ21 = 14.43, P = 0.0001), indicating that the juveniles, emerging in early September (authors’ personal observations), were more abundant in shallow than in deep water (July: 5 [-1, +2] and 11 [-2, +3] and September 6 [-2, +2] and 0 ind. 200/cm2, in shallow and deep water, respectively).
TABLE A1. Invertebrate abundance (individuals/tile [200/cm2]; least square mean ± SE) in shallow and deep water, under different nutrient enrichment levels in late July and early September.