A moderate-energy sandy beach receiving a hlgh natural organic load of stranded wrack was studied over 1 yr. Nitrogen input to the beach was calculated from wrack nitrogen content, an average standing stock of 72 kg m-' wet wrack and a turnover time of 9 d, estimated from litter bag (2mm mesh) experiments. Nitrogen in the beach was measured as bacterial biomass (mean 105 g N m-'), meiofauna biomass (mean 0.1 g N m-') and interstitial water inorganic nitrogen (mean 83 g N m-'). Wrack nitrogen input was in excess of 1 4 kg N m-' yr-l. Despite receiving a high nitrogen load, this beach dld not exhibit any net accumulation of nitrogen over the year. Marine sandy beaches receive a variety of organic materials from the sea: macrophyte wrack, dead animals and dissolved and particulate organics flushed through the sand during the process of swash-water filtration (Riedl 1971, McLachlan 1983). Incorporated with this is nitrogen which is released when the material is consumed or mineralised on the beach. The nitrogen released can either be returned to the sea, trapped in the beach by incorporation into microbial or fauna1 biomass or in inorganic form in the groundwater (Eagle 1983), or lost by denitrification. Coastal areas are generally regarded as nitrogen sinks (Nixon 1981). Koop et al. (1982) and Koop & Lucas (1983), studylng kelp wrack breakdown on an artificial sandy beach over 8 d, found most of the nitrogen to be converted to bacterial biomass and only 1.5% to return to the sea. Hennig et al. (1983) examined nutrient cycling in small laboratory sand columns over 90 min periods and found the sand to accumulate nitrogen, concluding that beaches are nitrogen sinks. It seems intuitively obvious, however, that all nitrogen supplied to a beach in organic or inorganic form must ultimately return to the sea (Oliff et al. 1970, Pugh 1976, Eagle 1983) or the beach would, on geological time scales, show signs of such accumulations. The present study represents an attempt to address this apparent paradox. We selected a beach receiving a relatively high O Inter-Research/Printed in F. R. Germany natural organic input, measured this input in terms of nitrogen at the beginning and end of a 1 yr period and recorded possible signs of accumulation of nitrogen in the beach fauna or groundwater. The outcome is discussed in the framework of a variety of beach types. Methods. The monitoring program was conducted on a stable, moderate-energy shore near Port Elizabeth, which consists of a rocky platform at the mid-tide level with a sandy beach overlying it, giving way to dunes just above the drift line. This beach has an average slope of XI, a width of 30 m and a mean particle size of 300 pm. Mean monthly sea temperature range is 15 to 21 'C and maximum spring tidal range 2.1 m. The beach receives a regular input of algae torn loose from the adjacent rocky coast. Field work stretched from March 1984 to March 1985. For parameters other than wrack input, sampling was conducted at the beginning and end of the year period. Wrack was collected every 3 to 4 wk. Four equidistant sites up the beach were marked for sampling, from Site 1, the mid-water of spring tides (MWS), to Site 4, the high water of springs (HWS). For meiofauna, sampling extended vertically in 15 cm intervals into the sand as far as the low tide water table. Four replicate cores (15 cm long X 10 cm2 were taken of which 2 were pooled. The fauna were narcotised with 7 O/ O MgC12 and fixed in 10 % formalin. Duplicate sand cores of 10 cm3 for bacterial counts were taken at 10 to 20 cm vertical intervals at each site using a modified syringe and stored in 2 O/ O filtered formalin. Groundwater was collected at the water table. To col!ect wrack, a metal hoop covering 0.1 m2 was placed on the sand surface every 1 m along a transect covering the length of the beach. All algae within the hoop were weighed using a Pesola spring balance. Subsamples of 100 cm2 were removed from the meiofauna cores, extracted by 4 decants through a 45pm screen, stained in rose bengal and counted to major taxa. Counts were corrected for 90 O/O extraction 192 Mar. Ecol. Prog. Ser. efficiency (McLachlan 1978). For bacterial counts duplicate 0.5 g samples of wet sand were sonicated 3 times each for 5 min in 5 m1 filtered 2 % formalin at a sonication frequency of 25 kHz. Of the 15 m1 supernatant 0.5 m1 was filtered through a 0.2 pm Nuclepore polycarbonate filter and stained with DAPI (Coleman 1980). Bacteria were then counted using an epifluorescence microscope. Porosity of the sand was measured as the volume of water needed to saturate 100cm3 oven-dried sand. Groundwater samples were analysed for NH3-N, NO2-N, NO3-N and total nitrogen on a Technicon Autoanalyser 11. At the beginning and end of the program, wrack was separated into component species and the total wet mass of each obtained. Subsamples were then used to estimate percent moisture by drying at 70°C and nitrogen by the microkjeldahl method (Holme & McIntyre 1971). A litter bag experiment estimated the breakdown rates of algal wrack exposed to desiccating elements on the beach. One hundred 50 g samples of freshly collected algae from the nearby rocky shore were placed in nylon bags of 2 mm mesh and strewn over the beach. Bags were retneved at daily intervals and weighed to obtain wet mass. The experiment was carried out in early spring using the more abundant red alga, Hypnea rosea. Results. Wrack was distributed over the whole beach face but often concentrated towards the upper shore. The total volume of the triangular wedge of sand between the rock base and the drift line marker was 27.5 m3 per running metre. Based on a mean porosity of 45 O/O by volume this reduces to 12.4 m3 pore space, mostly filled with interstitial water, and 14.1 m3 sand grains. The volume of interstitial water held at low tide was estimated to be 9.4 m3, allowing for ca 3 m3 air space above the water table. The water table at the time of sampling was close to the surface and showed evidence of groundwater seepage as salinities from 0 to 34%0 were regularly recorded during wrack collections, particularly at upper tide levels. At initial and final sampling (March 1984, 1985) all salinities were above 16%. Reduced layers were noted at the 2 middle sampling sites, on both occasions within 10 cm of the sand surface. Bacterial densities were fairly uniform across the beach and there were no significant differences between sampling sites or depths on either occasion (2way ANOVA, p<0.05), although densities tended to be,higher at lower shore levels. Mean numbers per g wet ' sand were 2 orders of magnitide lower during March 1985 (3.62 X 10" than March 1984 (2.4 X 108) probably due to different wrack input immediately preceding sampling. There were large differences in groundwater nutrient levels between sites and dates. The residence time of water in the beach is 12 to 24 h (McLachlan 1979, 1982), sufficient for the oxidation of much ammonium to nitrate. The concentrations of all 3 oxidation states of nitrogen were significantly higher during 1985 than 1984 (2-way ANOVA, p<0.001). Due to analytical problems, total nitrogen analysis was only done at the end of the study period. At this time total nitrogen values were 2.5 to 4.4 times the total inorganic nitrogen values, suggesting that 59 to 77% (mean 67 %) of the nitrogen in the interstitial water was in organic form. Since this includes organisms such as bacteria, Protozoa and meiofauna, approximately 50 % of the non-living nitrogen in the interstitial water may have been in inorganic form during the time of final sampling. The standing stock of wrack recorded along the transect over the year is summarised in Table 1. The mean quantity of wet wrack on the beach was 72 kg Table 1. Standing stocks of algal wrack recorded over 12 mo Date Total wet mass Tide Transect length