We have carried out a series of experiments aimed in large part at testing the 
two hypotheses outlined above. We have proceeded by 1) determining the 
concentrations of elements forming relatively soluble sulfides (Mn and Fe) and 
elements forming relatively insoluble sulfides (Ni, Cu and Cd), along with 
nutrients, pH, total CO-, (TCO->), SO^ - , and S in anoxic Narragansett Bay 
sediments; 2) carrying out tracer experiments to determine the rate at which 
organisms enhance the benthic flux of chemicals by pumping water across the 
sediment-water interface, and using these results, along with diflusive flux 
estimates, to predict total fluxes across the sediment-water interface; 
3) measuring benthic fluxes directly via bell jar experiments, and using the 
results both to check the prediction of the pore water data and flux model, and 
to make direct determinations of metal benthic fluxes; and 4) working out the 
mass balance of heavy metals in Narragansett Bay at one point in time, using 
the Bay itself as a gigantic bell jar and thereby checking on the generality of 
our benthic flux measurements. 
In this paper we summarize the results of this work. 
PORE WATER WORK 
Samples were collected in plexiglas box cores or PVC pipe. One centimeter 
sediment slices were placed in polyethylene centrifuge bottles in a helium 
atmosphere and centrifuged at in situ temperatures to separate pore waters. 
Supernatant waters were filtered through acid-washed Nuclepore filters. The 
concentration of TCO-, was determined by gas chromatography. Sulfate was 
measured by BaSo^ coprecipitation using ^ ^Ba as a tracer. Sulfide and 
ammonia were determined colorimetrically by the methods of Cline (3) and 
Solorzano (10), respectively. Phosphate and silicate were determined by a 
modified autoanalyzer procedure (11, 12). Dissolved manganese and iron were 
determined by flameless atomic absorption spectrometry using either a 
Perkin-Elmer 503 or 360 AAS with model 2100 graphite furnace deuterium arc 
background corrector. Standards were prepared in Sargasso Sea water. Further 
details of sample collection and the above analytical methods are given by 
McCaffrey (8). 
Cd and Ni were determined by flameless atomic absorption after sample 
concentration using a modification of the Co-APDC coprecipitation technique 
of Boyle and Edmond (2). A 0.2 ml aliquot of a 4mM CoCl-> solution is added 
to the pore water sample at pH 2, followed by addition of 0.4 ml of a 2% W/V 
solution of APDC. The sample is shaken vigorously and allowed to sit for up to 
30 minutes. It is then filtered through a 2.5 cm diameter, 0.4 jum micron 
Nuclepore filter held by an acid-washed Millipore Filter apparatus. The filter is 
washed with several ml of deionized water, and placed in a polypropylene vial 
containing 1.0 ml of redistilled 3N HNO^. The precipitate is dissolved by 
10 
