When using the diffusive gradients in thin-films (DGT) technique in well-stirred solutions, the diffusive boundary layer has generally been ignored on the assumption that it is negligibly thin compared to the total thickness of delta g, i.e., the sum of the thickness of the prefilter and diffusive gel. Deployment of devices with different diffusive layer thicknesses showed that the thickness of the DBL was approximately 0.23 mm in moderate to well-stirred solutions, but substantially thicker in poorly or unstirred solutions. Measurement of the distribution of Cd in the DGT resin gel at high spatial resolution (100 microm) using laser ablation inductively coupled plasma mass spectrometry showed that the effective sampling window had a larger diameter (2.20 cm) than the geometric diameter of the exposure window (2.00 cm). Lateral diffusion in the gel, which had previously been neglected, therefore increased the effective surface area of the device by approximately 20%. The concentrations measured by DGT agreed well with the known concentrations in standard solutions for all diffusion layer thicknesses, when the effective area and the appropriate diffusive boundary layer (DBL) were used. The extent of the error associated with neglecting the DBL and using the geometric window area depends on the gel layer thickness and the true thickness of the DBL, as determined by the deployment geometry and flow regime. When DGT measurements were made in well-stirred solutions using a 0.80-mm diffusive gel, the effect of neglecting the DBL and using the inappropriate geometric area offset each other, with the error being <+/-10%. For precise measurements, and especially work involving speciation or kinetic measurements, where DGT devices with different diffusive gel layer thicknesses are deployed, it is necessary to use the effective area and the appropriate DBL thickness in the full DGT equation, which allows for the use of layer-specific diffusion coefficients.

Measurements of trace metal species in situ in a softwater river, a hardwater lake, and a hardwater stream were compared to the equilibrium distribution of species calculated using two models, WHAM 6, incorporating humic ion binding model VI and visual MINTEQ incorporating NICA−Donnan. Diffusive gradients in thin films (DGT) and voltammetry at a gel integrated microelectrode (GIME) were used to estimate dynamic species that are both labile and mobile. The Donnan membrane technique (DMT) and hollow fiber permeation liquid membrane (HFPLM) were used to measure free ion activities. Predictions of dominant metal species using the two models agreed reasonably well, even when colloidal oxide components were considered. Concentrations derived using GIME were generally lower than those from DGT, consistent with calculations of the lability criteria that take into account the smaller time window available for the flux to GIME. Model predictions of free ion activities generally did not agree with measurements, highlighting the need for further work and difficulties in obtaining appropriate input data.

Trace metals were measured in situ in a freshwater river draining a peat catchment (DOC = 15 mg L(-1)) using diffusive gradients in thin-films (DGT) devices with a range of gel layer thicknesses (0.16-2.0 mm). The reciprocal of the accumulated mass of each metal varied linearly with the thickness of the diffusive layer. These plots allowed calculation of the thickness of an apparent diffusive boundary layer (ADBL). A constant value was obtained from the plots of Cd, Pb, and Zn. The observed increase in the ADBL for the other metals (Mn

Several techniques for speciation analysis of Cu, Zn, Cd, Pb, and Ni are used in freshwater systems and compared with respect to their performance and to the metal species detected. The analytical techniques comprise the following:  (i) diffusion gradients in thin-film gels (DGT); (ii) gel integrated microelectrodes combined to voltammetric in situ profiling system (GIME−VIP); (iii) stripping chronopotentiometry (SCP); (iv) flow-through and hollow fiber permeation liquid membranes (FTPLM and HFPLM); (v) Donnan membrane technique (DMT); (vi) competitive ligand-exchange/stripping voltammetry (CLE−SV). All methods could be used both under hardwater and under softwater conditions, although in some cases problems with detection limits were encountered at the low total concentrations. The detected Cu, Cd, and Pb concentrations decreased in the order DGT ≥ GIME−VIP ≥ FTPLM ≥ HFPLM ≈ DMT (>CLE−SV for Cd), detected Zn decreased as DGT ≥ GIME−VIP and Ni as DGT > DMT, in agreement with the known dynamic features of these techniques. Techniques involving in situ measurements (GIME−VIP) or in situ exposure (DGT, DMT, and HFPLM) appear to be appropriate in avoiding artifacts which may occur during sampling and sample handling.

Cross-flow ultrafiltration techniques and a high-temperature combustion (HTC) method were used to investigate the distributions and fluxes of dissolved (DOC) and colloidal organic carbon (COC) in the Gulf of Mexico and in the Middle Atlantic Bight. Concentrations of DOC in both regions decreased from 580 PM in surface waters to -45 PM in deep waters and showed large vertical gradients in the subsurface layer. The vertical distributions of DOC were oceanographically consistent. A conservative mixing behavior of DOC was observed in slope waters, and water mixing processes were important factors in controlling the distribution of DOC in the ocean. Calculated downward fluxes of DOC are comparable to those measured for particulate organic C. Size fractionation results revealed that COC (1 kDa-0.2 pm) comprised -4O-50% of the total DOC in seawater off Cape Hatteras, while it comprised -30-40% in the Gulf of Mexico. Highmolecular-weight COC,, (10 kDa-0.2 pm) represented 4-l 0% of the total DOC in both study areas. Concentrations of COC, , COC3, and COC,, and their percentages in the total DOC decreased from nearshore to offshore and from surface to deep waters. The COC fractions seem to partition in a predictable way in seawater, with DOC concentration as a master variable. On average, 4-7% of the total DOC was in the COC,, fraction, 7-14% was in the 3-lo-kDa fraction, and -24% was in the I-3-kDa fraction, leaving 55-65% in the < I-kDa fraction.

Thiolic compounds are important metal-complexing ligands as well as important components of the global sulfur biogeochemical cycle. A lack of information on the concentration and distribution of thiols in natural waters, especially in the dissolved fraction, is still a major impediment to a complete understanding of the role of thiols in these biogeochemical processes. The concentrations of dissolved, colloidal, and particulate thiols were measured along a salinity gradient in estuarine waters off of Galveston Bay, Texas. The majority of thiols were present in the dissolved fraction, although more thiolic species were detected in the particulate phase. Dissolved glutathione was present at higher concentrations (0.23 to 6.23 nM) than was the particulate glutathione (0.094 to 0.72 nM). Most ␥-glutamylcysteine was present in the particulate phase, with concentrations as high as 2.24 nM in the middle of the estuary. Phytochelatin-2 was ubiquitous in surface waters, with chlorophyll a-normalized concentrations of up to 6.3 mol g Chl a Ϫ1 . A major thiol peak was present in Lower Galveston Bay and a minor peak in Upper Galveston Bay, and in both regions, 5-6 mol of ␥-glutamylcysteine were produced per mole of glutathione. This bimodal distribution indicates in situ production of thiols from two different phytoplankton communities in Galveston Bay during this period.Sulfur is an essential element and is required for protein synthesis by all organisms. The metabolic studies of sulfurcontaining species in natural waters have branched into two major areas, which are concerned with the following: (1) the global biogeochemical cycle of volatile sulfur species, such as dimethyl sulfide (DMS), carbonyl sulfide (OCS), and related compounds (Watts 2000) and (2) the detoxification properties of thiolic compounds, such as glutathione (GSH), cysteine (Cys), and sulfide, for ameliorating oxidative stress caused by trace metals, radicals, and other xenobiotic compounds (Grill et al. 1985;Giovanelli 1987). These two research areas are related at the molecular level in organisms during the synthesis and transformation of amino acids. Production of DMS is closely related to methylation reactions during methionine synthesis (Grone and Kirst 1992) and to the detoxification processes involved in GSH transformation (Meister and Anderson 1983). In water, organic matter-mediated photoreactions of dissolved organic sulfur compounds lead to the formation of OCS and related compounds (Zepp and Andreae 1995).GSH, as one of the most abundant low-molecular weight (LMW) thiols in animals, plants, and bacteria, has been shown to play an important role in protecting cells against oxidative stress, radiation damage, and elevated levels of heavy metals (Giovanelli 1987). The use of GSH (rather than Cys) as the major active reduced sulfur species is thought to have evolved when organisms adapted to the oxic atmo-1 Corresponding author (tangd@tamug.tamu.edu). AcknowledgmentsThis manuscript greatly benefited from discussions with J. Pinckney about...

In view of conflicting reports regarding the performance of DGT in low ionic strength solutions (I < 1 mM), further investigations have been carried out. Minimal washing of the diffusive gel and deployment in 1.0 and 10 mM NaNO3 solutions containing Cu and Cd gave the theoretical response of 1 for [C](DGT)/[C](SOLN), where [C](DGT) is the concentration of metal measured by DGT and [C](SOLN) is the concentration of metal measured directly in the solution by an appropriate analytical method. Erroneously high values for [C](DGT)/[C](SOLN) were obtained when these same gels were deployed at I = 0.1 mM, presumably due to a net negative charge on the gel, attributable to the presence of initiation products of polymerization. However, washing the diffusive gels completely, where the storage solution pH equaled that of deionized water, gave values of approximately 0.5 for [C](DGT)/[C](SOLN) from deployments at I = 0.1 mM, consistent with the lower measured value of the diffusion coefficients at this ionic strength. These results can be explained by the presence of a net positive charge on the gel when it is exhaustively washed, which reduces the effective diffusion coefficient of metal ions by changing their concentration at the gel-solution interface (Donnan partitioning). Diffusive gel equilibration experiments showed the presence of low capacity sites capable of binding metals irrespective of ionic strength. This binding within the diffusive gel does not affect most DGT measurements, as short (4 h) deployments at concentrations of 10 ppb gave theoretical results. Incomplete washing of the resin-gel caused a 5-15% measurement error and a decrease in precision, even at ionic strengths of 10 mM. A high level of accuracy and precision (typically <5%) was maintained during all aspects of this work, even at ionic strengths of 0.1 mM, in contrast to previous results. This is attributable to three factors: (1) exhaustive washing and conditioning protocols, (2) improvements to the DGT sampling device, and (3) low and reproducible blanks due to ultraclean handling procedures. Provided effective diffusion coefficients measured at the same ionic strength are used, the established DGT theory is obeyed irrespective of ionic strength.

A simple method for the analysis of polyacrylamide diffusive gradients in thin film (DGT) gels by laser ablation inductively coupled plasma mass spectrometry (LA-ICPMS), employing a novel use of (115)In internal standardization, has been developed. This method allows the determination of Co, Ni, Cu, Zn, Cd, and Pb concentrations (at the DGT filter face) or fluxes in sediments at a spatial resolution of 100 microm. Single-layered gels, using an optimized laser defocus of 4000 microm at 400 mJ power, showed high precision (generally approximately 10%) and a linear response during solution deployment. Of the elements Sc, In, Ba, La, Ce, and Tb, Ba most closely tracked variations in laser energy and showed the highest analytical precision but could not be used as an internal standard due to its elevated presence in natural sediments. Therefore, internal standardization, necessary to normalize data collected on different days, was carried out using (115)In contained within a second layer of backing gel and dried along with the analyte layer as a dual-gel disk. This multilayered gel standard required a laser defocus setting of 1000 microm and a laser power of approximately 800 mJ. Analytical precision for a 64-spot ablation grid at 100-microm spacing was approximately 10%. Verification of this method was carried out on DGT sediment probes deployed in Priest Pot (English Lake District). Results obtained by conventional slicing techniques and aqueous elution agreed with laser ablation results when the different sampling areas were considered. The elution results varied by a factor of <2, whereas the laser ablation technique showed a variability of approximately 4, indicating localized elevated concentrations of Co. This higher resolution LA-ICPMS method could ultimately lead to an improved understanding of the geochemical processes responsible for metal uptake and release in sediments.