SCOPE 48 - Sulphur Cycling on the Continents

7.1 INTRODUCTION

Over the past 15 years, research impelled by the lake acidification issue has led to an improved understanding of the biogeochemistry of S in lakes. This research has shown that

1. the S cycle in small lakes has been affected by increased atmospheric deposition (Rudd, Kelly and Furutani, 1986; Nriagu and Soon, 1985; Fry, 1986) with attendant effects on the acid-base chemistry of lakes (Landers et al., 1988; Brakke, Landers and Eilers, 1988);
2. under certain conditions biologically mediated SO₄^2- reactions, particularly bacterial SO₄^2- reduction, can significantly affect lakewater SO₄^2- chemistry (Kelly et al., 1987; Baker and Brezonik, 1988);
3. these reactions produce alkalinity and may counteract acidification (Cook et al., 1986);
4. historical changes in lakewater SO₄^2- concentrations may be recorded in sediments (Mitchell et al., 1988);
5. the diagenesis of S in freshwater sediments is more complex than previously thought (David and Mitchell, 1985; Rudd, Kelly and Furutani, 1986).

The purpose of this chapter is to examine the controls on biogeochemical reactions of S in the water column and sediments of temperate dimictic lakes. The processes causing fluxes of S between the water column and the sediments and the atmosphere also will be examined. Because the primary reactions that S undergoes are closely linked to organic matter, this chapter will also discuss the interactions of S with the elements predominant in organic matter (carbon, nitrogen, and phosphorus). We gratefully acknowledge Shirley Richards for allowing us to use her unpublished data. This is publication No.3569 of the Environmental Sciences Division, ORNL.

7.2 REACTIONS AND FORMS OF SULPHUR

A schematic diagram of the S cycle in small lakes is presented in Figure 7.1, and the microorganisms performing the conversions of S in aquatic systems are presented in Table 7.1. Sulphur is supplied to lakes from atmospheric deposition in the form of SO₄^2- and SO₂ and from terrestrial portions of watersheds, primarily in the form of SO₄^2-, but also as organic S (David and Mitchell, 1985; Mitchell, David and Harrison, this volume). In northern temperate regions, the export of S from terrestrial ecosystems is in balance with atmospheric inputs, so that the SO₄^2- concentration in lakes tends to be directly related to the rate of atmospheric deposition (Sullivan et al., 1988). In the water column, SO₄^2- is taken up by plankton during assimilatory SO₄^2- reduction and converted into carbon-bonded S compounds (e.g. the S- containing amino acids cysteine, methionine, and cystine). Algae may release dimethylsulphoniumpropionate (DMSP), which is broken down to dimethyl sulphide (DMS), but this mechanism for producing DMS is presently well established only for marine algae (Andreae and Barnard, 1984; Andreae and Jaeschke, this volume). Ester sulphate (C-OSO₃) compounds also may be produced by plankton (Schiff and Hodson, 1973; David and Mitchell, 1985). The various organic-S compounds may be recycled in the water column or may settle out of the water column where they either accumulate or are recycled.

                      
 Figure 7.1. Schematic diagram of the sulphur cycle in small lakes. Polysulphides are indicated by S_x^2-

Table 7.1. Sulphurusing bacteria in aquatic systems

Lakewater SO₄^2- is conveyed by diffusion and bioturbation to the anoxic portions of the lake where it undergoes reactions, the most important of which is bacterial SO₄^2- reduction forming H₂S (Table 7.1) .Some SO₄^2- is assimilated by known biosynthetic pathways into bacterial biomass. H₂S may either react to form metal sulphides, or diffuse to oxic zones where it is oxidized chemically or by chemoautotrophic or photo autotrophic S-oxidizing bacteria to S⁰, S_x^2- (polysulphides) or SO₄^2- (Table 7.1; Luther and Church, this volume). Elemental sulphur may also serve as an electron acceptor in the anaerobic oxidation of organic matter (Table 7.1) .

Reduced sulphur (H₂S, S⁰, S_x^2-) may react with humic substances (Casagrande and Ng, 1979; Casagrande et al., 1979; Vairavamurthy and Mopper, 1987, 1989; Francois, 1987; Luther and Church, this volume). The organic-S compounds formed are of unknown chemical make-up but may include thiols, mercaptans, and thioethers. Breakdown of organic-S compounds by heterotrophic bacterial metabolism may supply H₂S to the anoxic zones (Dunnette, 1973, 1989; Zinder and Brock, 1978; Jones, Simon and Roscoe, 1982).

Reduced S species may undergo chemical or biological oxidation (Wetzel, 1975; Luther and Church, this volume). Chemoautotrophic S-oxidizing bacteria are primarily aerobic taxa that can oxidize reduced S inside or outside of the cell (Table 7.1 ; Wetzel, 1975). The anaerobic phototrophic S-oxidizing bacteria are located on the sediments or in a narrow layer in the water column, where light, redox, and pH conditions are optimal (Table 7.1 ; Parkin and Brock, 1981).

Sulphur gases, such as dimethyl sulphide, carbonyl sulphide, methane thiol, and dimethyl disulphide, may be formed in the water column and sediments. The processes by which these gases are formed have not been adequately studied in freshwater systems, but they appear to be linked both to the decomposition of organic-S compounds in the sediments and to either decomposition or algal release in the water column (Richards, Kelly and Rudd, 1991).

 The total amount of S in a particular sediment horizon is the net accumulation of particulate forms deposited at the time the sediment was laid down and S incorporated in the sediments following diffusion and reaction in the horizon. The palaeolimnological interpretation of sediment S profiles is confounded by the many mobile forms of S (SO₄^2-, H₂S, and soluble organic- S compounds) (Holdren et al., 1984; Carignan and Tessier, 1988).

7.2.1 INCORPORATION OF SULPHUR BY BIOTA

All living organisms require S as a minor nutrient, in roughly the same atom proportion as phosphorus (Howarth and Stewart, this volume). Sulphur is present in freshwater algae at a ratio of about 1 S atom to 100 C atoms (0.15 to 1.96% by dry weight), and the S content varies with species, environmental conditions, and season (Baker et al., 1989; King and Klug, 1982). Vascular plants, algae, and bacteria (except some anaerobes that require S^2-) have the ability to take up, reduce, and assimilate SO₄^2- into amino acids and convert SO₄^2- into ester sulphate compounds.

For most lakes, SO₄^2- concentrations in aerobic habitats are high enough to supply all of the S required by planktonic algae and bacteria. In anaerobic habitats, SO₄^2- is depleted but H₂S is usually present and can be used in biosynthesis. Because of this abundant supply, aquatic biota apparently do not take up S in excess of that required for growth, thereby maintaining relatively constant concentrations of S in organic matter. An exception to this generalization about storage is the phototrophic sulphur bacteria that store elemental sulphur (Table 7.1) .This stored S⁰ is for later use as an electron donor when other donors become unavailable, rather than for biosynthesis.

In the absence of changes in the major nutrients (P, N), changes in lakewater SO₄^2- concentrations will not likely lead to changes in algal S uptake and sedimentation, because the nutritional requirement of aquatic biota for S has not been affected. For example, experimental additions of SO₄^2- to lakes did not stimulate the algal uptake of S, as indicated by the similarity between C : S ratios in seston of acidified and control lakes at the Experimental Lakes Area (ELA), northwestern Ontario (J. Crusius, Columbia University, unpub. data) and by C: S ratios in seston of the control and treatment basins in Little Rock Lake, Wisconsin (molar C : S ranged from 99 to 144; Baker et al., 1989). Knauss and Porter (1954) observed less than a doubling of SO₄^2- content in Chlorella when SO₄^2- concentrations were increased from 40 to 1400 mg (S) l^-1 (1.3 to 44 mM). Also, examination of SO₄^2- uptake by a floating vascular plant over a 3000-fold variation in medium SO₄^2- concentration showed that total SO₄^2- uptake varied by only three-fold to four-fold, with the variation reflected in a change in inorganic SO₄^2- in the tissue (Datko et al., 1978).

The lowest SO₄^2- concentration in the studies cited above is about 1 mg (S) l^-1 (30 µM) (Crusius, unpub. data; Baker et al., 1989). The algal uptake of SO₄^2- has not been studied in lakes having concentrations below this value. SO₄^2- concentrations of 0.1mg (S) l^-1 (30 µM) were reported for Lake Victoria (R. Hecky, Freshwater Institute, Winnipeg, unpub. data), and concentrations below 0.024 mg (S) l^-1 (0.75 µM) have been observed in some lakes remote from human influence [e.g. in the mountainous western United States (Eilers et al., 1989) and on the Kenai Peninsula, Alaska (Landers et al., 1989)]. For lakes with low concentrations of SO₄^2-, the supply of SO₄^2- in the water column may limit the growth of aquatic biota, but this has not been demonstrated.

From the evidence on S uptake summarized above, the total assimilatory uptake of S by algae and bacteria and the accumulation of seston S in sediments can be expected to be related more to biomass or the trophic status of the lake than to the concentration of lakewater SO₄^2- .Increases in major nutrient (e.g. P) inputs to a lake may lead to an increased sedimentation of both organic C and organic S, without changing the C: S ratio of the sedimenting material.

Organic-S compounds may be recycled within the water column and the sediments. Ester sulphate compounds (C-OSO₃) may be hydrolysed to SO₄^2- (King and Klug, 1980; Luther and Church, this volume); S in amino acids and proteins may be broken down to form H₂S and other gaseous forms of S (Dunnette, 1973, 1989; Zinder and Brock, 1978; Jones, Simon and Roscoe, 1982; Andreae and Jaeschke, this volume), some of which may then undergo oxidation. The few studies of organic-S recycling have shown that between 25 and 80% of the seston S falling on the sediment-water interface is decomposed and re-enters the water column (Smith and Klug, 1981; King and Klug, 1982; David and Mitchell, 1985; Baker et al., 1989). King and Klug (1982) estimated that less than 50% of the total S and 75% of the protein S was mineralized. David and Mitchell (1985) observed that half of the carbon- bonded S was recycled at the sediment-water interface (upper 1 cm). These studies assumed that the recycling could be determined from the comparison of rates of seston S inputs and total accumulation of organic S in the sediments. SO₄^2- reduction was assumed to be insignificant and, therefore, not a source of organic S formation (David and Mitchell, 1985; Baker et al., 1989), or the possibility of this source was not considered (Smith and Klug, 1981; King and Klug, 1982). Because some of the organic S may be derived, not from seston, but from sulphides produced from bacterial SO₄^2- reduction (Rudd, Kelly and Furutani, 1986; Luther and Church, this volume), these recycling values may be underestimates. Measurement of seston S input and sediment accumulation of organic S in lakes with extremely low SO₄^2- concentrations is required to avoid this additional source of organic S.

7.2.1.1 Ester Sulphates

SO₄^2- may also be taken up by aquatic biota and converted into ester sulphate compounds (C-OSO₃) (Schiff and Hodson, 1973) in the water column (David and Mitchell, 1985). Ester sulphate compounds may also be formed within the sediment from SO₄^2- (Landers and Mitchell, 1988), although no evidence exists on whether the ester sulphate compounds are extracellular or intracellular. Measurements of the ester sulphate content of seston range from 32% (King and Klug, 1982) to 44-59% (David and Mitchell, 1985). The percentage of S in ester form in surficial sediments is similar or smaller than seston (35% , King and Klug, 1982; 22% , David and Mitchell, 1985). Ester hydrolase activity is present in sediments, showing that esters can be broken down biologically and can provide a source of SO₄^2- for SO₄^2- reduction (King and Klug, 1980). Mineralization of ester sulphate has been estimated by comparing seston input to sediment accumulation, yielding 43% (David and Mitchell, 1985) and < 50% (King and Klug, 1982) of the input ester sulphate being decomposed.

7.2.2 BACTERIAL SULPHATE REDUCTION

Dissimilatory SO₄^2- and S⁰ reduction is the process by which bacteria use SO₄^2- and other oxidized forms of S (oxidation state of O or greater; e.g. S₂O₃^2- and S⁰) as terminal electron acceptors in the catabolism of organic matter (Table 7.1). The SO₄^2-- and So-reducing taxa are strictly anaerobes and require organic matter or hydrogen as the electron donors (Pfennig, Widdel and Truper, 1981).

Assimilatory SO₄^2- reduction is also performed by microorganisms in both the water column and the sediments for synthesis of S-containing compounds, primarily amino acids. In aerobic portions of the water column, assimilatory reduction will predominate over dissimilatory reduction, because the latter process is carried out only by obligate anaerobes. No measurements have been made of assimilatory reduction in sediments, but assimilatory reduction is probably minor compared to dissimilatory reduction. For example, in cultures of Desulphovibrio, only 1 mol of S was assimilated for every 400 mol of SO₄^2- reduced to H₂S for the purpose of meeting energy requirements (R. Cuhel and C. Taylor, pers. comm.). Of course, in sediments, non-SO₄^2-- reducing microorganisms are also assimilating S for cell synthesis. The energetic metabolism of these other bacteria is about four times that of the SO₄^2- reducers (Kelly, Rudd and Schindler, 1988), but their S assimilation would still be small compared to SO₄^2- reduction. Assimilatory reduction is linked to population size and cell growth. Thus, it should not respond to  increases in SO₄^2- concentrations as does dissimilatory SO₄^2- reduction (as explained below).

The rate of bacterial SO₄^2- reduction is one of the better known processes in the S cycle in lakes. This rate has been measured directly by ³⁵S tracer (release of produced [³⁵S] H₂S or uptake of ³⁵SO₄^2-) (Stuiver, 1967; Ivanov, 1968; Sorokin, 1975; Jones, Simon and Roscoe, 1982), depletion of SO₄^2- over time in the water column (Stuiver, 1967; Cook and Schindler, 1983; Kelly, Rudd and Schindler, 1988), and calculations of SO₄^2- fluxes from porewater gradients (Holdren et al., 1984; Carignan, 1985; Rudd et al., 1986) . It is important to note that ³⁵SO₄^2- reduction rates represent a lower limit if the rates are measured without recovering all of the possible end-products (e.g. release of only [³⁵S] H₂S; see also Howarth and Teal, 1979; Giblin and Wieder, this volume). For the majority of lakes studied, the rates of bacterial SO₄^2- reduction range from 0.02 to 0.2 mmol m^-2 day^-1 (0.6 to 6 mg (S) m^-2 day^-1) (Table 7.2). The lowest rates reported, 0.02 mmol m^-2 day^-1 (0.06 mg (S) m^-2 day^-1), were measured in Little Rock Lake (Cook et al., 1987; Baker et al., 1989) and Sagamore Lake, and the highest (450 mmol m^-2 day^-1; 14 g (S) m^-2 day^-1) were measured in an impoundment with high concentrations of SO₄^2- (3 to 20 mM; 96 to 640 mg (S) l^-1) derived from acid mine drainage (Herlihy et al., 1987).

Table 7.2. Rates of bacterial sulphate reduction in lake sediments Reduction

Rates of bacterial SO₄^2- reduction are affected by organic carbon quantity and quality (Rudd, Kelly and Furutani, 1986; Sweerts, Rudd and Kelly, 1986; Giblin, Peterson and Fry, 1989) and also by the concentration of SO₄^2-(Cook and Schindler, 1983; Kelly and Rudd, 1984; Lovley and Klug, 1986). In the following two sections (7.2.2.1 and 7.2.2.2.), the influence of organic carbon supply and SO₄^2- concentrations on bacterial SO₄^2- reduction are described. Section 7.2.2.3 discusses the concept of the 'limiting' factor .

7.2.2.1 SO₄^2- Concentrations and Bacterial SO₄^2-  Reduction

In marine systems, which have SO₄^2- levels as high as 28 mM (900 mg (S) l^-1), SO₄^2- concentrations are high enough that bacterial SO₄^2- reduction rates are independent of SO₄^2- (Boudreau and Westrich, 1984; Capone and Kiene, 1988; Giblin and Wieder, this volume). In contrast, in lakes with SO₄^2-concentrations below 1 to 2 mM (32 to 64 mg (S) l^-1), bacterial SO₄^2- reduction is typically first order with respect to SO₄^2- concentrations (Ramm and Bella, 1974; Kelly and Rudd, 1984; Lovley and Klug, 1986). For these systems, increases in SO₄^2- concentrations (caused, for example, by increases in atmospheric SO₄^2- deposition) will cause increases in the rates of SO₄^2- reduction. For example, in Lake 223, ELA, experimental additions of sulphuric acid caused lakewater SO₄^2- concentrations to increase three-fold, causing bacterial SO₄^2- reduction rates to increase both in the anoxic hypolimnion and in the sediments throughout the lake (Figure 7.2; Cook and Schindler, 1983; Cook et al., 1986). In Lake 223, the accumulation of total S in the sediments increased primarily as a consequence of increased SO₄^2- reduction (Figure 7.3a). Similar findings on the relationship between SO₄^2- concentrations and SO₄^2- reduction rates have been reported in limnocorals (Perry, Baker and Brezonik, 1985; Schiff and Anderson, 1987) and in another whole-lake experiment (Rudd et al., 1990).

Rates of bacterial SO₄^2- reduction for these systems can increase with increases in SO₄^2- concentrations, because the supply of organic carbon is  large enough that only SO₄^2- is limiting the rate of SO₄^2- reduction in a particular lake. Prior to the experimental acid additions in Lake 223, organic matter decomposition in littoral sediments and in the hypolimnion occurred primarily by three processes: oxic decomposition, bacterial SO₄^2- reduction, and methanogenesis (Figure 7.4). In anoxic zones, decomposition was carried out primarily by methanogens (Kelly et al., 1982; Kelly, Rudd and Schindler , 1988). With the SO₄^2- additions to Lake 223, bacterial SO₄^2- reduction was stimulated at the expense of methanogenesis (Cook and Schindler, 1983). This stimulation is well documented in anaerobic mixed-population systems and is related to the energy yield for bacterial SO₄^2- reduction being greater than that for methanogenesis (Lovley and Klug, 1986; Kuivila et al., 1989;

                    
  
 Figure 7.2. Lakewater SO₄^2- concentration and a real rates of bacterial sulphate reduction in the anoxic hypolimnion of Lake 223 in the Experimental Lakes Area of northwestern Ontario. Data from Cook and schindler (1983) and Cook et al., (1986)

Figure 7.3 Acid volatile and total sulphur profiles in the sediments of (a) Lake 223 and (b) Lake 227. From Cook,1981

Howarth and Stewart, this volume). As Lake 223 was acidified, a change occurred, not in the total amount of organic matter that decomposed, but in the fraction of decomposition performed by bacterialSO₄^2- reduction (Cook, 1981; Kelly et al., 1982; Kelly, Furutani and Schindler, 1984). In the hypolimnion of Lake 223 during manipulation, the molar ratio of methanogenesis to SO₄^2- reduction was 3.2 : 1.0 (Kelly et al., 1982; Kelly, Rudd and Schindler, 1988), suggesting thatSO₄^2- reduction rates could increase further if moreSO₄^2- were added. In many low-SO₄^2- freshwater lakes, most organic matter is decomposed in anoxic zones by methanogenic bacteria (Ingvorsen and Brock, 1982; Kelly et al., 1982; Kelly, Rudd and Schindler, 1988; Kuivila et al., 1989), so that organic matter will not generally limit bacterial SO₄^2- reduction in these systems. However, this does not mean that organic substrate availability plays no role in determining the rate of SO₄^2- reduction (see below).

The rate of supply of SO₄^2- to the zone of bacterial SO₄^2- reduction- typically the sediments-may limit the rate of bacterial SO₄^2- reduction. For most systems, SO₄^2- is supplied by molecular diffusion or bioturbation (see below). However, for Lake Anna, Virginia, which is affected by acid mine drainage, SO₄^2- is scavenged from the water column by iron oxyhydroxide particles, enhancing the rate of SO₄^2- supply to the reducing zone and apparently increasing the rate of bacterial SO₄^2- reduction as measured by ³⁵SO₄^2- (Herlihy et al., 1987; Mills, Bell and Herlihy, 1989). This system has the highest reported SO₄^2- reduction rate for any lake (Table 7.2).

 Figure 7.4. Schematic diagram of the primary reactions that decompose organic matter in lakes. Based on a figure from Martens and Jannasch (1983)

The lowest concentration at which SO₄^2- reduction can occur is uncertain. At low SO₄^2- concentrations (< 20 to 40 µM; < 0.64 to 1.3mg (S) l^-1), thermodynamic calculations suggest that methanogenic bacteria are energetically favoured over bacteria that reduce SO₄^2- (Cook, 1981). Rates of bacterial SO₄^2- reduction have been observed to be low at concentrations below 40 µM (1.3 mg (S) l^-1; Lovley and Klug, 1986), and the half- saturation constant for bacterial SO₄^2- reduction in cultures has been reported as ranging from 5 to 70 µM (0.16 to 2.2mg (S) l^-1; Ingvorsen and Jųrgensen, 1984). However, workers have reported SO₄^2- concentrations as low as 0.2 to 5 µM (0.006 to 0.16mg (S) l^-1) in anoxic porewaters (Cook, 1981; Holdren et al., 1984; Kelly and Rudd, 1984; Cook et al., 1986,1987; Rudd, Kelly and Furutani, 1986).

7.2.2.2 The Influence of Organic Carbon Supply on Bacterial SO₄^2- Reduction

The supply of organic matter to lake sediments influences the S reactions and retention in sediments. Gorham et al. (1974) found that the accumulation of S in sediments of lakes in the English Lake District was highly correlated to algal standing crop and the organic S content of the sediments. Although the forms of S accumulating and the mechanisms for their formation were not determined, Gorham et al. attributed this correlation to the occurrence of extensive anoxic zones that favour both preservation of seston S compounds and the reduction of SO₄^2- and subsequent retention of reduced S.

Sweerts, Rudd and Kelly (1986) observed that additions of flocculent organic-rich particles to surface sediments in situ caused an increase in the rates of oxygen consumption and bacterial SO₄^2- reduction. Because of the uptake of oxygen in the organic floc, the oxic-anoxic boundary moved closer to the surface of the consolidated sediment. The zone of SO₄^2- reduction also moved closer to the sediment surface steepening the vertical SO₄^2- concentration gradient, thereby causing an increase in the diffusive flux and the rate of SO₄^2- reduction.

At Lake 227, ELA, the rate of bacterial SO₄^2- reduction was stimulated by experimental eutrophication (Cook, 1981, 1984). Nitrogen and phosphorus additions, which were made to simulate cultural eutrophication, caused primary production rates and concentrations of chlorophyll a to increase (Fee, 1979; Schindler, 1985). The total sedimentation rate increased from 45 g (dry wt) m^-2 a^-1 (Kipphut, 1978) to 210 g (dry wt) m^-2 a^-1 (Anderson, Schiff and Hesslein, 1987) as a result of the fertilization. Anoxic conditions in the water column increased and extended throughout the hypolimnion (Cook, 1984). These changes led to an increase in the habitat for bacterial reduction and an increase in rates of bacterial SO₄^2- reduction. The sediment accumulation of total S at the deepest point in Lake 227 increased from 9 to 20 mmol m^-2 a^-1 (0.3 to 0.64 9 (S) m^-2 a^-1; Figure 7.3b; Cook, 1984) as a consequence of both increased SO₄^2- reduction and sulphide retention and an increased sedimentation of organic S derived from algae. Although complete analyses for S species were not made, acid-volatile S and authigenic pyrite were among the forms accumulating (Cook, 1984).

Data from both eutrophic and oligotrophic lakes also illustrate the effect of organic carbon supply on bacterial SO₄^2- reduction. Eutrophic lakes (Men-dota and Wintergreen) have much higher SO₄^2- reduction rates than oligotrophic lakes (Clearwater and Lake 223) even though the SO₄^2- concentrations in these four lakes are similar (Table 7.2).

A low rate of organic carbon supply may also limit the extent of bacterial SO₄^2- reduction. At McNearney Lake (northern Michigan, USA), Cook et al. (1990) hypothesize that the low supply of organic matter to the sediments causes the rate of bacterial SO₄^2- reduction and S accumulation to be low. The lake has been naturally acidic for the past several thousand years and is ultraoligotrophic, having very low levels of chlorophyll a (0.2 µg l^-1) and extremely low rates of bulk sedimentation (0.1 to 0.6 mm a^-1; Cook et al., 1990). The SO₄^2- concentration in McNearney Lake is 75 µM (2.4 mg (S) l^-1); porewater concentrations in this lake have not been measured. The small concentration of S in the sediments has remained relatively constant over the past few hundred years (Figure 7.5) and exhibits no changes at the time of increases in the anthropogenic polycyclic aromatic hydrocarbons or lead (Cook et al., 1990).

The stable S isotope content of sediments in McNearney Lake shows only a slight depletion of d³⁴S (< 4%o), indicating a small amount of bacterial SO₄^2- reduction and of sulphide retention (Figure 7.5 ; Fry, 1990) .Stable isotope data from other small oligotrophic lakes in which both high rates of bacterial SO₄^2- reduction and efficient retention of sulphides occur, indicate surface sediment d³⁴S depletions of 10 to 15%o compared with downcore d³⁴S values (Nriagu and Soon, 1985; Fry, 1986; Fry, 1990). Additional evidence for low rates of bacterial SO₄^2- reduction in McNearney sediments is the absence of a recent increase in S concentrations and the relatively high molar C : S ratios (range: 111 to 162; Cook et al., 1990). Retention of S by bacterial SO₄^2- reduction coupled with metal or organic S formation causes lower molar C : S ratios (< 75; Gorham et al., 1974; Nriagu and Soon, 1985; Mitchell et al., 1990). The sedimentary S accumulating in McNearney Lake is apparently derived from organic S formed in the water column, along with some limited amounts of S derived from bacterial SO₄^2- reduction. In McNearney Lake, Cook et al., (1990) hypothesized that the low rates at which organic matter is supplied to the sediments may cause most of the organic matter to be oxidized aerobically at the sediment-water interface. Apparently, little of the organic matter in the anoxic portion of the sediments is suitable for bacterial SO₄^2- reduction, despite high lakewater SO₄^2- concentrations. Other workers have observed that bacterial SO₄^2- reduction rates in certain lakes were low and attributed these rates to low supplies of organic carbon (David and Mitchell, 1985; Rudd et al., 1986).

Figure 7.5. Total sulphur profile and stable sulphur isotope content (d³⁴S) in McNearney Lake, Michigan. (Fry, 1990). The anthropogenic Pb and polycyclic aromatic hydrocarbon rises occurred at 3 to 4 cm in this core (Cook et al., 1990).

7.2.2.3 Is Sulphate Concentration or Organic Carbon Supply Limiting Sulphate Reduction? 

The limiting factor-the environmental parameter that is in least supply compared to the requirements of the organism-will determine the activity and growth rate of an organism. The previous two sections discuss evidence for both SO₄^2- and organic carbon being the factor limiting the activity of SO₄^2- reducing bacteria. Recent models of SO₄^2- removal in acid-sensitive lakes assume that bacterial SO₄^2- reduction is first order with respect to SO₄^2- concentration (Kelly et al., 1987; Baker and Brezonik, 1988) and thus consider SO₄^2- to be the limiting factor. Rate constants (which relate rates and concentrations) for the suite of lakes studied by Kelly et al. and Baker and Brezonik had a remarkably small range. However, as the model developers pointed out, this small range in rate constants occurs because the lakes to which the model was applied are oligotrophic and have similar rates of carbon supply to the sediments, as well as only a small component of SO₄^2- removal by algal uptake.

The Lake 227 experiment shows that when carbon supply changes, S retention in lake sediments also changes (Figure 7.3(b); Cook, 1981). The experiment by Sweerts, Rudd and Kelly (1986), in which organic-rich floc was added to the sediment surface of cores, partly explains this by showing that the increased organic carbon causes the oxic-anoxic interface within the sediment to move closer to the sediment surface. The location of this interface in the sediment, together with the SO₄^2- concentration, determines the concentration gradient and thus the flux of sulphate into the sediments. In a given lake, with a particular level of productivity, the average position of the oxic-anoxic boundary can be considered fixed at a certain depth by the supply of organic carbon. With increases in SO₄^2- concentration, both the concentration gradient and the rate of SO₄^2- reduction will increase because there is usually enough organic substrate to support increased SO₄^2- reduction. Therefore, SO₄^2- can only be considered limiting when the organic carbon supply is constant and at a high enough level.

The sediment depth of the oxic-anoxic boundary for ELA lakes is deeper in the winter than in the summer (C. A. Kelly, unpub. data), probably due to a combination of lower temperature and lower organic carbon supply. SO₄^2- reduction rates in summer are higher than average net annual rates, because of lower winter rates and also because some of the SO₄^2- reduced during the summer is apparently re-oxidized during the winter (Rudd, Kelly and Furutani, 1986). However, in the study cited, no sulphur was lost below the depth of 3 cm, which is probably the sediment depth below which oxygen does not penetrate in the winter. Even though the daily rate of SO₄^2- reduction varies throughout the year, and some reduced sulphur is re-oxidized, both of these phenomena are likely related to organic input to the sediments. For the suite of lakes examined by Kelly et al., (1987) and Baker and Brezonik (1988), the similar rate constants for in-lake SO₄^2- retention likely are a result of similar trophic status and similar organic matter sedimentation. These findings mean that neither the SO₄^2- concentration nor the organic carbon supply rate by itself is an absolute predictor of SO₄^2- reduction rates. Therefore, neither parameter strictly fits the definition of a limiting factor .

7.2.3 FATE OF SULPHIDE

After H₂S is produced in sediments, whether by SO₄^2- reduction or decomposition of organic sulphur, it either diffuses into the water column or remains in the sediments. H₂S accumulates in anoxic hypolimnetic water, but it is oxidized in aerobic waters and is present only in small concentrations in surface waters (Luther and Church, this volume). H₂S may escape to the atmosphere when the rate of gas exchange is high (e.g. fall overturn). Low pH conditions may also promote high rates of gas exchange, because at pH < 6 the solubility of H₂S(g) is low (Cook, 1981) and the rates of chemical oxidation are slow (Chen and Morris, 1972; Millero et al., 1987; Millero and Hershey, 1989). Sulphide may also undergo a number of reactions within the sediments, such as oxidation and metal sulphide and organic S formation (Luther and Church, this volume).

Early work on SO₄^2- reduction in lake sediments assumed that the products were H₂S and iron monosulphide (acid-volatile sulphides: Sorokin, 1975; Berner, 1984). Howarth and colleagues (Howarth, 1979; Howarth and Teal, 1979; Howarth and Merkel, 1984) showed that pyrite (FeS2) could also be formed rapidly as a result of SO₄^2- reduction in salt marsh sediments, and the same process was later shown to occur in freshwater sediments (Rudd, Kelly and Furutani, 1986). In addition, ³⁵S from ³⁵SO₄^2- was also shown to be incorporated into organic matter in lake sediments (Rudd, Kelly and Furutani, 1986), although this apparently does not occur to any significant extent in salt-marsh sediments (Howarth and Merkel, 1984). Because other phases besides acid-volatile sulphide may be formed, use of the production of only the acid-volatile sulphide to estimate bacterial SO₄^2- reduction rates will yield underestimates (Howarth, 1979; Giblin and Wieder, this volume). The conditions determining which reaction is followed (oxidation vs metal S formation vs organic S formation) are poorly understood, but their occurrence in sediments has been studied by several groups, as discussed below.

7.2.3.1 Formation of Metal Sulphides

In addition to H₂S, reduced forms of metals are produced in anoxic zones of sediments. Use of oxidized forms of metals as electron acceptors by bacteria releases sufficient Fe²⁺ and other metals (e.g. Cd²⁺ , Pb²⁺) to cause these environments to become supersaturated with respect to the sulphides of iron and other reduced metals (Davison, 1980; Cook, 1984; Lovley, 1987). For some lakes, precipitation of iron sulphides (amorphous iron monosulphide, hydrotroilite, mackinawite, pyrrhotite, greigite, and pyrite; Berner, 1984; Davison, 1980) may be an important mechanism for incorporating S into lacustrine sediments and removing S from the water column (Figure 7.3; Cook and Schindler, 1983; Cook, 1984; White et al.,1989; Giblin et al.,1990).

The factors controlling the formation of the various reduced S minerals are poorly known. Pyrite and acid-volatile metal sulphides (hydrotroilite, mack- inawite, and greigite) have been identified in lake sediments (Cook, 1981; Kelly and Rudd, 1984; Nriagu and Soon, 1985; Giblin et al., 1990), but determining the exact chemical form of reduced S present and conditions that lead to the formation of these species requires further research.

The flux of iron to the zone of highest sulphide concentrations in the sediments may limit the ability of sediments to sequester S (Schindler, 1985; Carignan and Tessier, 1988; Rudd, Kelly and Furutani, 1986; Herlihy et al., 1987; Giblin et al., 1990). Where measured, rates of iron accumulation in lakes are either comparable in magnitude to S accumulation or greater by five-fold to ten-fold (Table 7.3). Rudd, Kelly and Furutani (1986) found that the proportion of extractable Fe associated with S in several acidified Adirondack lakes was a small fraction of the total Fe. Much higher proportions of S- bound Fe were measured in the sediments of acidic lakes in Quebec (Carignan and Tessier, 1988), indicating that Fe supply could become limiting to further formation of iron sulphide compounds in those lakes. Giblin et al., (1990) found similar results for several New England lakes with organic-rich sediments and low concentrations of non-S-bound Fe. It should be noted that some of the total Fe measured may be unavailable for microbial reduction or may be bound within mineral lattices and released slowly (Lovley and Phillips, 1986). Fe²⁺ available for reaction with sulphide may also be supplied by diffusion from sediment horizons beneath the zone of SO₄^2- reduction. Concentration gradients of Fe²⁺ in the porewater are typically high and may, over long periods of time, cause the transport of Fe from the entire sediment column to the zone where sulphide is produced and iron sulphide is formed (Cook, 1984; Cook et al., 1987). Once the inventory of reactive iron is exhausted (Anderson and Schiff, 1987), the capacity for S retention by reaction with iron is limited by the rate of supply of Fe. If precipitation of metal sulphides or incorporation into organic forms does not occur, H₂S may build up to high concentrations or diffuse to oxic zones where it will eventually be oxidized back to SO₄^2-. A decrease in the capacity of the sediments to retain S would lead to greater cycling of S from the sediments and higher SO₄^2- concentrations in the water (Giblin et al., 1990).

Table 7.3. Rates of sulphur and iron accumulation in lake sediments

7.2.3.2 Reaction with Organic Matter

In addition to reacting with metals to form sulphide compounds, H₂S and S⁰  may react with organic matter to form organic-S compounds (e.g. thiols, mercaptans, and thioethers), as has been shown in both marine (Casagrande et al., 1979; Casagrande and Ng, 1979; Francois, 1987; Vairavamurthy and Mopper, 1987, 1989; Luther and Church, this volume) and freshwater sediments (Nriagu and Soon, 1985; Rudd, Kelly and Furutani, 1986). Organic-S compounds are produced either by the replacement of a carbon-bonded oxygen atom with an S(II) atom or by addition of H₂S to a carbon-carbon double bond (Anderson and Schiff, 1987; Luther and Church, this volume). The exact conditions under which organic S is formed from H₂S and the factors controlling the relative amounts of metal sulphide and organic S formed from H₂S are not known (Nriagu and Soon, 1985; Rudd, Kelly and Furutani, 1986; Carignan and Tessier, 1988). The quality and quantity of organic carbon in lake sediments may possibly limit the synthesis of organic-S compounds and may be responsible for the unexplained variability in organic S content in lake sediments.

Other factors affecting the relative reactivity of H₂S for iron or organic C may be the concentrations of Fe²⁺ and total sulphide in the porewater, and pH, which affects the portion of total sulphide that is S₂^- .Baker et al. (1989) showed that adding Fe²⁺ or lowering the pH decreased the proportion of ³⁵SO₄^2- reacting to form organic ³⁵S in sediment incubations. However, examination of ³⁵SO₄^2- incubation data from a variety of lakes studied by Rudd, Kelly and Furutani (1986) reveals no clear pattern. Determination of the factors controlling organic S formation from sulphide requires additional research.

Rudd, Kelly and Furutani (1986) added ³⁵SO₄^2- to epilimnetic sediments in situ and observed that the tracer initially was reduced and incorporated into both inorganic and organic S in the sediments. After an eight-month period, some of the tracer was lost from both the organic-S and inorganic-S pools through oxidation or volatilization (Rudd, Kelly and Furutani, 1986). More tracer was lost from the inorganic-S pools, leaving the organic form as the predominant phase, particularly in the upper sediment horizons. Other studies have not observed large increases in organic-S compounds in sediments following S amendments. Studies in which SO₄^2- has been added to whole lakes (e.g. Lake 223; Figure 7.3a) or to intact sediment cores (Giblin et al., 1990) revealed that the inorganic S phase exhibited increases in hypolimnetic sediments. In an examination of S sediment accumulation in the hypolimnetic sediments of a lake in the Adirondacks (New York, USA), White et al. (1989) showed that both inorganic-S and organic-S concentrations have increased, but inorganic S has increased more during the past 100 years. In an experimentally acidified lake, Kelly and Rudd (unpub. data) found that in epilimnetic cores, the sulphur increase near the surface was due to an increase in both inorganic and organic S. In hypolimnetic cores, the increase was due mostly to inorganic S. Thus, diverse patterns can be found within the same lake.

The study of the factors affecting organic-S accumulation has been impeded by the lack of a direct analytical method. Organic S is typically estimated by subtracting inorganic S (usually measured by using the chromium reduction method of Howarth and Merkel, 1984; see Giblin and Wieder, this volume) from total S. Total S may be measured on wet samples with an alkaline oxidation (Tabatabai and Bremner, 1970) followed by reduction to H₂S (Johnson and Nishita, 1952; Johnson and Ulrich, 1959). Because this method is labour intensive, total S is more frequently analysed by drying samples and combusting the S to SO₂, which can then be measured by fluorescence or infrared spectrophotometry .However, drying sediments can lead to signi- ficant and variable losses of total S (Amaral et al., 1989), especially when lyophilization is used as the drying method. Drying also leads to a change in stable S isotope content (Amaral et al., 1989), a result which shows that the loss is selective rather than equally distributed among the various S species. Other investigators have shown that drying does not lead to loss of inorganic- S content (Canfield et al., 1986; Carignan and Tessier, 1988; Morse and Cornwell, 1987). The stability of inorganic S with respect to drying suggests that the losses from the total S pool are from the organic fraction and that organic S estimated by difference may be too low. Organic-S volatilization upon drying is not as much of a problem in studies using ³⁵S. In these studies, the organic ³⁵S is measured by aqua regia digestion (Howarth and Teal, 1979), following removal of inorganic ³⁵S, with all analyses usually performed quickly on wet sediments. A standard technique is needed for the assay of organic S, as well as for all sedimentary S species, to allow intercomparison of results and determination of the factors affecting organic-S accumulation in lake sediments.

7.2.3.3 Oxidation to Elemental Sulphur and Higher Oxidation States

Dissolved and solid sulphide compounds are readily oxidized both chemically (Avrahami and Golding, 1968; Chen and Morris, 1972; Pankow and Morgan, 1980) and biologically (Schiff and Hodson, 1973; Roy and Trudinger, 1970). In both oxidation mechanisms, polysulphides, S⁰, S₂O₃^-, SO₄^2-, and SO₄^2- may be formed (Luther and Church, this volume). Although SO₄^2- is the most stable phase, the others are metastable forms that may persist for long periods of time.

Biological oxidation of sulphide and other reduced S species (thiosulphate, thiocyanate, elemental sulphur, and tetrathionate) is carried out by both the chemotrophic sulphur oxidizers, which may use reduced sulphur as an energy source and fix CO₂, and the phototrophic green and purple bacteria, which use reduced sulphur as an electron donor in the fixation of CO₂ (Table 7.1). In some lakes, these activities can be significant sources of organic matter productivity (Parkin and Brock, 1981), particularly meromictic salt lakes (Gorlenko, 1978). The relative potential of these reduced S species to be oxidized and the environmental factors that regulate the extent of oxidation have not been studied nearly as much as SO₄^2- reduction and are much less understood.

Oxidation of reduced S occurs at interfaces: the oxic-anoxic interface for chemical and chemotrophic oxidation and, in the anoxic zone, the light-dark interface for phototrophic oxidation (Table 7.1) .Such interfaces may occur within or on sediments or in the water column during stratification. Depending on the depth of light penetration, phototrophic bacteria may grow below the chemoautotrophic bacteria (Gorlenko, 1978). The phototrophic bacteria may migrate up and down in the water column, following the diurnal changes in depth of maximum concentration of reduced sulphur (Sorokin, 1970), or the bacteria may remain at fixed depth (Parkin and Brock, 1981). During the winter period, the oxic-anoxic interface in sediments moves downward, due to lower rates of microbial oxygen consumption, and reduced S that accumulated during the antecedent anoxic period is subject to partial oxidation (Rudd, Kelly and Furutani, 1986; Kling et al., 1991), with attendant effects on the partitioning of S species.

7.2.4 FORMATION OF VOLATILE SULPHUR SPECIES

The release of volatile S species is a potential pathway for export of S from lakes. Volatile S species include H₂S and a variety of organic-S compounds, including carbonyl sulphide (COS), dimethyl sulphide (DMS), methane thiol (MSH), dimethyl disulphide (DMDS), and carbon disulphide (CS₂). In general, these organic-S gases are more soluble than methane and even carbon dioxide, but they are volatile at lake temperatures and may contribute more to water-air fluxes than H₂S because of their greater stability in the presence of oxygen.

Most research on the water-air exchange of S gases has been concerned with DMS in marine systems (Andreae and Barnard, 1984; Andreae and Jaeschke, this volume) and with emissions from salt marshes, soils, and vegetation (e.g. Steudler and Peterson, 1984; Goldan et al., 1987; Giblin and Wieder, this volume; Andreae and Jaeschke, this volume). In marine systems, DMS is present in ocean surface waters at concentrations ranging from 0.031 to 34.3 nM (1 to 1100 ng (S) l^-1), with a mean of 2.6 nM (83 ng (S) l^-1) (Andreae and Barnard, 1984; Turner et al., 1988).

Few measurements have been made on volatile S production and release from freshwater systems. Holdway and Nriagu (1988) measured DMS concentrations in Lake Ontario surface waters ranging from 0.6 to 1.4 nM (20 to 45 ng (S) l^-1). Concentrations of DMS measured in Lake Mendota, a eutrophic Wisconsin lake, were much lower (0.017 to 0.16 nM; 0.5 to 5.2 ng (S) l^-1; Zinder and Brock, 1978), whereas two Antarctic lakes had higher concentrations (3.2 to 4.8 nM; 100 to 150 ng (S)  l^-1; Deprez, Franzmann and Burton, 1986). A recent survey of a variety of volatile S species in the epilimnia of Canadian Shield lakes during summer stratification showed average summer DMS values ranging from 0.48 to 1.6 nM (15 to 51 ng (S) l^-1; Richards, Kelly and Rudd, 1991; Table 7.4). This same study also showed low concentrations of COS (0.091 to 0.38 nM; 2.9 to 12 ng (S) l^-1) as well as MSH (0.018 to 1.1 nM; 0.58 to 35 ng (S) l^-1) and DMDS (0.02 to 0.097 nM; 0.64 to 3.1 ng (S) l^-1). Dimethyl disulphide may be produced by MSH oxidation after sample collection, so that measured concentrations of both gases may not reflect in situ concentrations. Much higher values of S gases were found in one shallow unstratified lake, with DMS averaging 6.6 nM (210 ng (S) l^-1), COS 0.61 nM (20 ng (S) l^-1), DMDS 1.4 nM (45 ng (S) l^-1), and MSH 8.6 nM (275 ng (S) l^-1) (Lake 114; Table 7.4). These concentrations were as high as those measured in a nearby bog pool, and in other Ontario bogs (Nriagu, Holdway and Coker, 1987), demonstrating a high potential flux for at least some freshwater systems.

The source of volatile S species in marine waters is primarily algal (Andreae and Jaeschke, this volume), but in freshwaters a mixture of water column and sediment sources may exist, depending on algal physiology and species and lake bathymetry .Some freshwater algal species have been shown to produce DMS under axenic growth conditions (Bechard and Rayburn, 1979). The absence of DMS production by many freshwater algae might be because dimethylsulphonium propionate (DMSP), the algal-produced precursor of DMS, is thought to be an osmolyte in high-salinity water. In support of this function of DMSP, Iversen, Nearhoof and Andreae (1989) observed that DMSP and DMS concentrations increased with increasing salinity along a brackish-marine salinity gradient. Howarth and Stewart (this volume) have suggested that the low rate of DMS production by algae in most freshwaters may also reflect a relatively high abundance of molybdenum (indicated by a low sulphate: molybdenum ratio; Howarth, Marino and Cole, 1988) in these waters.

Table 7.4. Estimates of gaseous sulphur flux and its significance in lake sulphate budgets. From Richards, Kelly and Rudd (1991)

Production of volatile S, primarily DMS, is also known to occur in decomposing freshwater algal cultures (Bechard and Rayburn, 1979) and in decomposing algal mats (Zinder, Deomel and Brock, 1977). Production of volatile S species during decomposition in other systems (e.g. bacterial cultures, soils, and manures) is well known (Bremner and Steele, 1978). Decomposition of methionine has been studied by adding it to anoxic sediments from a lake (Zinder and Brock, 1978) and a salt marsh (Kiene and Visscher, 1987) .In these studies the major volatile S product was MSH, along with small amounts of DMS. Both MSH and DMS were present in concentrations of 10 to 120 nM (320 to 3840 ng (S) l^-1) in the anoxic porewater of several lakes on the Precambrian Shield (Richards, Kelly and Rudd, 1991), at levels similar to marine porewater DMS concentrations (Andreae, 1985) and 2 to 14 times higher than in the overlying water. Bubbles produced from the sediments of an Arctic lake contained COS, DMS, and CS₂ (Hines and Morrison, 1989). Thus, in small shallow lakes such as Lake 114 (Table 7.4), volatile S release from sediments may be the primary source of S gases in surface water. However, the similarity of volatile S gas concentrations in most lakes with mean depths greater than a few metres (e.g. Lake 3O₂S and Lake 239, Table 7.4; Lake Ontario, Holdway and Nriagu, 1988), suggests that either algal production or some low rate of organic matter decomposition in the water column is supporting the concentrations observed in deeper lakes.

Richards, Kelly and Rudd (1991) examined the effect of concentration of SO₄^2- on surface water volatile S production by comparing two lakes with similar bathymetries but contrasting chemistries: one lake with low SO₄^2- and circumneutral pH and one with high SO₄^2- and low pH. Volatile S concentrations over the summer were virtually indistinguishable in surface waters of the two lakes (Figure 7.6), suggesting no effect of either SO₄^2- concentration or pH. However, the hypolimnetic production rate of DMS was much greater in the lake with greater SO₄^2- concentrations. If the absence of change in surface water volatile S compounds over the range of freshwater SO₄² concentrations is borne out, present-day measurement of volatile S compounds in the atmosphere may be used to estimate natural (non- anthropogenic) sources of atmospheric S from lakes.

Volatile S compounds are catabolized, as well as synthesized, in sediments. Zinder and Brock (1978) were only able to detect the production of volatile S compounds in sediments by suppressing breakdown of these compounds with chloroform, an inhibitor of C-1 metabolism. Both MSH and DMS are broken down to CO₂, CH₄, and H₂S (Zinder and Brock, 1978; Kiene et al., 1986). Oremland et al. (1989) isolated a methanogenic bacterium that was able to grow on DMS. Also, phototrophic purple bacteria have been shown to oxidize DMS to DMSO (Zeyer et al., 1987). Undoubtedly, other transformations of volatile S species are carried out by bacteria. Zinder and Brock (1978) suggested that sediments may be a sink for volatile S species, rather than a source. However, the porewater measurements mentioned above suggest that sediments are a net source of volatile organic-S species in lakes.

Figure 7.6. Surface water concentrations of inorganic and organic volatile S species in Lake 302 South and in Lake 226 Southwest during 1988. The SO₄^2- concentration in Lake 302 South was 106 to 135 µmol L^-1 and the pH was 4.5 to 4.6 during 1988. In Lake 226 Southwest, SO₄^2- was 26 to 28 µmol L^-1 and the pH was 7.0 during 1988. From Richards, Kelly and Rudd (1991)

The volatile organic-S species are present in porewater in nanomolar concentrations. In contrast, H₂S is often present in anoxic zones in micromolar concentrations (Cook, 1984; Nurnberg, 1984; Carignan, 1988). In oxic zones, H₂S is unstable and is oxidized with a half-life of 50 hours (Millero et al., 1987; Millero and Hershey, 1989) .The organic-S species are stable in cold oxic water for up to 24 hours (Richards, Kelly and Rudd, 1991). Because of its low concentration and chemical instability, there are probably few reliable measurements of H₂S concentration in oxic surface waters. Therefore, the importance of fluxes of H₂S in relation to the volatile organic-S species is difficult to assess.

The relative role of gaseous S flux in whole-lake budgets has recently been assessed in one lake and appears to be a minor term. As discussed below, up to 50% of incoming SO₄^2- may be removed from the water column by all mechanisms, and it is not known whether the sediments are the sole sink for this S. Some S lost from the water column could be converted into volatile compounds and lost to the atmosphere. However, in a whole-lake ³⁵SO₄^2- addition experiment, Stuiver (1967) was able to account for all of the ³⁵S added to the lake either in the sediments or the water column, indicating little if any volatilization. This agrees with Richards, Kelly and Rudd (1991) who compared estimates of volatile organic-S fluxes with SO₄^2- budgets for three Canadian Shield lakes (Table 7.4). These fluxes were insignificant compared to SO₄^2- reduction rates (Table 7.2) or to whole-lake S retention estimated from mass balances (Table 7.4).

An additional issue is the role of volatile S fluxes from lakes as a source of S to the atmosphere. The lake-air fluxes of organic-S gases for a shallow unstratified lake (Lake 114; Table 7.4) are similar on an areal basis to fluxes from tropical forests (Andreae and Andreae, 1988) and are about one-third of the mean areal flux from the ocean (Andreae and Raemdonck, 1983). The range of S fluxes from lakes (Table 7.4) is similar to the range measured from midcontinent soils in the United States (Goldan et al., 1987; Lamb et al., 1987). Because Lake 114 is located in a low SO₄^2- deposition area (13 to 43 µmol m^-2 day^-1; 0.42 to 1.4 mg (S) m^-2 day^-1; Linsey, Schindler and Stainton, 1987), as much as 10 to 30% of atmospheric S deposition falling directly on the lake surface may be returned to the atmosphere by gas exchange of organic-S compounds. Some areas of North America (e.g. Hudson Bay lowlands), northern Europe, and northern Asia contain many shallow unstratified lakes that may have rates of volatilization comparable to those for Lake 114. Also, lakes in which SO₄^2- concentrations are unusually high due to evaporation (> 20 g l^-1; 0.6 M) have recently been shown to have DMS fluxes to the atmosphere that are greater than those in the ocean (Richards, Rudd and Kelly, in review). The importance of the lake-to-airflux of S gases for the S budgets of these lakes and as an input to the atmosphere deserves further attention. However, because lakes cover a small percentage of the surface area of the continents, lakes are probably a very small source of sulphur to the atmosphere.

7.3 SULPHUR CYCLING IN WHOLE LAKES

The previous section describes the various reactions that S undergoes in small lakes. In this section, these reactions will be put into a whole-lake perspective by examining S budgets and a simple model relating inputs, outputs, and retention of S in small lakes.

7.3.1 WHOLE-LAKE SULPHUR BUDGETS

Construction of whole-lake S budgets has emphasized inputs and outputs of SO₄^2- , primarily because SO₄^2- is the most abundant and most commonly measured S species. David and Mitchell 1985 were the only workers to examine other forms of S in addition to SO₄^2-; they found that organic S was a minor constituent in the inputs and outflows of South Lake (New York).

From a whole-lake perspective, SO₄^2- is removed from the water column and retained in the lake sediments by two processes: bacterial SO₄^2- reduction, followed by iron sulphide or organic sulphide formation, and sedimentation of organic S from the water column. The net sedimentation of organic S is difficult to quantify because rates of mineralization are not known and are indirectly estimated, as discussed above. Net retention of sulphides derived from bacterial SO₄^2- reduction is also difficult to quantify. In systems in which bacterial SO₄^2- reduction has been measured directly, oxidation of reduced S species after formation may be significant and is rarely measured (Rudd, Kelly and Furutani, 1986). Usually net loss of S from the water column is calculated from a mass balance by difference, and includes both organic S sedimentation and accumulation of S formed from bacterial SO₄^2- reduction.

Whole-lake S budgets have been calculated for six lakes that have been investigated for the mechanisms of S removal (Table 7.5). For three lakes in midcontinental North America (Lake 223, Lake 3O₂N, and Lake 3O₂S, all of which are located at ELA), SO₄^2- is removed from the water column, primarily by bacterial SO₄^2- reduction, at large areal rates. Part of the S removal is due to the enhancement of bacterial SO₄^2- reduction from experimental additions of SO₄^2-. Rates of SO₄^2- reduction for these three lakes are higher than that for Little Rock Lake (Wisconsin) and South Lake (New York) (Table 7.5). Interestingly, although seasonal rates of bacterial SO₄^2- reduction in Lake 227 were quite high because of stimulation by eutrophication (73 mmol m^-2 a^-1; 2.3 g (S) m^-2 a^-1; Table 7.2), the reduced S was not sequestered efficiently by the hypolimnetic sediments, because the annual rate of SO₄^2- reduction and retention was estimated to be only 2 mmol m^-2 a^-1 (0.06 g (S) m^-2 a^-1. Apparently, a fraction of the S derived from bacterial SO₄^2- reduction remained in the hypolimnion as H₂S, to be oxidized to SO₄^2- at fall and spring overturn (Cook, 1981).

The rates of S removal due to bacterial SO₄^2- reduction were variable among the lakes, yet rates of organic-S sedimentation were similar among the lakes, ranging from 3 to 11 mmol m^-2 a^-1 (0.1 to 0.35 g (S) m^-2 a^-1) (Table 7.5). The rates of organic-S sedimentation for Lake 3O₂N and 3O₂S are gross sedimentation, because estimates of recycling at the sediment-water inter- face were not made.

Except for South Lake, these lakes retain S at relatively high rates (Table 7.5), so that a large proportion of the SO₄^2- input to the lake is removed from the water column. Differences among lakes in the proportion of input SO₄^2- retained are related to differences in morphometry and hydrology of the lakes and are discussed below.

Table 7.5. Whole-lake annual sulphur budgets for six lakes, showing the processes by which sulphur is removed from the water column and retained in the sediments on a net basis. Units are mmol m^-2 a^-1

7.3.2 MODELLING SULPHATE REMOVAL IN LAKES

An apparent dichotomy that arose during research on SO₄^2- removal in acidified lakes was the great difference among lakes in the importance of in- lake SO₄^2- retention in whole-lake S budgets. As mentioned above, in the first ELA acidification experiment (Lake 223), 40% of the SO₄^2- input to the lake was biologically removed from the water column and retained in the lake sediments (Cook et al., 1986). This finding was in marked contrast to some atmospherically acidified lakes where little or none of the SO₄^2- input was removed within the lake (Wright, 1983; Galloway et al., 1983). This difference between the experimentally and atmospherically acidified lakes was not solely due to the presence of an anoxic hypolimnion in Lake 223, because the anoxic hypolimnion is less than 2% of the area of Lake 223. Furthermore, SO₄^2- reduction was also occurring in littoral sediments of Lake 223 at rates similar to those in the profundal sediments (Kelly and Rudd, 1984; Cook et al., 1986). When SO₄^2- reduction rates were measured directly in a variety of acidified lakes in North America and northern Europe, areal rates were found to be fairly similar from lake to lake (Rudd et al., 1986); the SO₄^2- reduction rates collected in Table 7.2 support this broad similarity among many lakes. The relatively small range in SO₄^2- reduction rates led to the hypothesis that the most important factors controlling whole-lake SO₄^2- retention in acidified lakes were water residence time and mean depth (i.e. the extent of contact between the water and the sediments) (Kelly et al., 1987; Baker et al., 1986).

The models developed for SO₄^2- removal in lakes with different residence times and mean depths are mathematically similar to those for phosphorus removal (e.g. Dillon, 1974). Conceptually, however, the models are different in that the phosphorus removal is essentially a water column (algal) process, whereas usually SO₄^2- removal via bacterial reduction is a sediment process (Cook et al., 1986; Rudd et al., 1990). The key parameters for the S models are water residence time, mean depth, and a mass transfer coefficient that quantifies the areal rate of SO₄^2- reduction by the sediments (Kelly et al., 1987; Baker, Brezonik and Pollman, 1986).

The model formulation assumes that SO₄^2- reduction rate is first order with respect to SO₄^2- concentration. Sulphate reduction in sediments has been shown to be first order within a lake as SO₄^2- concentration increases. In comparing many lakes, Rudd et al. (1986) found that the concentration-rate relationship was weaker among lakes than that within one lake; however, the concentration range of the measurements was relatively small (30 to 120 µM; 0.1 to 3.8 mg (S) l^-1) and thus was not suitable for observing differences among lakes. The average mass transfer coefficient [S_s = annual rate of S reduction (mmol m^-2 a^-1) divided by the SO₄^2- concentration (mmol m^-3)] calculated from these direct measurements was 0.36 to 0.17 m a^-1 (Kellyet al., 1987). This mass transfer is analogous to a piston velocity for SO₄^2- flux from the water column to the sediments. The proportion of input SO₄^2- removed within the lake (Rs) is then calculated from

where q_a is the areal outflow (m a^-1) and is equal to mean depth (m) divided by the outflow-based water residence time (year). S_s can also be calculated for lakes where SO₄^2- removal has been measured by mass balance budgets, by rearranging Equation (7.1). When calculated this way, good agreement was reached by two independent groups: 0.54 to 0.13 m a^-1 (n = 8; Kelly et al., 1987) and 0.52 to 0.34 m a^-1 (n = 14; Baker and Brezonik, 1988). Use of these values predicts SO₄^2- removal very well (Figure 7.7; Kelly et al., 1987; Baker and Brezonik, 1988).

The cause of the slightly higher mean mass transfer coefficients calculated from the whole-lake budgets relative to the direct measurements of SO₄^2- reduction is expected because of the contribution of algal uptake and sedimentation. In two experimental lake basins at ELA, where one had elevated SO₄^2- concentrations and one did not, the contribution of gross S sedimenta- tion to SO₄^2- removal was 14% in the low SO₄^2- basin and 6.5% in the high SO₄^2- basin (Rudd et al., 1990).

Figure 7.7. In-lake retention of SO₄^2- as a function of lakewater residence time. The model-predicted values are derived from Kelly et al. (1987) by using a mass transfer coefficient of 0.5 m a^-1

This model, and these mass transfer coefficients, are only suitable for relatively oligotrophic lakes without extensive anoxic hypolimnia (Kelly et al., 1987). One might expect lakes with large anoxic hypolimnia to have much larger mass transfer coefficients [e.g. Lake Malawi, Africa (Kelly et al., 1987)], because of the more rapid SO₄^2- transport from eddy diffusion across the oxic-anoxic interface in the sediments. However, Lake 227, which has a seasonally anoxic hypolimnion, has a relatively high mass transfer coefficient on a seasonal basis (5.5 m a^-1; Cook, 1981), yet has a low mass transfer coefficient derived from an annual whole-lake budget (0.14 m a^-1; Table 7.5). This annual mass transfer coefficient is smaller than that for oligotrophic lakes. In Lake 227, SO₄^2- was being reduced at high rates yet the efficiency of retention of reduced S was low. The H₂S produced from bacterial SO₄^2- reduction remained in the anoxic water column where it was oxidized to SO₄^2- at spring and fall overturn.

Lake Anna also does not fit the model because SO₄^2- is conveyed to the sediments by sedimentation of particles and not by diffusion, giving this lake an unusually high mass transfer coefficient (Mills, Bell and Herlihy, 1989). Lakes with low SO₄^2- reduction rates [e.g. South Lake (David and Mitchell, 1985) and McNearney Lake, discussed above], are also exceptions to the model. In these lakes, algal uptake and sedimentation may be the dominant mechanism of S retention. No predictive model has been developed for these lake systems and would be difficult to construct because the removal is probably related to algal biomass accumulation, which would be best predicted by phosphorus loading. Also, no a priori method exists to identify lakes that have low SO₄^2- reduction rates.

Another approach to determining the coefficients for SO₄^2- mass transfer is to use the net rate of accumulation of total S (from all mechanisms, including algal sedimentation and bacterial reduction) in the sediments as the measure of net SO₄^2- removal

where the sediment accumulation rate is in mmol m^-2 a^-1 and SO₄^2- is in mmol m^-3. The use of this technique in lakes experiencing changes in atmospheric deposition introduces unce.rtainty because the accumulation rate must be averaged for several decades, a period during which the lakewater SO₄^2- concentration has likely changed. Norton et al. (1988) measured S accumulation rates and SO₄^2- concentrations in a number of lakes in eastern North America. From the data of Norton et al., which include 15 lakes not used in the whole-lake budget approaches discussed above, we calculated a mass transfer coefficient of 0.8 to 0.5 m^-2 a^-1, using Equation (7.2). This is a slightly higher mean value than coefficients calculated from whole-lake budgets (0.5 to 0.3 m^-2 a^-1; Kelly et al., 1987; Baker and Brezonik, 1988), perhaps because the sediment accumulation rates were measured at the deepest points in the lakes, where S accumulation is likely to be greater than the average lake value. However, the overlap of the sediment accumulation and whole-lake budget data is strong support for the SO₄^2- removal model.

Extrapolation of in-lake SO₄^2- retention from a particular lake or group of lakes requires knowledge of mean depth and residence time. In-lake SO₄^2- retention will be larger for lakes with longer residence times and smaller for shorter residence time lakes (Equation (7.1); Figure 7.7). Norton et al. (1988) examined S accumulation rates and concluded that, in general, these rates represent a small component of the whole-lake S budget that would remove less than 10 to 20 µM (0.32 to 0.64mg (S) l^-1) of SO₄^2-. However, their data set included many lakes with short residence times, and their conclusions probably would have been different had more lakes with long residence times been included.

Some workers (Norton et al., 1988; Shaffer et al., 1988) have tried to simplify further the models of SO₄^2- removal by substituting the ratio of the watershed area to lake area and some generalized estimate of runoff for measurements of areal water loading (mean depth divided by water residence time) .These approaches can lead to serious errors unless the value used for runoff is close to the true net value. For example, the estimated q_a values using 1 m a^-1 precipitation (Norton et al., 1988) are an order of magnitude larger than the literature values of qa for two of the lakes examined by Norton : et al. [McNearney Lake (Cook et al., 1990) and Hustler Lake (Kanciruk et al., 1986)]. The use of too high a value for net precipitation leads to an underestimate of the in-lake SO₄^2- retention, as was done by Norton et al. (1988)., However, even though the most accurate calculation requires good hydrologic data, a rough estimate of water residence time will at least indicate whether SO₄^2- retention should be an important or a negligible process in a given lake.

7.4 INTERACTIONS OF THE C, N, P, AND S BIOGEOCHEMICAL CYCLES IN LAKES

The biogeochemistry of S is closely linked to biogeochemistry of C, N, and P (Howarth and Stewart, this volume). Without the link to the C cycle, the behaviour of S in small lakes would be conservative. In turn, sulphur's chemical forms, concentrations, and reactions affect the cycles of the other elements. Because of these interactions, the biogeochemistry of C, N, P, and S should not be examined in isolation.

A number of interactions of the S and C biogeochemical cycles have been treated above in Section 7.2. For example, methane fluxes in lakes may be affected by changes in the S cycle. Increasing SO₄^2- concentrations may cause increases in catabolism by bacterial SO₄^2- reduction and decreases in that by methanogenic bacteria (Figure 7.4; Kelly et at., 1982; Kelly, Furutani and Schindler, 1984). In addition to the effect of SO₄^2- on the rate of methane formation, SO₄^2- may increase the rate of anoxic methane oxidation, potentially causing a reduction in the flux of methane from the sediments to the water column and to the atmosphere.

SO₄^2--reducing microorganisms may be able to metabolize organic molecules (e.g. alkanes and low-oxygen content compounds) that are not suitable for mixed populations of methanogenic bacteria. Thus, increasing the relative importance of bacterial SO₄^2- reducers may result in more decomposition and recycling of organic matter and less burial of organic matter in lake sediments, although this has only been shown in peatland soils (Yavitt, Lang and Wieder, 1987).

Whereas addition of SO₄^2- to lakes probably only affects the pathway of decomposition (SO₄^2- reduction vs methanogenesis) rather than the total amount, additions of sulphuric acid may affect total decomposition if the pH becomes low enough. For example, acidification with sulphuric acid to pHs below 5 may inhibit decomposition (Kelly, Furutani and Schindler, 1984).

Another interaction of the S and C cycles in the production of HCO₃^- during bacterial SO₄^2- reduction. This production of HCO₃^- can affect the alkalinity of lakes and can counteract the effects of atmospheric deposition (Cook et al.,1986; Schindler et al.,1986; Howarth and Stewart, this volume). Increases in SO₄^2- deposition will cause increases in bacterial SO₄^2- reduction and increases in production of HCO₃^-, thereby providing a homeostatic mechanism for maintaining alkalinity concentrations and slowing the rate of acidification (Cook et al., 1986). The efficiency with which this homeostatis maintains alkalinity depends on the amount of SO₄^2- retained by the lake, which is a function of SO₄^2- concentration and hydraulic residence time, and the ability of the sediments to sequester reduced S [S(II)] in an inorganic or organic S form. In the Lake 223 experiment, the sediments retained nearly all of the S produced by bacterial SO₄^2- reduction, much of it as iron sulphides (Figure 7.3(a); Cook et al., 1986). These SO₄^2- losses counteracted the acidification by neutralizing between 66 and 81% of the added acid. The efficiency of S retention in the lake sediments also appears to be high in the Lake 302S experiment (Rudd et al., 1990). The rate of supply of iron to the sediments may eventually limit the ability of sediments to sequester S and produce alkalinity (Schindler, 1985; Carignan and Tessier, 1988; Giblin et al., 1990).

Alkalinity is also produced during ester sulphate formation, which does not involve reduction of SO₄^2-. Theoretical considerations show that ester sulphate formation may produce either one equivalent of alkalinity per SO₄^2- equivalent incorporated (Kelly and Rudd, 1989) or two equivalents when the esters contain base cations (Baker et al., 1989; Urban and Baker, 1989). However, the chemical identities and characteristics of these esters are unknown, and further research is needed on the relationship between ester sulphate formation and alkalinity production.

Changes in the S cycle caused either by increased inputs of nutrients (eutrophication) or by increased atmospheric SO₄^2- deposition may result in an increased rate of sulphide production. Accumulation of produced H₂S, a toxic substance, in the water column may have severe effects on a variety of organisms. In some lakes, sulphide is used by chemosynthetic and photosynthetic anaerobic bacteria in the production of organic matter (Howarth and Stewart, this volume). Primary production by these organisms should increase with increased rates of bacterial SO₄^2- reduction and H₂S accumulation.

The formation of organic-S compounds from reduced S species, which has been demonstrated for marine sediments (Francois, 1987; Vairavamurthy and Mopper, 1987, 1989; Luther and Church, this volume), is another interaction of the carbon and sulphur cycles. Although the formation of organic-S compounds in lacustrine sediments has not received adequate study, the synthesis of organic-S compounds may be limited by the amount and types of organic compounds present, as well as by the amount of reduced S, particularly polysulphides (Vairavamurthy and Mopper, 1987, 1989). Thus the trophic state of a lake may affect not only the rate of SO₄^2- reduction, as discussed above, but also the distribution of S into the various end-products.

There is at present no evidence that primary productivity is ever limited by availability of SO₄^2-. In most lakes, we can assume that productivity is determined by the supply of other nutrients. However, S may play an indirect role in controlling the supply of nutrients by affecting the mobility of P. Under conditions of increased SO₄^2- concentrations, enhanced iron sulphide formation may remove iron from the iron-phosphorus cycle and make it unavailable to bind phosphate (Wetzel, 1975; Baccini, 1985). Increased fluxes of phosphorus to the water column may result in increases in primary productivity. However, in a set of enclosure experiments in a softwater lake, Curtis (1989) found that P release was not related to sulphide accumulation of Fe(III) accumulation. In these experiments, low H⁺ concentrations alone or in combination with high sulphide concentrations were the factors controlling P release from sediments. In contrast, Caraco, Cole and Likens (1989) presented evidence for enhanced P release in lake sediments from New England with relatively low Fe content and relatively high S content. Caraco, Cole and Likens examined the relationship between P release and total sediment respiration, in a large set of lakes with a wide range of SO₄^2- concentrations. Relative P release was significantly correlated to SO₄^2- concentration under both oxic and anoxic conditions, suggesting that iron sulphide formation may be a controlling factor. However, sediment Fe alone did not explain the variation in relative P release. After SO₄^2-, conductivity was the next best predictor of relative P release. The importance of conductivity suggests that the observation could also be related to the findings of Mayer and Kramer (1986), who showed that softwater lake sediments adsorb p more efficiently than sediments from hardwater lakes. As Caraco, Cole and Likens (1989) state, the mechanism behind the empirical relationship between p release and SO₄^2- has yet to be sorted out.

Increasing SO₄^2- concentrations may affect the rates of nitrogen fixation, an important source of nitrogen for many lakes. Due to a similarity in stereochemistry between SO₄^2- and molybdate, SO₄^2- blocks the uptake of molybdenum, a necessary micronutrient for nitrogen-fixing organisms (Howarth, Marino and Cole, 1988; Howarth and Stewart, this volume). Inhibition of nitrogen fixation in freshwaters may alter the supply of nitrogen and cause a shift in nutrient limitation from phosphorus to nitrogen if the sulphate: molybdenum ratio is sufficiently high. Such an effect has been demonstrated in sulphate-rich salt lakes (Marino et al., 1990).

Another effect of the N cycle occurs when SO₄^2- enters freshwater as sulphuric acid and the pH of lake is lowered below about 5.4. When this occurs, nitrification is inhibited, causing accumulation of NH₄⁺ (Rudd et al., 1988). In addition, because the N cycle is blocked and does not proceed from NH₄⁺ to NO₃^-, denitrification decreases, leading to a build-up of fixed nitrogen.

7.5 SUMMARY

The key features of the S cycle in dimictic freshwater lakes are assimilatory reduction and uptake of SO₄^2- in the water column by algae and bacteria; dissimilatory SO₄^2- reduction by bacteria in anoxic zones; reaction of reduced S species with organic matter and reduced metals (e.g. Fe²⁺); and production of S gases (COS, DMS, DMDS, and MSH), probably accompanying decomposition of S-bearing organic matter .

For many lakes, water column concentrations of SO₄^2- are great enough to provide all of the S required by algae and bacteria. Uptake of SO₄^2- by algae appears to be independent of SO₄^2- concentrations from 0.03 to 43 mM (1. to 1400 mg (S) l^-1). Seston S undergoes decomposition in the sediments, providing a supply of S and volatile S compounds to the overlying water. The production of S gases in lakes is independent of SO₄^2- concentrations and pH. DMS, COS, DMDS, and MSH are sparingly soluble and chemically reactive in water and thus are present in lake water in low (< 10 nM; < 320 ng (S) l^-1) concentrations. Water-air gas exchange of S gases may contribute significantly to atmospheric S concentrations, particularly in areas remote from industrial sectors and marine S sources.

Rates of bacterial SO₄^2- reduction depend on SO₄^2- concentrations and organic matter supply. Experimental additions of SO₄^2- to mesocosms and whole lakes stimulated the rate of SO₄^2- reduction, indicating that other factors (e.g. organic matter) were not limiting bacterial activity for these oligotrophic systems. However, the trophic status of lakes may also influence the extent of SO₄^2- reduction, with ultraoligotrophic systems not producing enough organic matter to support bacterial SO₄^2- reduction. Experimental eutrophication with P and N stimulated bacterial SO₄^2- reduction in Lake 227, probably because of the increased supply of labile organic matter .

The capacity of sediments to form metal sulphide compounds depends on the supply of reducible metals, especially Fe. Lakes studied in north central and eastern North America have sufficient supply of reducible Fe from sedimentation, so that iron is not limiting the formation of Fe-sulphides. For lakes studied in Quebec, rates of Fe accumulation in sediments are comparable in magnitude to rates of S accumulation, which suggests that Fe may be limiting the formation of Fe-sulphide compounds. Hydrogen sulphide may also react to form organic-S compounds. The factors controlling the relative amounts of metal sulphides and organic S formed H₂S are not known.

Sulphur budgets for five lakes show that rates of organic-S formation were similar, but retention of S produced from bacterialSO₄^2- reduction was highly variable, with 4 to 60% of the input S retained by this mechanism. Differences in the percentage of input S retained among these five lakes are primarily related to differences among lakes in water residence time and bathymetry .A simple steady-state model, based on inputs, outputs, and first- orderSO₄^2- reduction, successfully explains differences in the retention of inputSO₄^2-.

REFERENCES

Amaral, J A., Hesslein, R. H., Rudd, J. W. M. and Fox, D. E. (1989). Loss of total sulfur and changes in sulfur isotopic ratios due to drying of lacustrine sediments. Limnol. Oceanogr., 34,1351-8.

Anderson, R. F. and Schiff, S. L. (1987). Alkalinity generation and the fate of sulfur in lake sediments. Can. J. Fish. Aquat. Sci. , 44 (Suppl. 1), 188-93.

Anderson, R. F., Schiff, S. L. and Hesslein, R. H. (1987). Determining sediment accumulation and mixing rates using ²¹⁰Pb, ¹³⁷Cs, and other tracers: problems due to postdepositional mobility or coring artifacts. Can. J. Fish. Aquat. Sci., 44, (Suppl. 1),231-50.

Andersson, B. I. (1985). Chemical composition of water across the sediment-water interface in the acidified Lake Gårdsjön, SW Sweden. Ecol. Bull. (Stockholm), 37, 263-75.

Andreae, M. O. (1985). Dimethylsulfide in the water column and the sediment porewaters of the Peru upwelling area. Limnol. Oceanogr., 30, 1208-18.

Andreae, M. O. and Andreae, T. W. (1988). The cycle of biogenic sulfur compounds over the Amazon Basin. 1. Dry season. J. Geophys. Res., 93, 1487-97.

Andreae, M. O. and Barnard, W. R. (1984). The marine chemistry of DMS. Mar. Chem., 14, 267-79.

Andreae, M. O. and Jaeschke, W. A. This volume.

Andreae, M. O. and Raemdonck, H. (1983). Dimethyl sulfide in the surface ocean and the marine atmosphere: a global view. Science, 221, 744-7.

Avrahami, M. and Golding, R. M. (1968). The oxidation of sulfide ion at very low concentrations in aqueous solutions.J. Am. Chem. Soc., 1968, 647-51.

Baccini, P. (1985). Phosphate interactions at the sediment-water interface. In: Stumm, W. (Ed.). Chemical Processes in Lakes. Wiley Interscience, New York, pp. 189-205.

Baker, L. A. and Brezonik, P. L. (1988). Dynamic model of in-lake alkalinity generation. Water Resources Res., 24, 65-74.

Baker, L. A., Brezonik, P. L. and Edgerton, E. S. (1986). Sources and sinks of ions in a softwater acidic lake in Florida. Water Resources Res., 22, 715-22.

Baker, L. A., Brezonik, P. L., and Pollman, C. D. (1986). Model of internal alkalinity generation: sulfate retention component. Water Air Soil Pollut., 31, 89-94.

Baker, L. A., Urban, N. R., Brezonik, P. L. and Sherman, L. A. (1989). Sulfur cycling in a seepage lake. In: Saltzman, E. and Cooper, W. (Eds). Biogenic Sulfur in the Environment. American Chemical Society, Washington DC, pp. 79-100.

Bechard, M. J. and Rayburn, W. R. (1979). Volatile organic sulfides from freshwater algae. J. Phycol., 15, 379-83.

Berner, R. A. (1984). Sedimentary pyrite formation: an update. Geochim. Cosmochim. Acta, 48, 605-15.

Boudreau, B. P. and Westrich, J. T. (1984). The dependence of bacterial sulfate reduction on sulfate concentration in marine sediments. Geochim. Cosmochim. Acta, 48, 2503-16.

Brakke, D. F., Landers, D. H. and Eilers, J. M. (1988). Chemical and physical characteristics of lakes in the northeastern United States. Environ. Sci. Technol., 22, 155-63.

Bremner, J. M. and Steele, C. G. (1978). Role of microorganisms in the atmospheric sulfur cycle. In: Alexander; M. (Ed.). Adv. Microb. Ecol., Vol. 2, pp. 155-201.

Canfield, D. E., Raiswell, R., Westrich, J. T., Reaves, D. M. and Berner, R. A. (1986). The use of chromium reduction in the analysis of reduced inorganic sulfur in sediments and shales. Chem. Geol., 54, 149-55.

Capone, D. G. and Kiene, R. P. (1988). Comparison of microbial dynamics in marine and freshwater sediments: Contrasts in anaerobic carbon catabolism. Limnol. Oceanogr., 4, 725-49.

Caraco, N. F., Cole, J. J. and Likens, G. E. (1989). Evidence for sulphate-controlled phosphorus release from sediments of aquatic systems. Nature, 341, 316-18.

Carignan, R. (1985). Quantitative importance of alkalinity flux from the sediments of acidic lakes. Nature, 317, 158-60.

Carignan, R. (1988). Seasonal dynamics of sulfate and hydrogen sulfide near the sediment-water interface of an oligotrophic acid lake. Verh. Int. Ver. Theor. Angew. Limnol., 23, 106-15.

Carignan, R. and Tessier, A. (1988). The co-diagenesis of sulfur and iron in acid lake sediments of southwestern Quebec. Geochim. Cosmochim. Acta, 52, 1179-88.

Casagrande, D. J. and Ng, L. (1979). Incorporation of elemental sulphur in coal as organic sulphur. Nature, 282, 598-9.

Casagrande, D. J., Idowu, G., Friedman, A., Rickert, P., Siefert, K. and Schlenz, D. (1979). H₂S incorporation in coal precursors: origins of organic sulphur in coal. Nature, 282, 599-600.

Chen, K. Y. and Morris, J. C. (1972). Kinetics of oxidation of aqueous sulfide by O₂. Environ. Sci. Technol., 6, 529-37.

Cook, R. B. (1981). The biogeochemistry of sulfur in two small lakes. Ph.D. Thesis, Columbia University, New York, 248pp.

Cook, R. B. (1984). Distributions of ferrous iron and sulfide in an anoxic hypolimnion. Can. J. Fish. Aquat. Sci., 41, 286-93.

Cook, R. B. and Schindler, D. W. (1983). The biogeochemistry of sulfur in an experimentally acidified lake. Ecol. Bull. (Stockholm), 35,115-27.

Cook, R. B., Kelley, C. A., Schindler, D. W. and Turner, M. A. (1986). Mechanisms of hydrogen ion neutralization in an experimentally acidified lake. Limnol. Oceanogr., 31, 134-48.

Cook, R. B., Kelley, C. A., Kingston, J. C. and Kreis, R. G. (1987). Chemical limnology of soft water lakes in the Upper Midwest. Biogeochemistry, 4, 97-117.

Cook, R. B., Kreis Jr, R. G., Kingston, J. C., Camburn, K. E., Norton, S. A., Mitchell, M. J. , Fry, B. and Shane, L. C. K. (1990). Paleolimnology of McNearney Lake: an acidic lake in northern Michigan. Paleolimnology, 3, 13-34.

Curtis, P. J. (1989). Effects of hydrogen ion and sulphate on the phosphorus cycle of a Precambrian Shield lake. Nature, 337, 156-8.

Datko, A. H., Mudd, S. H., Giovanelli, J. and Macnicol, P. K. (1978). Sulfur- containing compound in Lemna perpusilla 6746 grown at a range of sulfate concentrations. Plant Physiol., 62, 629-35.

David, M. B. and Mitchell, M. J. (1985). Sulfur constituents and cycling in water, seston, and sediments of an oligotrophic lake. Limnol. Oceanogr., 30, 1196-207.

Davison, W. (1980). A critical comparison of the measured solubilities of ferrous sulphide in natural waters. Geochim. Cosmochim. Acta, 44, 803-8.

Deprez, P. P., Franzmann, P. D. and Burton, H. R. (1986). Determination of reduced sulfur gases in Antarctic lakes and seawater by gas chromatography after solid adsorbent preconcentration. J. Chromatogr., 362, 9-21.

Dillon, P. J. (1974). A critical review of Vollenweider's nutrient budget model and other related models. Water Resources Bull., 10, 969-89.

Dunnette, D. A. (1973). Chemical ecology of hydrogen sulfide production in freshwater lake sediment. Ph.D. Dissertation, University of Michigan, 123pp.

Dunnette, D. A. (1989). Origin of hydrogen sulfide in freshwater sediments. In: Saltzman, E. and Cooper, W. (Eds). Biogenic Sulfur in the Environment. American Chemical Society, Washington DC, pp. 72-8.

Eilers, J. M., Brakke, D. F., Landers, D. H. and Overton, W. S. (1989). Chemistry of lakes in designated wilderness areas in the western United States. Environ. Monit. Assess., 12, 3-21.

Fee, E. J. (1979). A relation between lake morphometry and primary productivity and its use in interpreting whole-lake eutrophication experiments. Limnol. Oceanogr., 24, 401-16.

Francois, R. (1987). A study of sulfur enrichment in the humic acid fraction of marine sediments during early diagenesis. Geochim. Cosmochim. Acta, 51, 17-27.

Fry , B. (1986). Stable sulfur isotopic distributions and sulfate reduction in lake sediments of the Adirondack Mountains, New York. Biogeochemistry, 2, 329-43.

Fry, B. (1990). Sulfur stable isotope studies of lake cores from the Paleoecological Investigation of Recent Lake Acidification project. J. Paleolimnol., 4, 23-29.

Galloway, J. N. , Schofield, C. L. , Peters, N. E. , Hendrey, G. R. and Altwicker, E. R. (1983). Effect of atmospheric sulfur on the composition of three Adirondack lakes. Can. J. Fish. Aquat. Sci., 10, 799-806.

Giblin, A., Peterson, B. J. and Fry, B. (1989). Controls on sulfur storage in lake sediments: the importance of iron and organic matter. Abstracts of the Annual Meeting, American Society of Limnol. and Oceanogr., Fairbanks, Alaska, June 1989.

Giblin, A. E. and Wieder, R. K. This volume.

Giblin, A. E., Likens, G. E., White, D. and Howarth, R. W. (1990). Sulfur storage and alkalinity generation in New England lake sediments. Limnol. Oceanogr., 35, 852-69.

Goldan, P. D., Kuster, W. C., Albritton, D. L. and Fehsenfield, F. C. (1987). The measurement of natural sulfur emissions from soils and vegetation: three sites in the eastern United States revisited. J. Atmos. Chem., 5, 439-67.

Gorham, E., Lund, J. W. G., Sanger, J. E. and Dean, W. E. (1974). Some relationships between algal standing crop, water chemistry , and sediment chemistry in the English Lakes. Limnol. Oceanogr, 19, 601-17.

Gorlenko, V. M. (1978). Phototrophic sulfur bacteria in salt meromictic lakes and their role in the sulfur cycle. In: Krumbein, W. E. (Ed.). Environmental Biogeoche mistry and Geomicrobiology. Vol. 1: The Aquatic Environment. Ann Arbor Science, Ann Arbor, MI, pp. 109-19.

Herlihy, A. T., Mills, A. L., Hornberger, G. M. and Bruckner, A. E. (1987). The importance of sediment sulfate reduction to the sulfate budget of an impoundment receiving acid mine drainage. Water Resources Res., 23, 287-92.

Hines, M. E. and Morrison, M. C. (1989). Emissions of biogenic sulfur compounds from Alaskan tundra. Eos Trans. Am. Geophys. Union, 70, 284.

Holdren, G. R., Brunelle, T. M., Matisoff, G. and Whalen, M. (1984). Timing the increase in atmospheric sulphur deposition in the Adirondack Mountains. Nature, 311, 245-8.

Holdway, D. A. and Nriagu, J. O. (1988). A purge and trap gas chromatographic method for dimethyl sulfide in freshwater. Intern. J. Environ. Anal. Chem., 32, 177-86.

Howarth, R. W. (1979). Pyrite: Its rapid formation in a salt marsh and its importance in ecosystem metabolism. Science, 203, 49-51.

Howarth, R. W., Marino, R. and Cole, J. J. (1988). Nitrogen fixation in freshwater, estuarine, and marine ecosystems. 2. Biogeochemical controls. Limnol. Oceanogr. 33, 688-701.

Howarth, R. W. and Merkel, S. (1984). Pyrite formation and the measurement of sulfate reduction in salt marsh sediments. Limnol. Oceanogr., 9, 598-608.

Howarth, R. W. and Stewart, J. W. B. This volume.

Howarth, R. W. and Teal, J. M. (1979). Sulfate reduction in a New England salt marsh. Limnol. Oceanogr., 28, 70-82.

Ingvorsen, K. and Brock, T. D. (1982). Electron flow via sulfate reduction and methanogenesis in the anaerobic hypolimnion of Lake Mendota. Limnol. Oceanogr., 27, 559-64.

Ingvorsen, K. and Jųrgensen, B. B. (1984). Kinetics of sulfate uptake by freshwater and marine species of Desulfovibrio. Arch. Microbiol., 139,61-6.

Ivanov, M. V. (1968). Microbiological Processes in the Formation of Sulfur Deposits. Israel Programme for Scientific Translations, Ltd. 298pp.

Iversen, R. L., Nearhoof, S. L. and Andreae, M. O. (1989). Production of dimethyl sulfonium propionate and dimethylsulfide in estuarine and coastal waters. Limnol. Oceanogr., 34,53-67.

Jensen, H. S. and Andersen, F. O. (1987). Sulfate reduction in a coastal lake during summer stratification. Arch. Hydrobiol., 108, 465-81

Johnson, C. M. and Nishita, H. (1952). Microestimation of sulfur in plant materials, soils, and irrigation waters. Anal. Chem., 24, 736-42.

Johnson, C. M. and Ulrich, A. (1959). Analytical methods for use in plant analysis. Calif Agric. Exp. Stn. Bull., 766, 55.

Jones, J. G., Simon, B. M. and Roscoe, J. V. (1982). Microbiological sources of sulphide in freshwater lake sediments. J. Gen. Microbiol , 128, 2833-9.

Kanciruk, P., Eilers, J. M., McCord, R. A. Landers, D. H., Brakke, D. F. and Linthurst, R. A. (1986). Characteristics of Lakes in the Eastern United States. Vol. III: Data Compendium of Site Characteristics and Chemical Variables. EPA/600/4- 86/007c. US Environmental Protection Agency, Washington DC.

Kelly, C. A., Furutani, A. and Schindler, D. W. (1984). Effects of lake acidification on rates of organic matter decomposition in sediments. Limnol. Oceanogr., 29, 687-94.

Kelly, C. A. and Rudd, J. W. M. (1984). Epilimnetic sulfate reduction and its relationship to lake acidification. Biogeochemistry , 1, 63-77.

Kelly, C. A. and Rudd, J. W. M. (1989). Reply to comment by Urban and Baker , Limnol. Oceanogr., 34, 1146-8.

Kelly, C. A., Rudd, J. W. M. and Schindler, D. W. (1988). Carbon and electron flow via methanogenesis,SO₄^2-, NO^3-, Fe³⁺, and Mn⁴⁺ reduction in the anoxic hypolimnia of three lakes. Arch. Hydrobiol. Beih. Ergebn. Limnol., 31, 333-44.

Kelly, C. A., Rudd, J. W. M., Cook, R. B. and Schindler, D. W. (1982). The potential importance of bacterial processes in regulating the rate of lake acidification. Limnol. Oceanogr., 27, 868-82.

Kelly, C. A., Rudd, J. W. M., Hesslein, R. H., Schindler, D. W., Dillon, P. J., Driscoll, C. T., Gherini, S. A. and Hecky, R. E. (1987). Prediction of biological acid neutralization in acid-sensitive lakes. Biogeochemistry, 3, 129-40.

Kiene, R. P. and Visscher, P. T. (1987). Production and fate of methylated sulfur compounds from methionine and dimethylsulfonium propionate in anoxic salt marsh sediments. Appl. Environ. Microbiol., 53, 2426-34.

Kiene, R. P., Oremland, R. S., Catena, A., Miller, L. G. and Capone, D. G. (1986). Metabolism of reduced methylated sulfur compounds in anaerobic sediments and by a pure culture of an estuarine methanogen. Appl. Environ. Microbiol , 52, 1037-45.

King, G. M. and Klug, M. J. (1980). Sulfhydrolase activity in sediments of Winter- green Lake, Kalamazoo County, Michigan. Appl. Environ. Microbiol., 39, 950-6.

King, G. M. and Klug, M. J. (1982). Comparative aspects of sulfur mineralization in sediments of a eutrophic lake basin. Appl. Environ. Microbiol., 43, 1406-12.

Kipphut, G. W. (1978). An investigation of sedimentary process in lakes. Ph.D. Thesis, Columbia University, New York, 180pp.

Kling, G. W., Giblin, A. E., Fry, B. and Peterson, B. J. (1991). The role of seasonal turnover in lake alkalinity dynamics. Limnol. Oceanogr., 36, 106-22.

Knauss, H. J. and Porter, J. W. (1954). The absorption of inorganic ions by Chlorella pyrenoidosa. Plant Physiol , 29, 229-34.

Kuenen, J. G. and Tuovinen, O. H. (1981). The genera Thiobacillus and Thiomicorspira. In: The Prokaryotes. Springer-Verlag, Berlin, pp. 1023-36.

Kuivila, K. M., Murray, J. W., Devol, A. H. and Novelli, P. C. (1989). Methane production, sulfate reduction, and competition for substrates in the sediments of Lake Washington. Geochim. Cosmochim. Acta, 53, 409-16.

Lamb, B., Westberg, H., Allivine, G., Bamesberger, L. and Guenther, A. (1987). Measurement of biogenic sulfur emissions from soils and vegetation: application of dynamic enclosure methods with Natusch filter and GC/FPD analysis. J. Atmos. Chem., 5, 469-91.

Landers, D. H. and Mitchell, M. J. (1988). Incorporation of ³⁵SO₄^2- into sediments of three New York lakes. Hydrobiol., 160, 85-95.

Landers, D. H., Overton, W. S., Linthurst, R. A. et al., (1988). Eastern Lake Survey. Environ. Sci. Technol., 22, 128-35.

Landers. D. H., Ford, J., Eilers, J., Mitch, M., Morrison, M., Newell, A. and Hurley, K. (1989). Regional evaluation of lakes on the Kenai Peninsula of Alaska. Abstracts of the Annual Meeting, American Society of Limnology and Oceanography. Fairbanks, Alaska, June 1989.

La Riviere, J. W. M. and Schmidt, K. (1981). Morphologically conspicuous sulfur- oxidizing eubacteria. In: The Prokaryotes. Springer-Verlag, Berlin.

Linsey, G. A., Schindler, D. W. and Stainton, M. P. (1987). Atmospheric desposition of nutrients and major ions at the Experimental Lakes Area in northwestern Ontario, 1970-1982. Can. J. Fish. Aquat. Sci., 44, (Suppl. 1), 206-14.

Lovley, D. R. (1987). Organic matter mineralization with the reduction of ferric iron: a review. Geomicrobiol. J., 5, 375-99.

Lovley, D. R. and Klug, M. J. (1986). Model for the distribution of sulfate reduction and methanogenesis in freshwater sediments. Geochim. Cosmochim. Acta, 50, 11- 18.

Lovley, D. R. and Phillips, E. J. P. (1986). Availability of ferric iron for microbial reduction in bottom sediments of the freshwater tidal Potomac River. Appl. Environ. Microbiol., 52, 751-7.

Luther, G. W. and Church, T. M. This volume.

Marino, R., Howarth, R. W., Shamess, J. and Prepas, E. (1990). Molybdenum and sulfate as controls on the abundance of nitrogen-fixing cyanobacteria in saline lakes in Alberta. Limnol. Oceanogr., 35, 245-59.

Martens, C. S. and Jannasch, H. W. (1983). Cycling of metabolizable C, N, P, and S in organic-rich marine sediments. In: Bolin, B. and Cook, R. B. (Eds). The Major Biogeochemical Cycles and their Interactions. Wiley, Chichestesr, pp. 509-15.

Mayer, T. and Kramer, J. R. (1986). Effect of lake acidification on the adsorption of phosphorus by sediments. Water Air Soil Pollut., 31, 949-58.

Millero, F. J. and Hershey, J. P. (1989). Thermodynamics and kinetics of hydrogen sulfide in natural waters. In: Saltzman, E. and Cooper, W. (Eds). Biogenic Sulfur in the Environment. American Chemical Society, Washington DC, pp. 282-313.

Millero, F. J., Hubinger, S., Fernandez, M. and Garnett, E. S. (1987). Oxidation of H₂S in seawater as a function of temperature, pH, and ionic strength. Environ. Sci. Technol., 21, 439-43.

Mills, A. L., Bell, P. E. and Herlihy, A. T. (1989). Microbes, sediments, and acidified water: the importance of biological buffering. In: Rao, S. S. (Ed.). Acid Stress and Aquatic Microbial Interactions. CRC Press, Boca Raton, Florida, pp. 1-19. Mitchell, M. J., David, M. B. and Harrison, R. B. Chapter 9 in this volume.

Mitchell, M. J., David, M. B. and Uutala, A. J. (1985). Sulfur distribution in lake sediment profiles as an index of historical depositional patterns. Hydrobiol. , 121, 121-7.

Mitchell, M. J., Owen, J. S. and Schindler, S. C. (1990). Factors affecting sulfur incorporation into lake sediments: paleoecological implications. J. Paleolimnol., 4, 1-22

Mitchell, M. J. , Schindler, S. C. ,Owen, J. S. and Norton, S. A. (1988). Comparison of sulfur concentrations within lake sediment profiles. Hydrobiol. , 157, 219-29.

Morse, J. W. and Cornwell, J. C. (1987). Analysis and distribution of iron sulfide minerals in recent anoxic marine sediments. Mar. Chem., 22, 55-69.

Norton, S. A., Mitchell, M. J., Kahl, J. S. and Brewer, G. R. (1988). In-lake alkalinity generation by sulfate reduction: a paleolimnological assessment. Water Air Soil Pollut., 39, 33-45.

Nriagu, J.O., Holdway, D. A. and Coker, R. D. (1987). Biogenic sulfur and the acidity of rainfall in remote areas of Canada. Science, 237, 1189-92.

Nriagu, J. O. and Soon, Y. K. (1985). Distribution and isotopic composition of sulfur in lake sediments of northern Ontario. Geochim. Cosmochim. Acta, 49, 823-34.

Nurnberg, G. K. (1984). Iron and hydrogen sulfide interference in the analysis of soluble reactive phosphorus in anoxic waters. Water Res., 18, 369-77.

Oremland, R. S., Kiene, R. P., Mathrani, I., Whiticar, M. J. and Boone, D. R. (1989). Description of an estuarine methylotrophic methanogen which grows on dimethyl sulfide. Appl. Environ. Microbiol., 55, 994-1002.

Owen, J. S. and Mitchell, M. J. (1989). Sulfur biogeochemistry of an acidic lake in the Adirondack Region of New York. In: Rao, S. (Ed.). Acid Stress and Aquatic Microbial Interactions. CRC, Boca Raton, Florida, pp. 59-68.

Pankow, I. F. and Morgan, J. J. (1980). Dissolution of tetragonal ferrous sulfide (mackinawite) in anoxic aqueous systems. 2. Implications for the cycling of iron, sulfur, and trace metals. Environ. Sci. Technol., 14, 183-6.

Parkin, T. B. and Brock, T. D. (1981). The role of phototrophic bacteria in the sulfur cycle of a meromictic lake. Limnol. Oceanogr., 26, 880-90.

Perry, T. E., Baker, L. A. and Brezonik, P. L. (1985). Comparison of sulfate reduction rates in laboratory microcosms, field mesocosms, and in situ at Little Rock Lake, Wisconsin. Lake Reservoir Manage., 2, 309-12.

Pfennig, N. and Biebl, H. (1981). The dissimilatory sulfate-reducing bacteria. In: The Prokaryotes. Springer-Verlag, Berlin, pp. 941-7.

Pfenning, N., Widdel, F. and Truper, H. G. (1981). The dissimilatory sulfur-reducing bacteria. In: The Prokaryotes. Springer-Verlag, Berlin, pp. 726-940.
Ramm, A. E. and Bella, D. A. (1974). Sulfide production in anaerobic microcosms. Limnol. Oceanogr., 19, 110-18.

Richards, S. R., Kelly, C. A. and Rudd, J. W. M. (1991). Organic volatile sulfur in lakes of the Canadian Shield and its loss to the atmosphere. Limnol. Oceanogr. 36, 468-482.

Richards, S. R. , Rudd, J. W. M. and Kelly, C. A. (in review). Organic volatile sulfur in systems ranging in sulfate and dissolved salt concentrations over five orders of magnitude. Limnol. Oceanogr., in press.

Roy, A. B. and Trudinger, P. A. (1970). The Biochemistry of Inorganic Compounds of Sulfur. Cambridge University Press, 400pp .

Rudd, J. W. M. , Kelly, C. A. and Furutani, A. (1986). The role of sulfate reduction in the long term accumulation of organic and inorganic sulfur in lake sediments. Limnol. Oceanogr., 31, 1281-91.

Rudd, J. W. M., Kelly, C. A., St Louis, V., Hesslein, R. H. and Holoka, M. H. (1986). Microbial consumption of nitric and sulfuric acids in acidified north temperate lakes. Limnol. Oceanogr., 31, 1267-80.

Rudd, J. W. M., Kelly, C. A., Schindler, D. W. and Turner, M. A. (1988). Disruption of the nitrogen cycle in acidified lakes. Science, 240, 1515-17.

Rudd, J. W. M., Kelly, C. A., Schindler, D. W. and Turner, M. A. (1990). A comparison of the acidification efficiencies of nitric and sulfuric acids by two whole- lake addition experiments. Limnol. Oceanogr. 35, 663-679.

Schiff, S. and Anderson, R. F. (1987). Limnocorral studies of chemical and biological acid neutralization in two freshwater lakes. Can. J. Fish. Aquat. Sci., 44, (Suppl. 1), 173-87.

Schiff, J. A. and Hodson, R. C. (1973) .The metabolism of sulfate. Annu. Rev. Plant Physiol., 24, 381-414.

Schindler, D. W. (1985). The coupling of elemental cycles by organisms: evidence from whole-lake chemical perturbations. In: Stumm, W. (Ed.). Chemical Processes in Lakes. Wiley-Interscience, New York, pp. 255-50.

Schindler, D. W., Turner, M. A., Stainton, M. P. and Linsay, G. A. (1986). Natural sources of acid neutralizing capacity in low-alkalinity lakes of the Precambrian Shield. Science, 232, 844-7.

Shaffer, P. W., Hooper, R. P., Eshleman, R. N. and Church, M. R. (1988). Watershed vs in-lake alkalinity generation: a comparison of rates using input- output studies. Water Air Soil Pollut.,39, 263-73.

Smith, R. L. and Klug, M. J. (1982). Reduction of sulfur compounds in the sediments of a eutrophic lake basin. Appl. Environ. Microbiol., 41, 1230-7.

Sorokin, Y. I. (1970). Interrelations between sulphur and carbon turnover in meromictic lakes. Arch. Hydrobiol., 66, 391-446.

Sorokin, Y. I. (1975). Sulphide formation and chemical composition of bottom sediments of some Italian lakes. Hydrobiol., 47, 231-40.

Steudler, P. A. and Peterson, B. J. (1984). Contribution of gaseous sulphur from salt marshes to the global sulphur cycle. Nature, 311, 455-7.

Stuiver, M. (1967). The sulfur cycle in lake waters during thermal stratification. Geochim. Cosmochim. Acta, 31, 2151-67.

Sullivan, T. J., Eilers, J. M., Church, M. R., Blick, D. J., Eshleman, K. N., Landers, D. H. and DeHaan, M. S. (1988). Atmospheric wet sulphate deposition and lakewater chemistry. Nature, 331, 607-9.

Sweerts, J. P., Rudd, J. W. M. and Kelly, C. A. (1986). Metabolic activities in flocculent surface sediments and underlying sandy littoral sediments. Limnol. Oceanogr., 31, 330-8.

Tabatabai, M. A. and Bremner, J. M. (1970). An alkaline oxidation method for the determination of total sulfur in soils. Soil Sci. Soc. Am. Proc., 34, 62-5.

Turner, S. M., Malin, G., Liss, P. S., Harbour, D. S. and Holligan, P. M. (1988). The seasonal variation of dimethyl sulfide and dimethylsulfonium propionate concentrations in nearshore waters. Limnol. Oceanogr., 33, 364-75.

Urban, N. R. and Baker, L. A. (1989). Alkalinity generation associated with the formation of ester sulfates. Limnol. Oceanogr., 34, 1144-6.

Utainier, R. Y., Pfenning, N. and Truper, H. G. (1981). Introduction to the phototrophic prokaryotes. In: The Prokaryotes. Springer-Verlag, Berlin, pp. 197-211.

Vairavamurthy, A. and Mopper, K. (1987). Geochemical formation of organosulphur compounds (thiols) by addition of H₂S to sedimentary organic matter. Nature, 329, 623-5.

Vairavamurthy, A. and Mopper, K. (1989). Mechanistic studies of organosulfur (thiol) formation in coastal marine sediments. In: Saltzman, E. and Cooper, W. (Eds). Biogenic Sulfur in the Environment. American Chemical Society, Washington DC, pp. 231-42.

Weissner, W. (1981). The Family Beggiatoaceae. In: The Prokaryotes. Springer-Verlag, Berlin, pp. 380-9.,

Wetzel, R. G. (1975). Limnology. Saunders Co., Philadelphia, 743pp.

White, J. R., Gubala, C. P., Fry, B., Owen, J. and Mitchell, M. J. (1989). Sediment biogeochemistry of iron and sulfur in an acidic lake. Geochim. Cosmochim. Acta, 53, 2547-59.

Wright, R. F. (1983). Input-output budgets at Langtjern, a small acidified lake in southern Norway. Hydrobiol., 101, 1-12.

Yavitt, J. B., Lang, G. E. and Wieder, R. K. (1987). Control of carbon mineralization to CH₄ and CO₂ in anaerobic, Sphagnum-derived peat from Big Run Bog, West Virginia. Biogeochemistry, 4, 141-57.

Zeyer, J., Eicher, P., Wakeham, S. G. and Schwarzenbach, R. P. (1987). Oxidation of dimethyl sulfide to dimethyl sulfoxide by phototrophic purple bacteria. Appl. Environ. Microbiol , 53, 2026-32.

Zinder, S. H. and Brock, T. D. (1978). Methane, carbon dioxide, and hydrogen sulfide production from the terminal methiol group of methionine by anaerobic lake sediments. Appl. Environ. Microbiol., 35, 344-52.

Zinder, S. H., Deomel, W. N. and Brock, T. D. (1977). Production of volatile sulfur compounds during the decomposition of algal mats. Appl. Environ. Microbiol., 34, 859-60.