| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Plant and Soil Science, University of Vermont, Burlington, VT 05405
* Corresponding author (dross{at}uvm.edu)
Received for publication November 28, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The potential mechanisms of nitrate consumption in peatlands include denitrification, assimilation by plants and microbes, and dissimilatory reduction to ammonium (Urban et al., 1988). Although there is considerable uncertainty surrounding the relative importance of the various nitrate removal mechanisms, most researchers agree that the major processes are assimilation by plants and denitrification. In peatland ecosystems, soil anoxia and the availability of labile compounds favors the denitrification process, and, consequently, can result in quantitatively important losses of nitrogen from these systems (Morris, 1991; Kent, 1994). In addition, researchers have frequently identified denitrification as the most important mechanism of NO3– removal in wetlands under evaluation for water quality amelioration (Jansson et al., 1994), and in riparian zones (Hill, 1996). Often, the rapid loss of nitrogen from wetlands is considered to be a beneficial function, as the process can protect wetlands from the eutrophying effects of too much nitrogen. However, increased denitrification rates will also increase outputs of nitrogenous gases, which may have potentially harmful effects on the atmosphere. The extent to which denitrifiers are able to process additional nitrogen inputs may be the determining factor in the long-term effects of added nitrogen on threatened peatland ecosystems.
Nitrification may also be an important process in wetland soils as it links N mineralization and atmospheric NH4+ inputs to potential N loss from the system through denitrification. While nitrification and denitrification processes have traditionally been regarded as separate, occurring in different layers of water, soils, and sediments, recent research has revealed the situation may not be so clear-cut. For example, studies have shown that nitrite-producing autotrophs have mechanisms for coping with oxygen limitation, while a number of heterotrophic nitrifiers have been found to be able to denitrify aerobically (Kuenen and Robertson, 1994). As a result, determination of nitrification rates in soil may help to evaluate the potential of that soil for loss of nitrogen through denitrification.
Few studies have attempted to quantify denitrification rates in natural ecosystems, partly because the low rate of gas flux from these systems makes it difficult to measure. Most studies on denitrification in peatlands have consisted of short-term measurements in the laboratory with small quantities of moss (Sphagnum spp.) or subsurface peat (e.g., Hemond, 1983; Hoffman et al., 2000; Klemedtsson et al., 1977; Muller et al., 1980), although a few studies have attempted to quantify rates in the field using acetylene inhibition methods (e.g., Urban et al., 1988). The main objective of this study was to utilize the acetylene inhibition technique in an effort to determine the contribution of denitrification to nitrogen loss from a natural peatland. In addition, the effects of moisture content, nitrate additions, and the degree of ombrotrophy–minerotrophy on denitrification rates were examined. Rates of net nitrification and the potential for net nitrification were also examined in an attempt to determine the importance of nitrification as a source of nitrate to denitrifiers in this peatland.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Molly Pond is bordered by a floating shrubland, which is dominated by leatherleaf [Chamaedaphne calyculata (L.) Moench], large cranberry (Vaccinium macrocarpon Aiton), and peat mosses (Sphagnum spp.). Other plant species found on the floating peat mat include bog cranberry (Vaccinium oxycoccos L.), pale laurel (Kalmia polifolia Wangenh.), bog rosemary [Andromeda glaucophylla (Link) DC.], pitcher plant (Sarracenia purpurea L.), sundew (Drosera intermedia Hayne), beak rush [Rhynchospora alba (L.) Vahl], and cotton grass (Eriophorum spp.) (Sanderson et al., 1993). The densest populations of pitcher plants are located in the northeast portion of the floating mat while other areas have limited or scarce densities of pitcher plants (Mouser, 2003). Where Molly Bog's open mat grades into forest, black spruce [Picea mariana (P. Mill.) B.S.P.] and tamarack [Larix laricina (Du Roi) K. Koch] can be found (Sanderson et al., 1993).
A hydrologic study of Molly Bog by Mouser (2003) determined that the bog is functioning close to a true ombrotrophic bog, with the majority of inputs from precipitation and losses through evapotranspiration. However, results of this study also showed that the local peat and regional water table are connected across the western side of the bog, but separated by a confining layer on the eastern side. The flow reversal that occurs in the western region of the bog during the wet season produces higher concentrations of chemical constituents in isolated areas of Molly Bog. This flow reversal has a large effect on the surface vegetation, leading to a lower density of pitcher plants in this region, which can be considered an indicator of the bog's varying trophic status (Mouser, 2003).
Sampling Strategy
For the present study, transects were established in locations based on results of the study by Mouser (2003). Transect I was established in the northeast portion of the bog (Fig. 1)
, the area with the highest density of pitcher plants and most ombrotrophic characteristics. Transect II was established in the northwestern portion of the bog, where pitcher plant densities are much lower, and some local peat and regional water table exchange has been shown to occur (Mouser, 2003). Transects ran from the center toward the perimeter of the bog. This design allowed for sampling of sites with varying moisture content, as moisture decreased with increasing distance from the pond. Transect I consisted of three sites spaced at 8-m intervals, and Transect II consisted of two sites, also spaced at 8-m intervals (in both cases, Site 1 was located closest to the pond). Water table depths during sampling periods were approximately 0, 7, and 9 cm for Transect I, Sites 1, 2, and 3, respectively, and 4 and 13 cm for Transect II, Sites 1 and 2, respectively. Establishment of a third site on Transect II did not occur because of a deeper water table at this location. Sampling occurred on 29 Apr.–1 May 2003 (April sampling), and on 25–26 May 2003 (May sampling). Molly Bog is a University of Vermont Natural Area, and, consequently, limitations were placed on the number of replicate samples that could be collected.
|
Denitrification Rate Measurements
Denitrification rates were measured using the acetylene inhibition technique in both field and laboratory studies (Yoshinari et al., 1977). The field, or in situ, method consisted of treatment of peat with acetylene using a static enclosure placed over the peat surface followed by determination of N2O emission. The laboratory, or ex situ, method consisted of incubation of intact peat cores with acetylene followed by analysis for N2O. The limitations of using the acetylene inhibition technique for measurement of denitrification (Aulakh et al., 1992; Hauck, 1986; Knowles, 1990; Mosier et al., 1990; Tiedje et al., 1989) were realized and addressed by employing the following methods: (i) using short incubation periods while also allowing sufficient time for gases to diffuse, (ii) using both an in situ and ex situ method of measurement, (iii) using high grade acetylene that was free of impurities, (iv) inserting chambers and removing cores in such a way as to minimize soil disruption, (v) monitoring temperature within field chambers to ensure it remained consistent with outside air temperature, (vi) moving N2O flux measurement chambers and core sampling sites to different locations within a field site on each sampling date, and (vii) adding acetylene saturated water to peat to enhance acetylene diffusion, as well as pumping with a syringe before gas sampling to enhance gas diffusion. The acetylene inhibition technique, using either the chamber or the core method, has proven to be a successful technique both because contamination by atmospheric N2 does not affect the analysis, and because a high sensitivity for N2O analysis is readily attained (Klemedtsson et al., 1977).
In Situ Method
In situ denitrification rates were measured on 29 Apr. and 1 May 2003 and on 25 May 2003. Measurement of rates was performed using 4-L chambers consisting of Nalgene low-density polyethylene jugs (Nalge Nunc International, Rochester, NY) with the tops removed (Freeman et al., 1997). A 1.03-cm (13/32-inch) hole was drilled in the base of each chamber and a septum was inserted to permit gas sampling and introduction. Two replicate chambers were installed at each of the five sites on each sampling date.
Before placement of the chamber over the peat surface, 60 mL of acetylene-saturated water was poured onto the sampling site. The objective of adding acetylene-saturated water was to enhance diffusion of acetylene to sites, as acetylene diffusion is limited in saturated soils (Hallmark and Terry, 1985; Terry et al., 1986). Acetylene was also introduced into the chamber headspace by placing a 250-mL beaker with approximately 200 mL of distilled-deionized water on the sampling site and adding approximately twenty-five 0.317- to 0.635-cm (1/8- to 1/4-inch) granules of calcium carbide to the water immediately before chamber installment. This quantity of calcium carbide was found to produce acetylene in amounts equal to approximately 10% of the headspace volume of the chamber. The chamber was then inverted and pushed into the peat, with the aide of a knife to prevent compaction, until the base of the chamber was below the water table. This method was designed to form an air-tight seal around the base of the chamber. The chamber was vented with a needle 20 min after it was installed to equalize the additional air pressure within the chamber caused by acetylene production. Gases within the chamber were sampled immediately on insertion of the chamber and again after 4 or 7 h. Gas diffusion was enhanced before each sampling by pumping with a 50-mL gas-tight syringe. Nine-milliliter gas samples were taken and placed in evacuated 9-mL glass vials.
Ex Situ Method
Ex situ denitrification measurements were made in the laboratory using the static core technique similar to that described by Tiedje et al. (1989). Five to seven cores were sampled from each site on the dates when the field chamber technique was used. Cores were collected by cutting peat with a knife around an 8-cm-diameter by 10-cm-high aluminum core that was pushed gently into the peat to prevent compaction. Cores were removed and sealed on the bottom with plastic wrap and elastic bands. If water was lost through the bottom of the core, it was replaced to its original level as estimated from the water level in the hole from which the core was removed, using solution from the hole. Cores were then placed in sealed plastic bags and transported on ice to the lab where they were stored at 5°C overnight before incubation the following day.
Incubation containers consisted of 1-L Nalgene polycarbonate straight-side jars (Nalge Nunc International) with polypropylene lids modified by drilling a hole in the lid and inserting a septum to permit gas sampling. Treatments consisted of two control cores from each site (C, control treatment), three cores with added acetylene from each site (A, acetylene treatment), and two cores with added nitrate and acetylene from four of the five sites (N, nitrate treatment). Cores with added nitrate received 10 mL of 0.1 M KNO3 (1 mmol N) distributed evenly throughout the core using a syringe with a long needle. Nine-milliliter gas samples were withdrawn from all cores, immediately following sealing of containers and before introduction of acetylene, and placed in 9-mL evacuated glass vials. Acetylene was generated by reacting calcium carbide with distilled-deionized water and was added to the headspace of designated cores after incubation containers were sealed. Fifty milliliters of acetylene (10% of headspace volume) was added to each container following removal of 50 mL of air. Gases were sampled again after 8 h of incubation in the dark at 10°C.
Gas Analysis
Samples were stored at room temperature and were mailed to the Institute of Ecosystem Studies in Millbrook, NY, for analysis by a Tracor 540 TCD/ECD tandem gas chromatograph (Thermo Electron Corp., Waltham, MA), with a Tekmar 7000 equilibrium headspace autosampler (Teledyne Technologies, Mason, OH). The column (stainless steel, 6 m long, 0.085-cm inner diameter, packed with Porapak Q), detector, and injector were run at 55°C. Nitrogen was used as a carrier gas, with a flow rate of 26 mL min–1. Gases were separated on the column and first passed through the nondestructive TCD and then through the 63Ni ECD. Standards were stored and analyzed to determine the effect of sample storage and transport on N2O concentrations. Results showed that vial leakage during storage and transport was insignificant.
Nitrification Rate Measurements
Initial and final pore water and peat nitrate and ammonium concentrations were determined for cores incubated in the denitrification study. Peat nitrate and ammonium concentrations were determined by KCl extraction. Initial peat concentrations were determined by sampling three replicates of 125-mL volumes of peat in the field on the dates cores were collected. The peat was sampled to a depth of 10 cm and placed in a 50-mL container with 250 mL KCl, shaken intermittently, and stored on ice until arrival at the lab for processing.
Final KCl extractions (60 mL peat to 120 mL KCl) were made on cores incubated for denitrification measurements. Separate extractions were done on saturated and unsaturated peat within each core that had an adequate peat volume in each zone. After shaking for at least 15 min, all extractions were centrifuged at 17400 x g for 10 min and filtered through GF/F 0.7-µm borosilicate glass filters (Whatman, Maidstone, UK) into 20-mL plastic vials, and frozen until analysis. Replicate samples of equal volume were weighed and dried for water content and dry weight determination with each initial and final KCl extraction.
Initial pore water nitrate and ammonium were determined from three replicate samples taken from each site on dates when initial KCl extractions were done. Pore water was collected and filtered through Whatman GF/F 0.7-µm borosilicate glass filters and frozen until analysis. Final pore water concentrations were determined from pore water drained from each core incubated in the denitrification study. Samples were filtered and frozen until analysis. All samples were analyzed for nitrate and ammonium on a Lachat QuickChem AE flow injection analyzer (Hach, Loveland, CO).
Two week-long aerobic peat incubations were also performed from 15–22 May 2003 and 25 May–1 June 2003. For each incubation, two replicate grab samples of peat were collected to a depth of 10 cm at four of the five sites. The peat was allowed to drain and excess water was consistently squeezed from samples. Samples were then homogenized and stored at 10°C for 1 wk in plastic bags that remained open to the environment. Homogenization was achieved by consistently breaking apart and mixing samples by hand. The KCl extractions were performed on Days 0, 1, 2, and 7. Lab extractions involved shaking 40 mL of peat with 80 mL of KCl, with subsequent centrifugation, filtering, and freezing, as described above. At the end of the incubation, two replicate subsamples were dried to determine dry weight and moisture content of each sample.
Calculations and Data Analysis
Denitrification rates were estimated from the change in N2O concentration over the incubation period. Calculations for cores were made using an equation developed by Mosier and Klemedtsson (1994), which involves the following parameters: (i) N2O evolved during the incubation period; (ii) the volume of the liquid phase in the core; (iii) the total gas volume within the incubation container; (iv) the temperature within the container; and (v) the Bunsen absorption coefficient that accounts for N2O dissolved in the water phase. The volume of the liquid phase in the core was calculated from an initial and final dry weight of the core. The total gas volume within the container was calculated using the measured bulk density of the core (dry weight/core volume), an average for published particle density values of Sphagnum peat (Heiskanen, 1992), and the volume of the core and incubation container. Calculations for chambers were made using an equation modified from Matthias et al. (1980), based on N2O evolved during the incubation period, the volume of headspace within the chamber, and the area of peat covered by the chamber.
Nitrate and ammonium concentrations in precipitation samples were used to calculate mean values of nitrate and ammonium in precipitation. Those mean values were then used to calculate an estimate of annual deposition of nitrate and ammonium per hectare at the study site. Calculated values for annual deposition compared well with data from the National Atmospheric Deposition Program's monitoring site in Underhill, VT, which is approximately 16 km (10 miles) from the study site.
The general linear model procedure in SAS (SAS Institute, 1996) was used for analysis of variance and covariance. Regression and correlation analyses, using Pearson's correlation coefficient, were also performed using SAS. Data were checked for normality using the Shapiro–Wilk test, and for homogeneity of variances using Levene's test. In some cases, data were log-transformed to satisfy normality. Pairwise comparisons were made using Fisher's least significant difference test. Differences between treatments and sites were analyzed with and without the effect of site or treatment used as a blocking factor in the model. The Kruskal–Wallis nonparametric test was performed in some cases where the assumptions of normality and homogeneity of variances were not met. Denitrification rates as well as changes in nitrate and ammonium concentrations were regressed with percent water by volume in each core to determine if significant relationships existed.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
|
While there appeared to be higher rates of N2O evolution from acetylene-amended versus unamended cores in Transect I during the April sampling (Table 1), the differences were not significant, and this trend did not continue through other sampling dates and transects. The lack of consistent differences observed between N2O evolution rates in unamended versus acetylene-amended cores appears to be due to similarity rather than high variability in the data, as nonparametric tests did not yield significant differences and mean rates do not appear to exhibit any consistent trends. It is believed that soil factors largely determine the ratio of N2 to N2O evolved during denitrification. Low pH (less than 5) leads to greater production of N2O versus N2 (Knowles, 1982; Morris, 1991; Parkin et al., 1985b), suggesting that N2O may be a major product of this process in acidic wetland soils (Hemond, 1983; Knowles, 1982; Parkin et al., 1985b).
Another potential explanation for the lack of significant differences between acetylene-amended and unamended cores may be that acetylene was not sufficiently blocking N2O reduction, which could be attributed to a lack of acetylene diffusion to sites of denitrification due to the high water content of cores. However, steps were taken to help overcome this potential problem, with the introduction of high concentrations of acetylene to the container, and the attempt to enhance diffusion by pumping with a 50-mL syringe. In addition, if inadequate diffusion were in fact the problem in these cores, we might have expected to see greater differences in rates between acetylene-amended and unamended cores in the drier versus the wetter sites, as Sites 2 and 3 were significantly drier than Site 1. However, this result was not observed.
It is also possible that acetylene was blocking autotrophic nitrification, leading to lower concentrations of nitrate available to denitrifiers in acetylene-amended cores, and subsequently, lower rates of denitrification. In addition, N2O production might be higher in unamended cores due to N2O produced as a by-product of autotrophic nitrification, a process which should be inhibited in the acetylene-amended cores. However, results from the analysis of net nitrification in this study did not support this hypothesis, as there were no differences found in net rates of nitrification between acetylene amended and unamended cores.
Effect of Transect Location and Sampling Date on Denitrification Rates from Cores
It was hypothesized that denitrification rates would be higher in the more minerotrophic transect, as pH was expected to be higher in this region, and the minerotrophic zone may also provide additional nutrients to microbes in this area due to the connection of the local peat and regional water table (Mouser, 2003). Surface pore water pH values were found to be significantly higher in the minerotrophic transect during the study period (Table 3), although the only instance where N2O evolution was found to be higher in Transect II was in the case of the control cores during the May sampling. Therefore, while there are some ground water inputs to the region where Transect II was located, these inputs did not significantly affect rates of denitrification in this area. These ground water inputs did lead to higher pH values in this region and differences in plant species found here (lower densities of pitcher plants), but did not lead to significant differences in nitrate concentrations in peat or pore water in this region, however, as all measured nitrate concentrations were below the detection limit.
There did not appear to be any consistent differences in N2O evolution rates between sampling dates. This is not surprising given that only two dates were sampled, and both were during the spring season. Most studies have found high rates of denitrification in spring and fall relative to summer or winter (Groffman, 1987; Hanson et al., 1994; Henrich and Haselwandter, 1997; Struwe and Kjoller, 1991; Tiedje et al., 1989). This has been attributed to a lack of competition for nitrate between denitrifiers and plants at these times, as well as the increased availability of carbon due to freezing and thawing events during these times (Groffman and Tiedje, 1989; Zak and Grigal, 1991). As a result, rates measured in this study may be representative of high-end fluxes from this ecosystem during a typical year.
Denitrification Rates from Field Chambers
The acetylene inhibition field chamber technique employed in this study to measure denitrification yielded very low rates of denitrification from the peatland, and these rates were not significantly different across sites. Chambers were deployed for 4 h during the April sampling, and rates of N2O evolution during this time were negative, which may be representative of a lack of N2O production, rather than N2O consumption, due to error around zero. The incubation period was increased to 7 h during the May sampling and positive results were obtained. As a result, data from the May sampling only are presented (Table 4), as the incubation time during the April sampling appeared to be insufficient. Rates measured from the chambers were much lower than rates measured from cores, and there are two possible explanations for the significant differences found between core and chamber rates. One is that gas diffusion between the peat and the chamber was hindered by the high water content of the peat, although we did attempt to minimize this problem by employing a longer incubation time during the May sampling and by adding acetylene saturated water to the study sites. The higher rates of N2O evolution measured during the May sampling, although significantly different only in Transect II, indicate that gas diffusion may well have been the problem as the longer incubation time seemed to lead to larger fluxes of N2O from the peat. Low rates of gas diffusion would have resulted in problems with acetylene diffusion into denitrification sites, as well as N2O diffusion out of denitrification sites. In addition, there may have been losses of N2O by lateral or downward diffusion beneath the chambers.
|
Nitrification
Rates of Net Nitrification
Rates of net autotrophic and heterotrophic nitrification could not be accurately determined from changes in peat or pore water nitrate concentrations because all initial nitrate concentrations were below the detection limit of the analytical instrument 
. Final concentrations of nitrate in peat and pore water (Table 5) were higher than initial concentrations in almost all cases, suggesting that nitrification was occurring to some extent. There was some variability in net nitrification rates, with most CV values below 100, although not nearly as much variability as occurred with denitrification rates. The very low nitrate concentrations found in peat and pore water are consistent with a study at Thoreau's bog in Massachusetts, which concluded that most of the nitrogen associated with or available to organisms in the bog was reduced nitrogen, as nitrate was undetectable much of the time in bog water (Hemond, 1983).
|
|
|
Potential Net Nitrification
Peat samples were drained and incubated aerobically on two separate dates to determine the potential for net nitrification in this peat soil. The average volumetric moisture content of samples during the two incubations was 22.0 and 16.6. All nitrate concentrations in both incubations were below the detection limit, and there were no significant changes in NH4+ concentrations over the course of either incubation (Table 8). These findings indicate that rates of net nitrification were negligible in this peat soil, even at a significantly reduced water content than is naturally present. This may have been due to an absence of nitrifier populations, or unsuitable soil conditions, such as low pH.
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Results suggest that increased chemical constituents (e.g., Ca2+) in the western region of this peatland, while influencing plant species composition (Mouser, 2003) and pH in that region, do not appear to be affecting rates of net nitrification or denitrification. This is not surprising given that the nitrate and ammonium peat and pore water concentrations were not higher in the western region of the bog. However, if the exchange between the local peat and regional water table became more substantial, or the regional water table nitrogen concentrations increased, it is possible that rates of denitrification and/or nitrification may increase in this region.
Rates of net nitrification in this peatland were extremely low, suggesting that nitrification may not be a significant source of nitrate to denitrifiers, based on our measured rates of denitrification. A more detailed study of gross rates of nitrification would be needed to further investigate this hypothesis. Results showed that there was very little free nitrate available to denitrifiers in peat extracts or in pore water. This, in combination with the low rates of net nitrification and the low rates of in situ denitrification measured in this study, suggests that inorganic N turnover in this wetland is low, and organic N may be the dominant form of soluble N in this peatland.
Moisture content of peat does not appear to be controlling rates of either net nitrification or denitrification in this peatland. It is likely that either substrate availability, pH, or oxygen status are the main limiting factors for these processes. However, on additions of nitrate to peat, there was a trend toward higher rates of denitrification in drier sites, a result that may be attributed to lower denitrification capacity in wetter sites due to a lower availability of nitrate in these areas and fewer denitrifying organisms. Although net nitrification rates were not significantly higher in drier sites, and all measured rates were very low, more nitrate may have been produced and rapidly cycled in drier sites.
In addition to denitrification and plant uptake, dissimilatory nitrate reduction to ammonium appeared to be functioning as a sink for nitrate. However, results did not show that this process was a significant competitor with denitrification for available nitrate in this peatland. Results support the hypothesis that nitrate is a limiting factor for denitrification in this peatland. Therefore, it is expected that increases in atmospheric nutrient inputs to this system will result in higher rates of denitrification, and, as a result, that the peatland has some potential to buffer the effects of increased nitrogen loadings. Although wetter sites may currently have a lower capacity for denitrification, increases in nitrogen inputs to these sites may stimulate a change in the microbial community and result in greater potential for denitrification. While ombrotrophic peatlands do not lose nitrogen through drainage into ground water, they do appear to provide conditions conducive to significant nitrogen loss through denitrification. As a result, the processes of denitrification and nitrogen assimilation by plants may serve to protect these threatened ecosystems from significant alterations as a result of increased atmospheric nitrogen deposition.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Related articles in JEQ:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
