<<

Rioi. Biockm. Vol. 23, No. 9, pp. 845-W. 1991 0038-0717/9153.00 +o.oo Printed in C&at Britain. All rightsmerved Copyright Q 1991 Pcrgamoa Pta pk

DENITRIFICATION: ENZYME CONTENT AND ACTIVITY IN DESERT

WILLIAM T. FWRRJoHN+ f.kPartment of Botany, Duke University, Durham. NC 27706, U.S.A.

(Accepted 8 March 1991)

!knnary-The content of denitrifying enzymes in upland desert soil was strongly associated with indices of N and C availability. Combinations of several predictors could explain 7 I % of the variance in cnxyrnc content in Chihuahuan desert soils and 87% of the variance in soils from various deserts in the southwestern U.S.A. A significant fraction of the enzyme content in wet desert soil is derived from a persistent pool of enzymes capable of tolerating extended periods of desiccation. The synthesis of new denitrifying enzymes appears to involve a complex interaction between available C. N. and soil moisture. The activity ofdenitrifying enzymes in desert soil was optimal at a pH of 7.0 and a temperature of 4OC. The Q,9 for denitrilication was 1.74. and the activation energy was about 41 kJ mot-‘. In addition, enzyme activity in freshly wet soil was not limited by NO; availability. and only slightly limited by the availability of C. Thus, wet desert soil appears to provide optimal conditions for several variables that can influence

INTRODUCTtON and Schlcsingcr. 1990). From a global perspective, emissions of N gas from arid land may Dcnitrification is the microbial reduction of NO; or amount to 30% of the total loss from terrestrial NO; to either N,O or Nr. This reaction allows a wide variety of to use NO, or NO, as the ccosystcms with most of this loss attributed to deni- terminal respiratory electron acceptor under low Or trification (Bowden, 1986). Few studies have investigated dcnitrification in conditions. The bacteria capable of denitrification are generally with a widespread distribution desert ecosystemsdespite its occurrence under appar- (Ticdje et al.. 1982; Tiedje. 1988). Many factors are ently unsuitable conditions, and its possible signifi- known to affect the rate of denitrification. These cance to desert fertility. Therefore, the purpose of this study was to investigate the factors affecting denitrifl- include pH. temperature. C and N availability, and the partial pressure of 0, (Firestone, 1982; cation in desert soils. The effect of several variables Knowles, 1982). In general, however, high rates of on both the production and activity of denitrifying denitrification are associated with wet, nutrient-rich enzymes will be considered. environments. Desert ecosystems would seem unsuitable for the MATERIALS AND METHODS process of denitrification. Most desert soils are hot, dry and nutrient-poor. Surface soil temperatures can Study sires exceed 60°C in the summer and drop below freezing To encompass the natural variability present in in the winter. Mean annual rainfall is generally desert landscapes, I sampled 3 bajadas derived from < 25 cm, and some regions can go for more than I2 different parent materials, in each of 3 deserts: the months without precipitation. Yet, the rate of deni- Chihuahuan; Mojave; and Great Basin. Soil samples trification in wet desert soil is comparable to, if not were also collected from nearby playas. The location greater than, that measured in more mesic ecosystems and features of each site are described in Table I. (Virginia ef al., 1982; W. T. Peterjohn. unpublished Ph.D. thesis, Duke University, 1990). Field methods In ecosystems where few nutrients are lost in I collected samples of surface soil (4-8 cm depth) stream flow, the production of gases by along transects oriented perpendicular to the length denitrification assumesgreater importance as a vector of each bajada (Peterjohn. lot. cit.). Between the of nutrient loss. A nitrogen budget for the Great transects additional samples were collected for a total Basin desert estimated that I9 kg N ha-’ yr-’ were of 107 samples from each bajada. Soils were also lost by denitrification (West and Skujins, 1977). This collected from 8 playas. All soils were collected amount represented 95% of annual losses and 65% during the dry season, sieved ( < 2 mm), and stored of total N inputs. On a regional scale, it appears that in water-tight containers. Equal mass subsamples of >77% of the N inputs to the deserts of the south- the soil from each transect were mixed to form 5 western U.S.A. are lost to the atmosphere (Peterjohn composite samples for each bajada, and I for each playa. An additional composite sample was created *Presentaddress: The Ecosystems Center, Marine Biologi- for each of the bajadas in the Chihuahuan desert cal Laboratory, Woods Hole, MA 02543, U.S.A. using qual mass from all I07 soil samples.

845 846 WI- T. Parartrorm

Table I. Location. geology. land use. and vegetation for each study site. Only the three most dominant woody pcranniafsare listed. tmporfanf values (i.v.) are based on the relative density and dominance of ench species (Mu&r-Dombob and Efienberg. 1974)

Vegetation Geology Dominant woody i.v. SWfC Site Location Deserf Mountain range Bedrock Landuse perennials (‘4)

KM. I 4OkmNNE Chihuahuan Dona Ana Rhyolite/Andasite Mixed Lorrea tridentata 47 Las truces FIorrntia cermto 21 Guteri;io sorothrae N.M. 2 10.4 km E ~ihuah~n Franklin Mixed sedimentary Grazed L. tridmtata ::: Anthony (mostly limcstonef Pnrrhcniwn infMunr 32 G. sarothroe 16 NM 3 4OkmNNE Chihuahuan Dona Ana Monxonite Ungrazed L. t&%vuoto 40 Las Cruccs Mt Summer-ford G. wathrae 37 Zinia awro~a Calif. I IJkm NNE Mojave Coxcomb Mctascdimcntary Ungramd Encelia farinosa 3: Desert Center Ambrosia domosa 17 L. tridentata I2 Calif. 2 22.5 km W Mojave Eagte Gneiss Ungraxed L. tridemato 40 Desert Center A. dwnom 13 Chrisothammu noureow 9 Calif. 3 9.6 km W Mojave Eagle Granite Ungram A. dnmom 42 Desert Center L. tridentata 27 HymenocIea saisoIo 14 NW. I 75.5 km N Greal Basin Sclenite Quartz monzonitel Grazed Atriples confirti/oIio Wadsworth Granodioritc Artemida spinescenr :I: Suaeda torreyana NW. 2 50.5 km ENE Grcaf Barin Stillwater Mixed scdimenfary Grazed Atriplex conferti/ulio 5: FJllOll Suoedo torreyona 30 SureohrrturIrrmictdtttur I6 NW. 3 3Q.Rkm NNW Great Basin Lake Basalt/Andesifc Mixed Sarcobatus boileyi Wadsworth A~femiiia .rp’“c~cenr :: Surc&rrnr ~erm;culatu.~ II

Laborutory methods sampfcs were removed at four, I h intervals. Gas Generut procedure. To investigate factors affecting samples along with certified standards were stored in denitrifying cnzymc production and dcnitrifying en- evacuutcd vials for later analysis of N*O using a gas zymc activity (DEA), a 2-step process was used. First, chromatograph equipped with a 61Ni electron capture dry soils were incubated under controlled conditions detector (Petcrjohn. lot. cir.). At the end of the assay to allow bacterial growth and the synthesis of dcnitri- procedure the air volume in each bottle was detcr- fying enzymes. Second, the enzyme content of incu- mined using a pressure transducer (Parkin er al., bated soils was fixed by adding ~hIoramphenico1 and 1984). All DEA values were corrected for N,D measured by an in viuo DEA assay (Smith and Tiedje, dissolved in the assay solution (Wilhem er d, 1977; 1979; Murray et al., 1989). By altering the conditions Tiedje, 1982). the dilution due to the overpressure, of the initial incubation, diffcrcnccs in the amount of and the amount of N*O removed by each sampling cnzymc produced could be measured. By altering the event. Since the DEA was determined under non- conditions of the DEA assay itself, differences in limiting conditions and in the presence of chforam- enzyme activity could be measured. phcnicol, differences in activities represent differences incuberion procedure. The standard incubation in the content of denitrifying enzymes present in the procedure was to place IO g of dry soil into a 120 ml soil immediately prior to the addition of the assay serum bottle, with enough distilled water to achieve solution (Smith and Tiedje, 1979; Tiedje, 1982; Tiedje a matrix potential of -0.05 MPa. The bottle was et al., 1989). capped and incubated aerobically for 3days at 30°C. A matrix potential of -0.05 MPa was chosen Enzyme content and production studies because it is typical of surface soils during wet periods Predictors of enzyme content. The standard soil in the Sonoran and Chihuahuan deserts (Young and incubation and DEA assay procedures (described Noble, 1986; Schlesinger er at., 1987). An equation above) were used to determine the denitrifying en- generated from 53 composite soil samples was used to zyme content. Soil variables potentially associated calculate the amount of water needed to achieve with denitrifying enzyme content were also measured -0.05 MPa in each soif (Peterjohn, for. cit.). in each soil (Peterjohn, toe. cit.). The variables in- Assay procedure. The standard procedure was to cluded: % sand, silt, and clay; pH; % total N, % total add 10 ml of an assay solution to the incubated soil, C, % carbonate C, % organic C, -N concen- make the headspace anaerobic by alternately evacuat- tration; -N concentration; Fday nitrifica- ing and flushing it with Ar, and then overpressure tion potential; wet soil CO* production; and the C: N the sealed bottles with 20 ml of CaC,-generated C2H, ratio. Bivariate plots between the denitrification en- and 60 ml of Ar. The assay solution contained zyme content and the other soil variables were used 10Om~ NaH,PO, buffer (pH = 7.00). 35.2m~ to identify outliers. and any linear or non-linear KNO,, 6.94 mM dextrose, and I .5 g chforamphenicol relationships. Simpte linear regression and stepwise I-‘. Preliminary work demonstrated that these con- multiple linear regression procedures were then used centrations of N and C did not limit enzyme activity. to establish the strength of these relationships. This Serum bottles were kept at 30°C. and 20mf gas analysis used all soil samples from the Chihuahuan De&&cation in desertsoils 847 desert, and the composite samples for each transect so each will be described separately. All experiments in the Mojavc and Great Basin deserts. used composite soils (107~sample composite) from Effect of C and N on enzyme production. Composite each bajada in the Chihuahuan Desert. samples from the 3 bajadas in New Mexico were used Effect of pH on DEA. To investigate the effect of in a factorial experiment to investigate the effect of pH on DEA, 5 g subsamples of each composite soil water, C, and N availability on denitrification enzyme were incubated anaerobically at -0.05 MPa for 3 production. The factorial experiment consisted of days at 3O’C. Following incubation. subsamples from assigning three subsamples from each soil to l-5 each composite soil were randomly assigned to I of treatments: (I) 1 ml of distilled water resulting in a I1 assay solutions, and DEA measured. Except for water potential of about -0.05 MPa; (2) I ml of a the pH buffer, the assay solution was identical to the solution containing 134.8 mg N I-’ as KNO,; (3) 1 ml standard solution described previously. Different pH of a solution containing 2940 mg C I-’ as dextrose; levels were attained by titrating a modified universal (4) I ml of a solution containing both C and N in buffer (MUB; Skujins et al., 1963; Tabatabai. 1982) the same amounts used in treatments 2 and 3 to one of the following pH values: 3, 4, 5, 6, 6.5, 7, (C:N = 21.8); and (5) a control with no additions to 7.5, 8. 8.5, 9 and IO. Since MUB is a relatively weak the dry soil. After mixing in treatment solutions, the buffer, 20 ml of the assay solution were added to the soils were incubated aerobically for 24 h at room soil and the final pH of the slurry was measured. Tbe temperature (23C). By altering the conditions of the final pH of the soil slurry was often higher than the initial incubation, the effect of these variables on pH of the assay solution due to the abundance of enzyme production could be studied. Following the CaCO, in desert soil. Three replicate samples for pH incubation period, DEA was determined using the levels 6. 7 and 8 were analyzed for each composite standard assay procedure as previously described. soil. One replicate was analyzed for the remaining pH A second factorial experiment was performed using levels. the composite sample from a single bajada in Efinct of temperature on DEA. To investigate the New Mexico (NM 3; Table I). The purpose of this effect of temperature on DEA, 5 g subsamplesof each cxpcriment was to dctcrmine whcthcr the manner of composite soil were incubated anaerobically at the treatment application could effect cnzymc pro- -0.05 MPa for 3 days at 30°C. Following incu- duction. This cxpcrimcnt was identical to the first bation, subsamplcs from each composite soil were with 2 cxccptions: soils wcrc incubated in pcrforatcd randomly assigned to I of 9 assay temperatures. and IOml plastic syringes rather than in l2Oml strum DEA was mcasurcd. The assay solution contained bottles; and 5 ml of the trcatmcnt solutions wcrc MUB (pH = 6.00). but was otherwise identical to the added to the top of each soil column rather than standard assay solution. Adjusting the buffer to a pH mixing 1 ml of the solutions into the soil. Thcsc of 6.00 ensured that the final soil slurry (20ml conditions were chosen bccausc they approximate solution: 5 g soil) had a pH bctwecn 7.06 and 7.27. those of a field expcrimcnt conducted on intact soil The assay tcmpcraturcs wcrc in the range IO-85C cores taken from the same site (Pctcrjohn. /oc. cit.). (SW Fig. 8). Three rcplicatc samples for the 20.40 and After the incubation, soils were transfcrrcd to strum 60°C treatments wcrc analyzed for each composite bottles and the DEA assayed in the standard manner. soil. One rcplicatc was analyzed for the remaining En:yme persistence in dry soil. A final cxpcrimcnt tcmpcrdturcs. examined the ability of dcnitrifying cnzymcs to pcr- E/j&t of C and N on DEA. A final factorial sist in dry desert soil. All soils in this study were experiment examined the effect of C and N avail- collected during the dry season and stored in this ability on enzyme activity in wet desert soil. Sixteen, condition (cu. 2% water) for over 2 yr. Comparison IO g subsamples from the composite soil of each site was made bctwccn the denitrifying enzyme content of wcrc incubated aerobically for 3 days at -0.05 MPa these soils and the enzyme content of the same soils and 30’C. Following incubation. 4 subsamples from after 3 days of simulated wet season conditions each site were randomly assigned to I of 4 treatments. (-0.05 MPa and 30-C). The enzyme content of the The treatments included: (I) an assay solution with dry soil was measured by the standard DEA assay supplementary C (6.94m~ dextrose) but no sup- without any preliminary incubation period. The en- plementary N; (2) an assay solution with supplemen- zyme content of wet soil was measured following the tary N (35.2m~ KNO,) but no supplementary C; standard incubation and assay procedures. All com- (3) an assay solution with both supplementary C posite soil samples from each transect in the Mojave and N in the same forms and amounts used in and Great Basin deserts were used. Soils from the treatments I and 2; and (4) a control assay solution Chihuahuan desert were not included because the with no supplementary C or N. All assay solutions factorial experiments performed on those soils contained 100 mM NaH,PO, buffer (pH = 7.00), and already included a dry vs wet soil comparison. 1.5 g chloramphenicol I-‘. The amounts of C and N, when added. were sufficient to saturate denitrification Enzyme actkity studies enzyme activity. All subsamples were slurried with Scparatc experiments examined the effects of pH. IO ml of the appropriate assay solution before the temperature. and N and C availability on denitrifying DEA was measured. enzyme activity. The general procedure involved in-

cubating subsamples of soil to provide replicates with RESULTS equivalent enzyme contents, and then altering the conditions of the DEA assay in order to measure the Enzyme content and production studies effect of a variable on enzyme activity. The specific Predictors of enzyme content. Indices of C and N incubation conditions depended on the experiment, availability were strong predictors of the denitrifying 848 WILLIAM T. R~XJOHN

-104 4 0 5 10 15 20 25

I-DAY PO7ENllAL NITRFIC*TlON (nip N 2 ‘1)

7-V POlENTUL NlllWN3TlON (ng N 2.‘) Fig. I. Relationship between denitrifying enzyme content Fig. 2. Relationship between dcnitrifying enzyme content (measured by DEA) and C availability (measured by CO, (measured by DEA) and NO; availability (measured by production) in bajada soils. Relationship for Chihuahuan potential ) in bajada soils. Relationship for desert soils (upper graph) has 2 outlicrs removed. Numbers Chihuahuan desert soils (upper graph) has 2 outlicrs rc- identify the site. Rclalionship for composite soils from all moved. Numbers identify the site. Relationship for com- deserts (lower graph) has I outlier removed. Letters identify posite soils from all deserts (lower graph) has I outlier the desert. removed. Letters identify the desert. enzyme content in wet bajada soils, both within the ation between enzyme content and CO* production Chihuahuan desert and between the major deserts of (r = 0.722; P < 0.0120). the southwestern U.S.A. (Figs I and 2). In the Effect of C and N on enzyme production. In the first Chihuahuan desert, enzyme content correlated experiment. where I ml of the treatment solutions strongly with CO, production, potential nitrification. was mixed into the soil, the following treatment and total N (Table 2). For all deserts, enzyme content effects were found to be significantly different: in composite samples correlated strongly with poten- water + NO; = water > control = water + C + NO; tial nitrifkation. CO* production, and silt content = water + C (P <0.05; Fig. 4). This pattern was (Table 2). Various combinations of predictors could found in all 3 Chihuahuan desert bajadas. In the explain 71% of the variation in enzyme content for second experiment, where 5 ml of the same treatment Chihuahuan desert soils, and 87% of the variation for composite samples from all deserts (Table 3). Although pH was not strongly associated with deni- Table 2. Predictors of DEA in desert soil under dmulated wet seawn trifying enzyme content. high values (DEA > IS ng N conditions based on simple linear rcgmrion. Only samples from bajadas arc in&dcd g-’ h-l) were found only in soils with a pH ranging between 7.5 and 8.2. This suggests that pH limits the Pearson category correlation Probability production of denitrifying enzymes beyond a certain VWiillC n cocflicicnt level range. Chihuahua” desert Playa soils frequently deviated from the patterns (2 oullien removed) found when only bajada soils were considered Co, production 319 0.795 O.ooOl (Fig. 3). Playas tended to have finer soil textures, Potential nitrifiution 319 0.670 0.0001 higher pH values, and extreme (high or low) denitri- Total N 319 0.636 0.0001 All dcscrtr fying enzyme content. When all playas were con- (I outlier removed) sidered, no strong predictors of denitrifying enzyme Potential nitrification 44 0.863 0.0001 content were identified. However, one of the playas CO, production 44 0.853 O.ooOl in the Chihuahuan Desert exhibited a strong associ- Silt 44 0.708 0.0001 Denitrifiation in desert soils 049

TAk3.PRdi&UlOfDEAilldaacWilundatillktCdWUXUOll ccmditio~ bwd on stqwine multipk linear rcgreuion. Dnly nmpks from hjadas are in&bded.Within acb categorythe model r* ValueI UC cumuktin C-WY Model ROlWbitity VuiAtc 9 kvel chiburblum desert (2 outlicn removed) cq ptod-ion 0.632 0.632 0.0001 Potential nitritiatioa 0.026 0.658 O.ooOl Sill 0.014 0.673 0.0002 organic c 0.027 0.700 0.0001 carbonate c 0.004 0.704 0.0436 PH 0.006 0.710 0.0146 =Y 0.002 0.712 0.1350 All deserts (I outlicr mnovcd) Potmtid nitrification 0.745 0.74s O.oooI Silt 0.063 0.808 o.@lO7 organic c 0.031 0.839 0.0079 Ammonium N 0.023 0.862 O.OIJJ PH 0.010 0.872 0.1004 solutions were added to the top of soil columns, the following treatment effects were significantly differ- ent: water + C + NO; * water + NO; > water = control = water + C (P < 0.05; Fig. 5). Significant differences for a given site were determined by l-way analysis of variance (ANOVA) followed by a Tukey’s c EJ SW1 multiple comparison of the means (SAS, 1982). Enzyme production (as measured by DEA) in both the first and second experiments was similar in the control (7.23 vs 8.43 ng N g-’ h-l). waler (9.59 vs 0 8 10 IS *o Rate (ng N tJ’ h.’ ) Fig. 4. Trcrtmcnt effects on the production of dcnitrifying P C-CWRdWl enzymes after I ml of treatment solution was mixed into soil. m O-OmmiM Soils were composite samples from 3 sites in the Chi- 4a huahuan desert. Bars with ditrcrcnt letters indicate signifi- cant differences (P c 0.05). Horizontal lines represent P I SEM.

II.88 ng N g-’ h-l), and water + C (5.69 vs 5.57 ng P N g-’ h-l) treatments. However, differences between wf 0 the first and second experiments were found for the I water + NO; (I 1.91 vs 30.46 ng N g-’ h-l), and 100 water + C + NO; (5.36 vs 89.73 ng N g-’ h-‘) treat- P% Ido op 0 *Q&i ments. In both experiments. more than 70% of the 0 w w 40 so 80 .lO ~rcwruenowyocQ.~d-1, enzyme content in the water treatments was originally

Control ;‘ a IZJ NM site 3 Water

Water+N b

Water+C : a I3

Water+C+N

20 40 60 00 100 Rate (ng N g” h-‘)

Fig. S. Treatment etlects on the production of dcnitrifying enzymes r!Ier 5 ml of treatment solution were poured on to Fig. 3. Bivariare plots of Co, production and potential soil columns. Soil was a composite sample from NM site 3. nitrification vs the dcnitrifying enzyme content of all soils Bars with different letters indicate significant differences (bajada + playa). (P ~0.05). Horizontal lina represent I SEM.

8.. UP-D 850 WILLLM T. RTERJOHX

present in the dry soil (controls) indicating that persistent denitrifying enzymes arc important in the Chihuahuan desert. Enqme persistence in &y soil. Persistent denitrify- ing enzymes are also important in the Mojave and Great Basin deserts. Typically, > 50% of the enzyme content in wet desert soil was present in the dry soil, even after storage for > 2 yr (Fig. 6). Wet and dry soil enzyme contents were strongly correlated (r = 0.939; P < 0.0001). However, the ratio of dry to wet soil enzyme content was not strongly associated with any of the measured soil parameters. CO, production was the best single predictor of the ratio of dry to wet soil enzyme content (r = 0.428; P < 0.0103). and was the 3 only variable entered in stepwise multiple linear regression. Fig. 7. The etkt of pH on denitrifying enzyme activity. Soils were composite samples from the Chihuahuan Desert. En:yme acricity studies Numbers identify the study site. The curve was generated Eflecr of pH on DEA. The dependence of DEA on from the theoretical equation for a diprotic system. the pH of the incubation solution was bell-shaped with a narrow optimum range centered around a pH of 7.0 (Fig. 7). Enzyme activity was reduced by at Votm= V,na,/(l+ (W’IIK,) + UW[H+l)). (1) least 50% if the pH was c6 or ~8. and very low where V,, = observed rate, V,,,,, = rate when all the activity was found when the pH was <5 or >9. enzyme is in the active form, [H+] = hydronium ion Good agreement was found between the observed concentration, and K,, and KU2= first and second data and the equation for a diprotic system where acid dissociation constants (Price and Stevens, 1982). successive pK, values are closer than 3.5 pH units The equation was fit using a multivariate nonlinear (Scgcl. 1975; Alberty, 1983). This equation is: procedure (SAS, 1982) which estimated V,, to bc 21.32 ng N g-’ h-‘, and K., and Kw2to be 5.89 x IO-’ and 1.47 x IO-‘. rcspcctivcly. F’ct o/ lrnlprrolltre on D&l. Increasing the tcm- pcrature of the assay solution cau.scd an increase in DEA. up to a tcmpcraturc of about 40°C. At tcm- pcraturcs higher than 4OC the cnzymc activity rapidly dccrcascd. and ccascd altogether when tcm- pcratures excccdcd 60 ‘C (Fig. 8). In the range from IO to 40 C. the obscrvcd activity is in good agreement with that predicted by the Arrhcnius equation. This equation is:

In(v,,,) = -(E,IR)(IIT) + C, (2) where Vah = observed rate, E. = activation energy, im 100 ma 400 so3 aa 700 mTSO(LoU (ngNg”h.‘) R = gas constant, T = temperature in “K. and C = a constant (Alberty, 1983). A linear regression pro- ccdure was used to estimate the slope (4935.7) and intercept (19.452) for Eq. 2. Since the slope estimates

Fig. 6. Denitrifying enzyme content in wet vs dry soil. Enzyme content was measured as DEA. Samples above the Fig. 8. The effect of temperature on denitrifying enzyme I:2 line had > 50% of their wet soil enzyme foment activity. Soils were composite samples from the Chihuahuan originally present in the dry soil. Only samples with a DEX desert. Numbers identify the study site. The curve was fit of <2OOng N g-’ h-’ are plotted in the lower graph. by eye. Denitrification in desert soils 851

EJR, multiplying by the gas constant provides an Table 4. Compatisott of the dmitrifyinp enzyme content estimate for the activation energy. For this study, the in soils of different habitats activation energy was about 41 kJ mol-‘. This value Enzyme content compares well with the range typical of enzyme as measured by DEA Habitat catalysis (20.9-62.8 kJ mol-‘; Segel. 1975). and is (ng N g-’ h-‘) very close to the value calculated by Parker et 01. Forest soils’ NC hardwood 84 (1983) for soil respiration in a wet Chihuahuan desert MI hardwood 84 soil (39.5 kJ mol-I). The Q,,, for DEA in Chihuahuan MI hardwoodt 190 desert soil was 1.74. NM aspen 364 Efect o/C and N on DEA. Enzyme activity showed Venauclan rainforest 392 Nigerian rainforest 224 a slight increase when supplementary C was added to Agricul~~nl soils the assay medium, either by itself or in combination GA corn* 28 with supplementary N (Fig. 9). No increase in activity Ml corn* 168 was detected when supplementary N was added to the IA corn* 280 MD corn; 126 assay medium. These results were consistent for all 3 MD corn. no-till: 285 sites in the Chihuahuan desert. Since the controls in MD corn. no-till$ ISI this experiment contained only the C and N made New Zealand pasture* 280 available during the 3-day incubation period. the Freshwater sediments’ White cedar swamp 1820 results indicate that under wet season conditions the Farm stream 2100 availability of N does not limit the rate of enzyme Eutrophic. pelagic lake 6440 activity, and the availability of C only slightly limits Desert soils’; enzyme activity. NM bajadas 9 CA bajjadas I3 NV bajadas 43 NM playas 192 CA playas 237 NV plilyas I63

lTicdjc cf RI. (ISilL tGrolTm;m and Ticdjc (IYX9). tl’arkm CI ol. (IYX7). q SITE 1 48;ukin and Robinson (IYHY). ?Pctcrjohn. iw. rir.

DISCUSSION. . En:yntc content und prodticlion The dcnitrifying cnzymc content of a soil is thought to bc a long-term intcgrativc index. rcflccting the history of cnvironmcntal variables that change rapidly at a given site (GrotTman and Ticdje. 1989; Ticdje Ed ul.. 1989). The content of denitrifying enzymes in soil strongly corrclatcs with the annual dcnitrification rate in forest ecosystems (Groflinan and Tiedje, l989), and has been used to predict the mean dcnitrification rate in an agricultural soil (Parkin and Robinson, 1989). The content of dcnitrifying enzymes in upland NTRXE- desert soils is low (Table 4). and is strongly associated

0 5 10 with indices of both C and N availability. This agrees with general theories suggesting that nutri- ent loss is greater in systems with higher nutrient availability (Vitousek ef al., 1979, 1982; Matson and .;‘;‘;‘;.;‘;‘;‘ ...... a , I . . . I . Vitousek. 1987). In addition, other studies demon- ...... strate higher rates of denitrification in sites with high ,a,,.,* ...... *.-.*.~.-.*.*. a nutrient availability (Melillo et al., 1983; Robertson .‘.‘.‘.‘.‘.‘.‘. cl SW 3 and Tiedjc. 1984, 1988). and that denitrifying enzyme . . ...*.,,***...... I. .,,,,,.,,.** ...... b content is strongly associated with soil carbon and .‘.‘.‘.‘.‘.‘.‘.‘.‘.‘.’ microbial biomass (Myrold and Tiedjc. 1985b; Tiedje I% et al.. 1982). In desert ecosystems. a significant fraction of the dcnitrifying enzyme content is dcrivcd from a persist- ent pool of enzymes capable of tolerating extended Rate (ng N g1 K1 ) periods of desiccation. Such desiccation tolerance Fig. 9. Treatment efkcts on denitrifying enzyme activily. has been demonstrated in other soils and allows Soils were composite samples from 3 sites in the Chi- denitrifiers to respond rapidly to favorable conditions huahuan desert. Ban with different letters indicate signifi- (Smith and Parsons, 1985; Groffman and Tiedje. cant differences (P < 0.05). Horizontal lines represent 1988; Martin er 01.. 1988). Rapid response by deni- I SEM. trifiers may be particularly important in desert 852 WILLIAM T. &TERJOHN ecosystems where moisture availability is highly both extremes and was low in soils with a pH > 8.2 episodic. The mechanism for desiccation tolerance of (Peterjohn, lot. cit.). In a broader survey of playas in denitrifying enzymes is currently not known. the southwestern U.S.A., Schlesinger and Peterjohn The production of denitrifying enzymes in desert (1988) measured pH values ranging from 8.22 to soils responded to experimental changes in water, C 10.91. Thus, pH may prevent significant rates of and N availability. In all cases. the addition of water denitrification from occurring in most playa soils. stimulated enzyme production. but the addition of C Denitrifiers may adapt to the environmental pH along with water prevented enzyme synthesis. The (Parkin er of., 1985). but this is unlikely to have effect of added C was probably due to immobilization occurred in desert soils because respiration rates in of available N and indicates that N availability is samples with a pH > 8.20 were very low (Peterjohn, necessary to initiate enzyme synthesis in desert soils. lot. cit.). Laboratory studies have also demonstrated that NO; The optimum temperature for denitrifying enzyme is either required for or greatly enhances the pro- activity in desert soils is close to 40’C. This is duction of denitrifying enzymes (Firestone, 1982; substantially lower than an optimum of 65’C re- Komer and Zumft, 1989). Enzyme production in ported in earlier studies (Nommik, 1956: Bremner response to N and C + N additions, however, was and Shaw, 1958b). but agrees with the optimum for sensitive to the exact nature of the treatment con- soil respiration measured in a Chihuahuan desert soil ditions. When 5 ml of treatment solution was poured (Parker ef al., 1983). Keeney er al. (1979) consider onto soil columns, both significant N and C + N 65’C too high for the true optimum of biological response occurred. However, when only I ml of the denitrification due to chemical decomposition of same solutions was mixed into the soil, enzyme NO; at temperatures greater than 5O’C. production showed no significant N limitation, and During the summer, surface soils in deserts can be the addition of C + N suppressed enzyme formation. > IO’C hotter than the air temperature, and soils The reason for the differential response to N and experience diurnal fluctuations of > 30‘C (Billings. C + N additions is not known, but the large diffcr- 1978; Nobel and Geller, 1987). Mean minimum and ences indicate complex interactive effects between C. maximum temperatures during July were 22.8 and N and soil moisture 46.2C at a depth of 5 cm in Chihuahuan desert soils (Wicrcnga. 1988). and about 22.2 and 60’C at a depth Enzyme activity of 2.54 cm in Great Basin dcscrt soils (Billings et (II.. Recent studies of in situ dcnitrification have shown 1954). For the same sites the mean minimum and that the temporal variability of cnzymc content in a maximum soil tcmperaturcs in January wcrc 1.3 and given soil is much lowcr than the temporal variability 13.2 ‘C in the Chihuahuan dcscrt. and about -3.9 of the actual dcnitrification rate (Smith and Parsons, and I5.6C in the Great Basin dcscrt. Tcmpcraturc 1985; Grotfman and Ticdjc, 1989; Parkin and ditfcrcnccs of IO‘C bctwccn cxposcd and shaded soils Robinson. 1989). This suggests that most of the have also been rcportcd in the Chihuahuan and variability in the actual dcnitrification rate at a given Sonoran dcscrts (Parker et ul., 1983; Nobel and site is due to changes in the activity of existing Gcllcr. 1987). Thus, tcmperaturc should have a enzymes rather than changes in the enzyme content strong temporal and spatial cffcct on the rate of of the soil. This is likely to also be the case in desert dcnitrification enzyme activity in desert soils. Few ecosystems, since desert soils contain a significant thcrmophilcs are found in dcscrt soils, and it is pool of persistent denitrifying enzymes, and only important to realize that most microbial activity small changes in enzyme content result from the occurs at the lower temperatures associated with addition of water. grcatcr soil moisture (Skujins. 1984). For example a Factors known to affect the activity of denitrifying single rain storm in the Great Basin desert dccrcascd enzymes include: pH; temperature; C and N concen- the maximum soil temperature at a depth of 2.54 cm tration; and the partial pressure of Oz. The optimal from 49 to 26’C in 5 days (Billings et al.. 1954). pH for DEA in Chihuahuan desert soil is about 7.0, In wet desert soils the availability of N was suffi- which agrees with other studies of dcnitrification cient to saturate denitritication enzyme activity, and (Wijlcr and Delwiche, 1954; Nommik, 1956; Bremner the availability of C only slightly limited enzyme and Shaw, 1958b; Klemedtsson er al., 1977; Fire- activity. Michaelis-Menton constants (K,,,) for NO; stone, 1982; Knowles, 1982). The pH of desert soils reduction by denitrification vary by >3 orders of in the southwestern U.S.A. ranges from 5.47 to 10.9 I, magnitude (Yoshinari er al.. 1977; Myrold and with a median value of 7.85 for upland soils Tiedje. 1985a; Murray el al., 1989). The higher K,,, (Schlesinger and Peterjohn, 1988). Most desert soils values reported for soils. however, may have resulted are highly buffered by CaCO,. so at any given from denitrification being limited by C or the rate of location there is little temporal change in pH. Thus, NO; diffusion (Firestone. 1982; Myrold and Tiedje, pH may have an effect on the spatial variability of 1985a; Murray er al.. 1989). Recent work is in closer DEA in deserts, but little effect on the temporal agreement with the lower values reported for isolated variability. cultures, which indicates that low values are probably Playas should be conducive to denitrification be- more appropriate. Low K,,, values (< 140 pM NO;) cause they often have fine textured soils, a high also indicate that enzyme activity is probably not moisture holding capacity, and relatively high nutri- limited by NO; in desert soils. For example, in ent availability. Indeed, playas in this study had saturated extracts of desert soils, a conservative esti- a higher average denitrifying enzyme content mate of the NO; concentration is 300~~, which is than upland desert soils (Table 4). The enzyme close to saturation for the lower values of K,,, content of individual playas, however, represented that have been reported (Klemmedtson er al., 1977; Denitrification in desert soils 853

Christensen and Tiedje. 1988; Murray et al., 1989; Acknowledgements--f thank Tony Ratbburn. Scott Nielsen er al.. 1990). Bridgbam. Larry Yungk. Jay O’NeilJ and Wa Wi for Several studies demonstrate that dcnitrification field and laboratory assistance. Valuable advia was pro- depends strongly on the availability of C (Bremner vided by William Schlesinger, Tim Parkin. Jim Siaiow, Phil and Shaw, 1958a. b; Bowman and Focht, 1974, Robertson and Pam Matson. Generous IogisticaJ support was nrovided bv Carol Wells. Curt Richardson. Bruce Burford and Bremner. 1975; Stanford er of., 1975; Co&. Walt Whitford and. Paul Fonteyn. Wi Westerman and Tucker, 1978; Reddy ef al., 1982). Schlesinger and Lisa D. Schlesinger provided valuable edi- These studies. however, do not separate enzyme torial comments. Financial support came from a National production (bacterial growth + enzyme synthesis) Science Foundation Dissertation Improvement Award, and from enzyme activity. I found little response of a NSF grant awarded to the Jomada Long Tetm Ecological enzyme activity to supplementary C and N in wet Research program. desert soils. Therefore, the large increase in total denitrification (enzyme production + activity) due to REFERENCES C + N additions in wet desert soils (Westerman and Tucker, 1978; Peterjohn. lot. cit.) is probably the Alberty R. A. (1983) Physical Ckemisrry. Wiley, New York. result of enhanced enzyme production rather than Billings W. D. (1978) Plunrs und rhe Ecosysfem. Wadsworth, Belmont. enhanced enzyme activity. Billings W. D., Humphrcys M. K. H. and Darling J. B. High availability of C and N for denitrification (1954) Environmental studies in the eold deserts and in freshly wet desert soil is consistent with many semidesertsof the western Great Basin of North America. studies that report a burst of nutrient availability and Report of Contract DA 44-IW-qm-1261, Quartermaster microbial activity after wetting dry soils (Myers and Research and Development Command, U.S. Army, McGarity, 1971; McKenzie and Kurtz. 1976; Patten Natick. er al.. 1980; Kroeckel and Stolp, 1986; Kieft et al.. Bonin P., Gilewicz M. and Bertrand J. C. (1989) ERects of 1987; GrofTman and Tiedje. 1988). Drying and wet- on each step of denitrification in Pseudomonas ting cycles are common in deserts, and they may naufica. Canadian Journal 0jMicrobiology 35. 1061-1064. Bowden W. D. (1986) Gaseous nitrogen emissions maintain nearly optimal substrate concentrations for from undisturbed terrestrial ecosystems: an assessment denitrification enzyme activity in wet dcscrt soil. of their impacts on local and global nitrogen budgets. Oxygen inhibits both the synthesis of dcnitrifying Biogcochemisrry 2, 249-279. enzymes and their subsequent activity (Firestone, Bowman R. A. and Focht D. D. (1974) The influence of 1982; Knowles. 1982: Parkin and Ticdjc, 1984; Bonin glucose and nitrate concentrations upon denitrification er rrl.. 1989; Korner and Zumft. 1989). Dcscrt soils arc rates in sandy soils. Soil Biokogy & Biochemisrry 6, often dry. poorly aggrcgatcd. and aerobic. Thus, 297-301. dcscrts would seem unsuitable for dcnitrification. Bremncr J. M. and Shaw K. (1958a) Denitrification in soil I. Howcvcr. the fact that dsnitritication occurs shortly Methods of invest& [ion. Journul o/ Agriculrurd Science 51, 22-39. after wetting indicates that anaerobic sites must Brcmner J. M. and Shaw K. (1985b) Denitritication in soil II. bccomc established quickly in dcscrt soils. Although Factors atl’ecting dcnitritication. lournal 01 Agriculrurd the eff’cct of O1 was not studied directly, it seems Science 51. 40-52. likely that the formation of anaerobic sites in dcscrt Burford J. R. and Bremner J. M. (1975) Relationships soils is closely coupled to soil moisture. tcmperaturc. between the denitrification capacities of soils and total, and available C. Soil moisture decrcascs the diffusion water soluble and readily decomposable soil organic rate of 0, by 4 orders of magnitude, and increased matter. & Biochemistry 7, 389-394. tcmperaturcs can significantly decrease the solubility Christensen S. and Tiedje J. M. (1988) Sub-parts-per-billion of 0, in water (Smith, 1980). Soil moisture, increased nitrate method: use of an N,O-producing denitrifier to convert NO; or “NO; to N1O. Applied and Enuiron- temperature, and available C all enhance respiration mental Microbiology 54. 1409-1413. rates and thus accelerate 0, depletion. In addition, Firestone M. K. (1982) Biological denitrification. In is important in the formation of soil Nitrogen in Agric&&l Soils -(F. J. Stevenson, Ed.). aggregates where anaerobic microsites can occur DD. 289-326. Soil Science Societv of America. Madison. (Sexstone er al., 1985; Tiessen and Stewart, 1988). &&man P. M. and Tiedje J. M. (1988) Denitrification Further research is needed before the role of 0, in hysteresis during wetting and drying cycles in soil. Soil controlling denitrification in desert ecosystems is fully Science Society of American Journal 52, 16261629. understood. Grogman P. M. and Tiedje J. M. (1989) Denitrification in From my investigations of denitrification in desert the north temperate forest soils: relationship between denitrification and environmental factors at the landscape ecosystems, deserts seem more suitable for this micro- scale. Soil Biology & Biochemistry 21, 621626. bial process than it would appear at first. Although Keeney D. R.. Fillery 1. R. and Marx G. P. (1979) Effect of the enzyme content of upland desert soils is low when temperature on the gaseous nitrogen products of denitrifi- compared to that in other habitats, the denitrification cation in a silt loam soil. Soil Science Society of Americu rates in wet desert soils are 2 than those measured Journal 43, I 124-l 128. in more mesic ecosystems (Peterjohn, lot. cit.). Thus, Kieft T. L.. Soroker E. and Firestone M. K. (1987) Micro- the nearly optimal conditions that exist for several bial biomass response to a rapid increase in water poten- factors that influence enzyme activity in wet soil tial when dry soil is wetted. Soil Biology & Biochemitrry apparently compensate for the low enzyme content. 19. 119-126. Klemedtsson L., Svensson B. H.. Lindherg T. and Rouwall From this study it is also evident that denitrification T. (19771 The use of acetylene inhibition of in desert ecosystems is triggered by rainfall and is reductase in quantifyinghenitrification in soils. Swedish tightly coupled to the complex interplay between Journal of Agriculrurol Research 7, 179-185. soil moisture, carbon. - and nitrogen availability, pH. Knowles R. (1982) Denitrihcation. Microbiological Reuiews temperature, and O*. 46. 43-70. 854 WILLtAMT. PETEIuoH?J

Korncr H. and Zumft W. G. (1989) Expression of denitrifi- Petetjohn W. T. and Schlesinger W. H. (1990) Nitrogen cation enzymes in response to the dissolved oxygen level loss from deserts in the southwestern United States. and respiratory substrate in continuous culture of Pseu- Biogeochemirrry 10. 67-79. domonas stuI:eri. Applkd and EnrironmenIal Microbiology Price N. C. and Stevens L. (1982) Fundamentals of &n:ymol- 55, 167&1676. ogy. Oxford University Press, New York. _ _ Kroeckel L. and Stolp H. (1986) Influence of the water Reddv K. R.. Rao P. S. C. and J~SSUD R. E. 119821. The regime on denitrification and aerobic respiration in soils. effect of carbon mineralization on dehitriftcation kinetics B&logy and Fertility of Soils 2, 15-2 I. _ in mineral and organic soils. Soil Science Society of McKenzie E. and Kurtz L. T. (1976) E&t of oretreatment . I American Journal 46, 62-68. on loss of nitrogen-i S-labelled nitrogen from Robertson G. P. and Tiedje T. M. (1984) Denitrification and waterlogged soil during incubation. Soil Science Sociery nitrous oxide production in successional and old-growth of American Journal 40, 534537. Michigan forests. Soil Science Society of America Journal Martin K., Parsons L. L., Murray R. E. and Smith M. S. 48. 383-389. (1988) Dynamics of soil denitrifier populations: relation- Robertson G. P. and Tiedje J. M. (1988) Deforestation ships between enzyme activity. most-probable-number alters denitrification in a lowland tropical rain forest. counts, and actual N gas loss. Applied und Environmental Nature 336. 756759. Microbiology 54. 27 I I-27 16. SAS (1982) Urer’s Guide: SruIisIics. SAS Institute. Cary, Matson P. A. and Vitousek P. M. (1987) Cross-system N.C. comparisons of soil nitrogen transformations and nitrous Schlesinger W. H. and Peterjohn W. T. (1988) Ion and oxide flux in tropical forest ecosystems. Global Biogeo- sulfate ratios in arid soils subject to wind erosion in the chemical Cycles I. 163-170. southwestern U.S.A. Soil Science Society of America Melillo J. M.. Aber J. D.. Steudler P. A. and Schimel J. P. Journal 52. 54-58. (1983) Denitrification potentials in a successional se- Schlesinger W. H.. Fonteyn P. H. and Marion G. M. (1987) quence of northern hardwood forest stands. Ecological Soil moisture content and plant transpiration in the Bulletins 35. 2 17-228. Chihuahuan Desert of New Mexico. Journal of Arid Mueller-Dombois 0. and Ellenbcry Cl. (1974) Aims and Ent~ir0nmmI.t 12, I 19-l 26. MeIhua!s of VegeraIion Ecolugy. Wiley. New York. Segcl I. H. (1975) Enzyme Kinerics: Behavior and Anuly.Tb Murrav R. E.. Parsons L. A. and Smith M. S. (1989) of Rapid Equilibrium and Steac{v -SIare En:yme Sytems. Kin&s of nitrate utilization by mixed pop&ions Wiley, New York. of dcnitrifying bacteria. Applkd cd Erwironmcwul Scxstonc A. J.. Rcvsbcch N. P., Parkin T. B. and Tiedje Microhiolugy 55. 7 I7 -72 I. J. M. (1985) Direct mcasuremcnt of oxygen proftics and Myers R. J. K. and McGarity J. W. (1971) Factors intlucnc- dcnitrifiwtion rates in soil aggregates. Suil Scirnce Suckry ing high dcnitrifying activity in the subsoil of solodizcd uf America Juurnal 49, 645 .65I. solonctz. Plunr and Suil 35. I45 I ho. Skujins J. (19X4) Microbial ecology of desert soils. AJr*unce.f Myrold I>. D. and Ticdjc J. M. (IOXSa) Diffusional con- i/t Alicrobial Ecolugy 7, 40 9 I. straints on dcnitrilication in soil. Soil Scicww SucicIy uj Skujins J. J.. Braal L. and McLarcn A. D. (1963) Anwricun Journal 49. 65 I 657. Charactcri7.tiun of phosphatase in a terrestrial soil Myrold 0. D. and Ticdjc J. M. (I9XSb) Establishment of stcrilizcd with an clcctron beam. Enzymulugia 25. dcnitrilication capacity in soil: clfccts of’ carbon. nitrate I25 -133. and moisture. Suil B&gy d Biuchcwristry 17. 819 822. Smith K. A. (1980) A model of the extent of anaerobic Nielsen L. P.. Christensen P. B.. Rcvsbcch N. P. and zones in aggregated soils and its potential application to Sorcnscn J. (1990) Dcnitrilication and oxygen respiration estimates of denitrification. Juurnal uf Suil Science 31, in biotilms studied with a microsensor for nitrous oxide 263-217. and oxygen. Microbial Ecology 19. 63 -72. Smith M. S. and Parsons L. L. (1985) Persistence of Nobel P. S. and Gcllcr G. N. (1987) Tcmpcruture modelling denitrifying enzyme activity in dried soils. Appkd and of wet and dry desert soils. Thr Journal of Ecolugy 75, Environ~tenIal Micrubiulugj 49. 3 16-320. _ 247-25n. Smith M. S. and Tiedie J. M. 119791 Phases ofdenitrification Nommik H. (1956) Investigations on denitrification in soil. following oxygen depletion in ioil. Soil Biology.~. & Biu- Acra Agriculfurue Scundinuvica 6. 195-228. chemisrri I I; 26 I-287. Parker L. W.. Miller J.. Steinberger Y. and Whitford W. G. Stanford 6.. Vander Pol R. A. and Dzienia S. (1975) (1983) Soil respiration in a Chihuahuan desert rangeland. Denitriticaiion rates in relation to total and extractabl; Soil Biology & Biochemisrry 15, 303-309. soil carbon. Soil Science Society of America Proceedings Parkin T. B. and Robinson J. A. (1989) Stochastic models 39, 284-289. of soil denitrification. Applied und Environmental Micro- Tabatabai M. A. (1982) Soil enzymes. In MeIhoa!r of Soil biology 55, 72-77. Analysis. Parr I/ (A. L. Page, R. H. Miller and D. R. Parkin T. B. and Tiedje J. M. (1984) Application of a soil Keeney. Eds), pp. 903-947. Soil Science Society of core method to investigate the effect of oxygen concen- America, Madison. tration on denitrification. Soil Biology & Biuchemisrry 16, Tiedje J. M. (1982) Denitrification. In Merhoak of Soil 331-334. Analysis. Parr I/ (A. L. Page. R. H. Miller and D. R. Parkin T. B.. Starr J. L. and Meisinger J. J. (1987) Influence Kenney, Eds), pp. 101 I-1026. Soil Science Society of of sample size on measurement of soil denitrification. Soil America, Madison. Science Society of America Journal 51. 1492-1502. Tiedje J. M. (1988) Ecology of denitrification and dissimila- Parkin T. B., Sexstone A. J. and Tiedje J. M. (1985) tory nitrate reduction to ammonium. In Biology of Adaptation of denitrifying populations-to low soil pH. Anaerobic Microorganisms (A. J. B. Zehnder. Ed.), AImlied and Environmental Microbiology 49. 1053-1056. pp. 179-244. Wiley, New York. Pa&n T. B.. Kaspar H. F.. Sexstone A. 7: and Tiedje J. M. Tiedje J. M.. Simkins S. and Groffman P. M. (1989) (1984) A gas-flow soil core method to measure field Perspectives on measurement of denitrification in the field denitrilication rates. Soil Biology & BiochemisIry 16, including recommended protocols for acetylene based 323-330. methods. Plant and Soil I IS. 261-284. Patten D. K., Bremner J. M. and Blackmer A. M. (1980) Ticdje J. M.. Sexstone A. J.. Myrold D. D. and Robinson Effects of drying and air-storage of soils on their capacity J. A. (1982) Denitrilication: ecological niches, compe for denitrification of nitrate. Soil Science Society of tition and survival. Anfonie van Leeuwenhoek Journal of America Journal 44, 67-70. Microbiology 48. 569-583. Denitrification in desert soils 855

T&en H. and Stewart J. W. B. (1988) Light and electron denitrification in a Sonoran desert soil. Soil Scirnce microscopy of stained mieroaggregates: the role of Society of America Jouwur~ 4f 59&599. organic matter and microbes in soil a~gation. Bio- Wierenga P. J. (1988) Weather data: daily and monthly geochemisrry 5. 3 12-322. means. March I983Gctober 1987. Jomada LTER site. Virginia R. A.. Jar&t W. M. and France-Vizcaino E. (1982) Research Report 88-m-02, NSF. Project No. BSR- Direct measurement of denitrification in a Prosopti 8612106, Department of Agronomy and Horticulture, (mesquite) dominated Sonoran Desert ecosystem. Oecolo- New Mexico State University. Las Cruces. gia 53, 120-122. Wijler J. and Delwiche C. C. (1954) Investigations on the Vitousek P. M., Gosz J. R.. Crier C. C.. Melillo J. M. and denitrifying process in soil. Pfant and Soil 5, 155-169. Reinen W. A. (1982) A comparative analysis of potential Wilhem E., Battino R. and Wilcock R. J. (1977) Low-press- nitrification and nitrate mobility in forest ecosystems. ure solubility of gases in liquid water. Chemical Reviews Ecologicaf Monographs 52. 155-177. Tl, 219-262. Vitousek P. M.. Gow J. R., Grier C. C.. Me&ho J. M.. Yoshinari T., Hynes R. and Knowles R. (1977) Acetylene Reiners W. A. and Todd R. L. (1979) Nitrate losses from inhibition of nitrous oxide reduction and measurement of disturbed ecosystems. Scirnte 204. 469-474. denitritication and in soil. Soil Biobgy West N. E. and Skujins J. ( 1977) The in North & Biochemistry 9. 177-183. American cold-winter semi-desert ecosystems. Oecologiu Young D. R. and Nobel P. S. (1986) Predictions of Planrarum 12. 45-53. soil-water potentials in the north-western Sonoran Westerman R. L. and Tucker T. C. (1978) Factors affecting Desert. Journal of Ecology 74. 143-154.