Production, oxidation and emission of methane in rice paddies

Production, oxidation and emission of methane in rice paddies

FEMS MicrobiologyEcology31 (1985) 343-351 Published by Elsevier 343 FEC 00045 Production, oxidation and emission of methane in rice paddies (Methan...

709KB Sizes 0 Downloads 46 Views

FEMS MicrobiologyEcology31 (1985) 343-351 Published by Elsevier


FEC 00045

Production, oxidation and emission of methane in rice paddies (Methanogenesis; methanotrophy; submerged soil; rhizosphere; salt effects)

A n n e t t e H o l z a p f e l - P s c h o r n , R a l f C o n r a d * a n d W o l f g a n g Seiler M a x - P l a n c k - l n s t i t u t f,',r C h e m i c , S a a r s t r a s s e 23, D - 6 5 0 0 M a i n z , F . R . G .

Received15 May 1985 Revision received27 August 1985 Accepted 2 September1985


2. I N T R O D U C T I O N

Production and emission of methane from submerged paddy soil was studied in laboratory rice cultures and in Italian paddy fields. Up to 80% of the CH 4 produced in the paddy soil did not reach the atmosphere but was apparently oxidized in the r h i z o s p h e r e . C H 4 emission through the rice plants was inhibited by an atmosphere of pure 0 2 but was stimulated by an atmosphere of pure N 2 or an atmosphere containing 5% acetylene. Gas bubbles taken from the submerged soil contained up to 60% CH 4, but only < 1% CH 4 after the bubbles had passed the soil-water interface or had entered the intercellular gas space system of the rice plants. CH 4 oxidation activities were detected in the oxic surface layer of the submerged paddy soil. Flooding the paddy soil with water containing > 0.15% sea salt (0.01% sulfate) resulted in a strong inhibition of the rates of methanogenesis and a decrease in the rates of CH 4 emission. This result explains the observation of relatively low CH 4 emission rates in rice paddy areas flooded with brackish water.

Rice paddy fields represent soil areas which are flooded for long periods so that anoxic soil conditions arise which allow anaerobic microbial metabolism such as denitrification, sulfate reduction and methanogenesis [1-3]. In principle, rice paddies are similar to the littoral of lakes and are characterized by the presence of plants and the occurrence of oxic and anoxic zones in the sediment. In contrast to profundal lake sediments [4] very little is known about methanogenesis in the littoral zone [5-7] or in rice paddies. With respect to rice paddies, most of our knowledge is based on studies by Koyama [8,9], Yamane and Sato [10-12] and Takai [13] who generally used air-dried soil samples, incubated under anaerobic conditions for prolonged periods. These experiments did not consider the influence of rice plants on the production, oxidation and emission of methane and thus may not be representative for natural conditions. L a b o r a t o r y studies indicated that the aerenchyma and intercellular space system of rice plants mediates the transport of C H 4 from the anoxic sediment into the atmosphere [14,15] and also allows the diffusion of oxygen into the roots [16,17]. Recent field studies have demonstrated that > 90% of the emitted C H 4 is transported by

* To whom correspondenceshould be addressed.

0168-6496/85/$03.30 © 1985 Federation of European MicrobiologicalSocieties

344 the rice plants [18,19]. In the absence of plants, C H 4 is released almost exclusively by emission of bubbles. Diffusional transport through the w a t e r / a i r interface does not contribute significantly to total C H 4 emission. De Bont et al. [15] reported on increased titers of methanotrophic bacateria in the oxic surface sediment as well as in the rhizosphere of paddy soil. From experiments inhibiting methanotrophic bacteria with acetylene the authors concluded that part of the methane is oxidized during its diffusion through the oxic surface layer of the paddy soil, but excluded significant oxidation during diffusion into the rice intercellular gas space system through the oxidized zone surrounding the rice roots. Field studies in rice-growing areas in California [20] Andalusia [18] and Italy [21] revealed different patterns and quantities of C H 4 emission which might have been due to different influences of soil and plants on the microbial production a n d / o r oxidation of methane. In particular, it has been suggested that the presence of brackish water in the Andalusian rice paddies has been the major reason for the lower C H 4 emission rates as compared to Italian paddy fields [21]. We now report on field and laboratory studies which were carried out to assess the relative importance of C H 4 production and oxidation within the rhizosphere on the emission of C H 4 from rice paddies into the atmosphere. The results show that a significant portion of the produced C H 4 is consumed and not released into the atmosphere. We further show that brackish flooding water significantly inhibits methanogenesis and C H 4 emission.

3. M A T E R I A L S A N D M E T H O D S 3.1. Cultivation of rice in the laboratory Glass beakers (800 ml) were filled with 600 g (dry wt.) of air-dried, sieved (3 m m mesh) paddy soil collected in the rice-growing area of Italy. The characteristics of the soil have been described [19]. The soil was then flooded with distilled water. The soil depth in the beaker was approx. 10 cm and was covered by approx. 5 cm of water. Rice seeds (var. Roma, type japonica) were obtained from the Italian Rice Research Institute in Vercelli. After

germination on filter paper soaked with water for 2 - 4 days, 10 seeds were planted in each beaker. The cultures were incubated in a water bath at 25°C and illuminated with incandescent light (Osram power star, 400 W) at a radiation density of approx. 200 W • m - 2 . The light regime was kept at a cycle of 13-14 h light and 10-11 h darkness. After 2 weeks of incubation, the tillers of the rice had reached a length of 15 cm and the redox potential of the soil was < - 2 0 0 mV (pH 6.8). Under these conditions the methane emission rates were in the order of 3.5 mg C H 4 per h per m 2 soil surface area. The C H 4 emission rates steadily increased within the following 2 weeks, and then remained relatively constant during the next 4 months, with values of approx. 10 mg- m -2 • h -~. To assess the influence of seawater or sulfate on methanogenesis and c n 4 emission, cultures were prepared by submerging the paddy soil in solutions of artificial seawater containing 0.15% or 0.3% seawater salt, or in a solution of 0.1% Na2SO 4. Rice seeds were grown as described for the controls prepared with distilled water. In another experiment, the supernatant flooding water of mature control rice cultures was exchanged for artificial seawater solutions. 3.2. Field studies Field studies were carried out in rice paddies of the Italian Rice Research Institute in Vercelli, located in the valley of the River Po. The field sites and the pattern of C H 4 emission during the growing season have been described [21]. 3.3. Measurement of C H 4 emission rates The rice cultures in the glass beakers were placed in a large glass incubation vessel (Fig. 1), gassed for 10 min with ambient air, 02 or N 2, and pressurized to 1.1 bar. Gas samples (10 ml) were taken repeatedly from the incubation vessel with gas-tight syringes and analyzed for C H 4 in a gas chromatograph. Emission rates were determined from the temporal increase of the C H 4 mixing ratio and the volume and pressure of the gas space of the incubation vessel. A total incubation time of 30 min was sufficient to allow the detection of a C H 4 emission rate of 0.2 # g - h -~, equivalent to 0.03 m g - h -1 - m -2 soil surface area. C H 4 emission

345 Septum


per g (dry wt.) soil. The lower detection limit for rates of methanogenesis was approx. 0.5 ng CH 4 • h -1 • g-1 dry wt. soil or 0.5 /Lg,h -~ for the total rate of CH 4 production in the laboratory rice cultures.


Seal _

+ Clamp sou

Fig. 1. Apparatus for the determination of CH 4 emission rates from laboratory rice cultures.

rates under field conditions were determined by a static box technique as described in detail by Seiler et al. [18] and Holzapfel-Pschorn and Seiler I21].

3.5. Measurement of methanogenesis in the fieid Sediment corers made out of glass were pressed into the submerged soil of the paddy field and closed with a septum (Fig. 2). After the corers had stayed in the field for different time periods they were removed and connected with a glass flask which had been flushed with CH4-free air (synthetic N2:O 2 mixtures free of hydrocarbons, Messer-Griesheim) (Fig. 2). The core of the submerged soil was then transferred into the flask by injecting CH4-free water through the septum of the glass corer. The gaseous and dissolved CH4 in


3.4. Measurement of methi~nogenesis in laboratory Rates of methanogenesis were measured in 120ml serum bottles into which cores of submerged soil (10 × 1 cm) were transferred. Cores were taken from the laboratory rice cultures by pressing a glass tube with an open valve at one end into the sediment until it was completely filled. Then the valve was closed, the core retrieved and pressed with N 2 into the serum bottle which was flushed with N 2. The bottles were stoppered, evacuated and refilled 6 times with N2, pressurized with N 2 to a final pressure of 1.4 bar and incubated at 25°C. The temporal increase of the CH 4 mixing ratio in the headspace of the bottles was measured by taking gas samples (1 ml) which were analyzed by gas chromatography. After a lag phase of 1-4 h, CH4 mixing ratios increased linearly with time for at least one day. The rate of methanogenesis was determined from the rate of increase of the CH 4 mixing ratio in the serum bottles, the volume and pressure of the headspace and the dry weight of the soil core in the serum bottle. The dry weight of the sediment was determined gravimetrically at the end of the incubation by drying at 108°C for 24 h. The sediment typically contained 0.5 g water


H20 Sediment B


Septum--~ " QIP lOcm


cm Fig. 2. Devices for the determination of the rate of'methanegenesis in paddy fields. (A) Position ~of closed corers in !he submerged paddy soft. (B) Dimensions of the glass corers closed on top with a septum. (C) Schema of the procedure for extracting CH 4 from the paddy sediment core.

346 the sediment core was extracted into the headspace by stirring the slurry for approx. 15 min. Due to the low solubility coefficient of C H 4 in water, more than 97% of the dissolved C H 4 is extracted into the gas phase. The amount of C H 4 extracted was calculated from the mixing ratio of C H 4 and the volume of the headspace. Tests showed that the time was sufficient for complete extraction and that no significant amounts of C H 4 were produced or consumed during the extraction procedure. Rates of methanogenesis were determined from the temporal increase of the C H 4 content in the soil cores by using at least 15 extractions during a total in situ incubation period of 4 - 6 h. The lower detection limit of the rate of methanogenesis in a core below a unit soil surface area was approx. 10 m g . m - 2 . h -1, the precision of the determination was < ___80%.

3. 6. Sampling of gas bubbles Gas bubbles present in the submerged soil of the paddy field were collected by placing a waterfilled funnel over the soil surface and stirring the submerged soil to force ebullition. The trapped gas bubbles were retrieved with an air-tight syringe through a septum fixed at the upper end of the funnel and analyzed immediately after sampling. Gas bubbles released from the submerged soil under undisturbed conditions were trapped by installing the funnel for several hours until sufficient gas had accumulated. Gas samples were also taken from the plant intracellular gas space system by inserting a syringe with a thin needle into the stem of the rice plant below the water surface and removing approx. 50-100 /~1 gas from the medullary cavity. 3. 7. Measurement of CH 4 consumption Sediment samples (approx. 10 ml) were taken from the paddy field and diluted with 40 ml of the flooding water. The slurry, which had an 02 concentration of approx. 8 ppm, was transferred into a glass syringe (50 ml) without gas bubbles. Then 1 ml of water containing dissolved C H 4 w a s injected and the syringe incubated at in situ tempeature (25-30°C). Subsamples (2 ml) of the slurry were taken at intervals using a 10-ml syringe. 8 ml CHa-free air was sucked into the syringe and the

dissolved C H 4 w a s extracted by heavy shaking. The amount of C H 4 dissolved in the sediment slurry was calculated from the mixing ratio of C H 4 and the volumes of the gas and aqueous phases in the syringe.

3.8. Gas analysis C H 4 was analysed in a gas chromatograph (GC-mini, Shimadzu) equipped with a flame ionization detector. The gas samples were injected through a sampling loop and separated on a molecular sieve column (13 × 6 0 / 8 0 mesh) 100 cm long at 70°C. The lower detection limit was approx. 3 × 10 - 3 p p m C H 4. CO 2 was analyzed by means of an infrared analyzer type U R A S (Maihak, Hamburg). Simultaneous presence of C H 4 did not interfere with the measurements of CO 2 within the actually observed range of mixing ratios. The concentration of 02 in water was determined with a polarographic 02 probe (Syland Scientific, Heppenheim).

4. R E S U L T S

4.1. Importance of C H 4 oxidation Rice cultures in beakers with submerged soil were incubated for 20 h in closed vessels containing air, oxygen or nitrogen gas atmospheres (Fig. 1). After renewal of the gas atmosphere, the rates of C H 4 emission were determined from the linear increase of the C H 4 mixing ratio in the incubation vessel during 30 min of incubation. The C H 4 emission rates measured by this procedure were due to plant-mediated transport and were not significantly influenced by bubble release or diffusional transport of C H 4 through the soil-water interface. The results are summarized in Table 1 and are given as percent of a control kept under ambient air. Relative to the control, incubation under pure 02 resulted in lower C H 4 emission rates, whereas incubation under pure N 2 resulted in higher C H 4 emission rates. These observations are explained by inhibition of methanogenesis or stimulation of methane oxidation under pure oxygen and stimulation of methanogenesis or inhibition of C H 4 oxidation under pure nitrogen a t m o s p h e r e s , respectively. Addition of 5%

347 Table 1 Influence of oxidation reactions on the flux of methane from rice paddy soil into the atmosphere Incubation of laboratory rice cultures in submerged soil

CH 4 emission (% of control under air) 0 2


20 h Under 02 or N2 Addition of 5% C2H 2 and 1 h incubation





acetylene to the atmosphere of the incubation vessel resulted in partial recovery of the O2-inhibited rate of CH 4 emission after 1 h of incubation, but did not affect the N2-stimulated rate of CH 4 emission. Since acetylene is a potential inhibitor of methane oxidation [22], the effects of the O 2 and N 2 atmospheres on the CH 4 emission were probably due to changes in the CH 4 oxidation process. The relative importance of CH 4 oxidation for the turnover of CH 4 in paddy soil was assessed by comparing the rates of methanogenesis and CH 4 emission (Table 2). The results show that the CH 4 emission rates accounted for o n l y 12-25% of the C H 4 production rate. The rest, about 80%, of the methane produced must have been oxidized before its release into the atmosphere. To determine the importance of CH 4 oxidation in rice paddies under field conditions, the rates of C H 4 emission were compared to the rates of methanogenesis per unit area. Methanogenesis was measured in the field by inserting dosed corers into the paddy soil (Fig. 2). The inserted corers did

not interfere with the in situ production of CH 4, but interrupted the escape of methane by ebullition or plant-mediated transport. Fig. 3 shows the results of an incubation experiment. The data show an increase in the total CH 4 content in the submerged soil with incubation time. Table 3 shows that only 60-90% of the CH 4 produced in the submerged paddy soil was released into the atmosphere. A further indication of the importance of CH 4 oxidation for the turnover of C H 4 in rice paddies was provided by the analysis of gas bubbles (Table 4). Gas bubbles, sampled directly from the submerged paddy soil, showed relatively large mixing ratios of CH 4 and CO 2. However, in the gas released by natural ebullition, the mixing ratios were 2 orders of magnitude lower. The low CO 2 mixing ratios may be due to dissolution into the water during the trapping period of several hours. However, this explanation does not apply for CH 4. Control experiments with CH 4 being added to the trapped ebullition gas to give a final mixing ratio of approx. 60% showed only an insignificant decrease in C H 4 with time, so that dissolution and oxidation of CH 4 in the water can be excluded. Therefore, the low CH 4 mixing ratios in ebullition gas must be due to the oxidation of CH 4 during transport out of the submerged soil. The same conclusion can be drawn from the low concentrations of CH 4 in the intercellular gas space of the rice plants. Whereas CO 2 mixing ratios were only half those found in the gas bubbles in the sediment, CH 4 mixing ratios were 50 times lower. 0







Table 2 Rates of methanogenesis and of CH4 emission in laboratory cultures of paddy soil planted with rice a Experiment CH4 production CH4 emission Percentage (/xg.h -m) (/Lg.h-1) emitted 1 2 3 Mean

462+ 29 197 244 ± 65 301 + 141

54 48 + 13 54 52+ 3

12 25 22 20 + 7

a Measurements were carried out on 2-3 repficate cultures incubated at 25°C in the light. The rates are given as means + SD in #g CH4 per culture per h.


16 Jul 84


"E /.OC E "0

r=0.49 P< 5 %





20C o


0 Time [ h ]

Fig. 3. Temporal increase in CH4 content in soil cores in situ.

348 Table 3


Production rate (mg. m - 2. h - 1)

Emission rate (mg. m - 2. h - l)

Percentage emitted

16 July 84 18 July 84 18 A u g 84 19Aug84

28 34 25 18

25 _4-15 305:9 15 5 : 1 125:2

91 90 60 66

Vercelli 21 Aug 1984

Oxygenated slurries of paddy sediment

Rates of methanogenesis and of methane emission in a rice paddy under field conditions a

1.5 J O-10cm depth %_


a Rates of CH 4 production were determined from the increase in C H 4 in cores of submerged paddy soil incubated in 15 replicates for up to 5 - 6 h under in situ conditions. The rates were calculated from the linear regression of the CH 4 contents with the incubation period (compare Fig. 3). Rates of C H 4 emission were determined using the static box technique and are given as means _+SD from 3 replicate measurements.

The direct demonstration of the existence of C H 4 oxidation activities could only be achieved in samples from the oxidized zone of the top 1 cm of the submerged paddy soil. When these samples were incubated as slurries containing dissolved 0 2 and C H 4 (Fig. 4), the CH 4 concentration in water decreased with time. CH 4 was not consumed, however, and small amounts were even produced when slurries were prepared from 10-cm-deep soil cores containing the anoxic part of the soil profile and parts of the rhizosphere. There were apparently enough anoxic particles in these samples to permit net CH 4 production, although oxygenated conditions had been established in the soil slurries. It is very difficult to design a direct test for CH 4 oxida-" tion activity in the narrow oxic layer around the rice roots because of the very active CH 4 production in the surrounding anoxic soil.





16o Time [rain ]

Fig. 4. Temporal change in C H 4 dissolved in slurries of submerged paddy soil samples.

4.2. Influence of brackish flooding water To assess the influence of seawater salts on methanogenesis and CH 4 emission, laboratory experiments were carried out in which cultures with Italian paddy soil and rice plants were flooded i



Anoxic paddy soil planted with rice and flooded with

10(] ~


s o l u t i o n s of a r t i f i c i a l

Table 4 c-

Mixing ratios of CH 4 in gas bubbles and rice intercellular space in rice paddies a


Source of gas


Bubbles in submerged paddy soil G a s in plant intercellular space Ebullition gas

CH 4 (%)


35.6+ 19.4

6.8 + 5.1

0.6 + 0.3

3.3 + 0.7

0.2 + 0.3

0.04 + 0.01


4C U.l



Salt concentration in flooding water •

a M e a n + S D of 5 - 1 0 measurements.


O ,:ter"

~ ~ ,


10 20 30 Time of Incubation [days]


Fig. 5. Change in CH 4 emission rate of laboratory rice cultures after replacement of the supernatant flooding water with artificial sea water solutions.

349 Table 5 Influence of sea salt or sulfate on the production and emission of methane from laboratory cultures of flooded paddy soil planted with rice a Experiment

CH 4 production (/~g.h -1)

% Of control

Control 0.15% Salt b (0.01% Sulfate)



0.30% Salt b (0.02% sulfate) 0.10% Sulfatee


95:7 45:2


4 2

CH4 emission (/Lg,h -1) 48 + 13 15:2

< 0.2 < 0.2

% Of control 100 3

<1 <1

a Measurements were carried out with 3 replicate cultures.

Rates are means+SD given as t~g C H 4 per culture per h after 100 days of incubation. b Paddy soil was submerged in artificial seawater solution. The sea-salt contained approx. 7% sulfate. Values in brackets give the sulfate concentration. c Paddy soil was submerged in Na2SO 4 solution.

with diluted artificial sea water. Fig. 5 shows that replacement of the flooding water with diluted artificial sea water resulted in a significant decrease with time in the CH 4 emissionrates. Table 5 summarizes the rates of mcthanogenesis and CH4 emission measured in laboratory rice cultures prepared by submerging the paddy soil in solutions of sea-salt or Na2SO4. The results clearly show that even small amounts of salt reduced the rate of methanogenesis, and that after an incubation period of 100 days, CH 4 emission rates were negligible compared to the control kept under freshwater conditions. A solution of 0.3% sea-salt, which contained sulfate at a total concentration of 0.02%, had an effect similar to that of a solution of 0.1% Na2SO4.

5. DISCUSSION Measurements of the CH 4 emission rates from rice paddies in Andalusia (Spain)and Vercelli (Italy) showed large differences in the total amount of CH 4 emitted during the vegetation period ([18,21]; unpublished results). The CH 4 emission rates obtained from the Andalusian rice paddies were a factor of 3-5 lower than those in the Italian

rice paddies. It has been suggested that the lower CH 4 emission from Andalusian rive paddies may have been due to lower organic carbon content a n d / o r the quality of the flooding water [21]. The Andalusian rive paddies are flooded with water from the Guadalquivir river, which is influenced by the inflow of Atlantic sea water through tidal movement. Hence, the flooding water has a relatively high content of sodium chloride and sulfate, with total salt concentrations varying between 0.6-3.5 g.1-1. Our experiments with artificial seawater or sulfate solutions indicate that the relatively low amounts of CH4 emitted from Andalusian rice paddies are due to inhibition of methanogenesis by the sulfate-containing sea-salt: It is a well-known phenomenon that methanogenesis in freshwater sediments is inhibited by the addition of sulfate [23-25]. This inhibition by sulfate has also been observed in flooded anoxic soils [26,27] and is confirmed by our experiments. Jakobsen et al. [26] reported a 50% decrease in methane yield when the soil was treated with 0.6%0 sulfate solution. Our studies even indicate a reduction, of 95% with only 0.17oo sulfate, which, however, was dissolved in diluted seawater. It is most likely that the sulfate present in the salt solution was inhibitory to methanogenesis. However, it is also possible that other ions present in the seawater or the salinity o f the solution had an additional inhibitory effect. This conclusion is in agreement with observations in Senegal rive paddies [28] where the number and activity of methanogenic bacteria were negatively correlated with the chloride content, and with measurements in coast wetlands [27] where CH 4 emission rates were negatively correlated with sulfate content and salinity. Although these results do not provide a physiological explanation of the inhibitory effect of salt and sulfate on methanogenesis in anoxic soils, they explain the observation that less methane is emitted into the atmosphere when rive paddie~ are flooded with brackish water. Methane emission from rice paddies is a function not only of the rate of methanogenesis but also of the rate of methane oxidation. There is evidence that methane is oxidized in paddy soil, so that a significant part of the methane produced is not emitted. This conclusion is drawn from results


of laboratory and field experiments showing: (1) C H 4 consumption in slurries of the oxic surface layer of submerged paddy soil; (2) inhibition of methane emission by an oxygen atmosphere and stimulation by a nitrogen atmosphere or an atmosphere containing acetylene; (3) significantly lower rates of C H 4 emission than of C H 4 production, both in the laboratory and in the field; and (4) lower mixing ratios of CH 4 in released gas bubbles and in gas sampled from the plants' intercellular gas space than in gas bubbles taken from the submerged paddy soil. C H 4 oxidation in the oxic surface layers of the paddy soil has already been demonstrated by De Bont et al. [15] by inhibiting c n 4 oxidation with acetylene. These authors also observed increased population densities of methanotrophic bacteria in the oxic surface layer, and also in the rhizosphere, indicating that C H 4 oxidation should take place at both locations. However, they did not observe a stimulation of C H 4 emission when they introduced 10% acetylene in a bag enclosing the phyllosphere of rice, and therefore concluded that C H 4 oxidation in the rhizosphere is not significant. In contrast to their results, we observed a rapid stimulation of the CH 4 emission from vegetated paddy soil after the addition of acetylene under conditions where C H 4 emission was almost exclusively due to plant-mediated transport. Consequently, we believe that the rhizosphere plays a significant role in C H 4 oxidation. Our conclusion is also in agreement with the observed imbalance of CH 4 production rates in the submerged soil and C H 4 emission rates into the atmosphere. Since the majority ( > 90%) of CH 4 emission proceeds via plant-mediated transport [18,19], the imbalance between C H 4 production rates and CH 4 emission rates must be due t o c n 4 oxidation occurring in the rhizosphere. The assumption of CH 4 oxidation processes in the rhizosphere is also supported by the observation that the CH 4 mixing ratios in gas from the plants' medullary cavity showed similar low values to those measured in gas bubbles which had passed the oxic surface layer of the paddy soil. Hence, we conclude that some of the CH 4 is oxidized by methanotrophic bacteria whenever it passes through an oxic zone, i.e., the oxic paddy surface layer or the oxic zone surrounding the rice roots.

At present, the data base is too limited to give a reliable quantitative estimate of the fraction of the CH_..4A. produced which is either released into the atmosphere or oxidized in the submerged soils. This applies particularly under field conditions, since the measurement of c n 4 production by inserting corers into the paddy soil resulted in a high variability of the data points, most probably due to spatial inhomogeneities of the field site. Hence, the precision of the in situ rates of methanogenesis determined from these data is very low. However, these figures represent the lower limits of the rates of methanogenesis, since it cannot be excluded that some of the C H 4 produced was oxidized within the soil corers due to transport of CH 4 bubbles into the oxic surface soil layers. Nevertheless, the balance between C H 4 production and emission clearly indicates that CH 4 oxidation reactions are important under field conditions and may account for > 10-40% of the C H 4 produced. More measurements of in situ rates of methanogenesis are needed to obtain a better knowledge of the temporal and spatial distribution of C H 4 production rates. Information on the distribution and activity of methanogenic and methanotrophic bacteria in paddy soil is also lacking. This knowledge will be necessary to explain the emission of C H 4 into the atmosphere, in particular the diurnal and seasonal changes of C H 4 emission rates from rice paddies, which have a significant impact on atmospheric chemistry and climate [29].

ACKNOWLEDGEMENTS We are grataeful to the Istituto Sperimentale per la Risicoltura (Dr. S. Russo) in Vercelli, Italy, for permitting and supporting our experiments in their rice paddies. We thank Carsten K6bbemann for technical assistance. This study was financially supported by the Bundesministerium for Forschung und Technologie (KBF 68). REFERENCES [1] Watanabe, I. and Furusaka, C. (1980) Adv. Microb. Ecol. 4, 125-168.

351 [2] Freney, J.R., Jacq, V.A. and Baldensberger, J.F. (1982) Dev. Plant Soil Sei. 5, 271-317. [3] Reddy, K.R. and Patrick, W.H. (1984) CRC Crit. Rev. Environ. Control 13, 273-309. [4] Nedwell, D.B. (1984) Adv. Microb. Ecol. 7, 93-131. [5] Zeikus, J.G. and Winfrey, M.R. (1976) Appl. Environ. Microbiol. 31, 99-107. [6] Jones, J.G. and Simon, B.M. (1981) J. Gen. Microbioi. 123, 297-312. [7] Jones, J.G. (1980) J. Gen. Microbiol. 117, 285-292. [8] Koyama, T. (1963) J. Geophys. Res. 68, 3971-3973. [9] Koyama, T. (1964) in Recent Researches in the Fields of Hydrosphere, Atmosphere and Nuclear Geochemistry (Miyake, Y. and Koyama, T., Eds.), pp. 143-177. Maruzen, Tokyo. [10] Yamane, I. and Sato, K. (1963) Soil Sci. Plant Nutr. 9, 32-36. [11] Yamane, I. and Sato, K. (1964) Soil Sci. Plant Nutr. 10, 127-133. [12] Yamane, I. and Sato, K. (1967) Soil Sci. Plant Nutr. 13, 94-100. [13] Takai, Y. (1970) Soil Sci. Plant Nutr. 16, 238-244. [14] Raimbauit, M., Rinaudo, G., Garcia, J.L. and Borean, M. (1977) Soil Biol. Biochem. 9, 193-196. [15] De Bont, J.A.M., Lee, K.K. and Bouidin, D.F. (1978) Ecol. Bull. 26, 91-96. [16] Van Raalte, M.H. (1941) Ann. Bot. Gard. Buitenzorg 51, 43-57.

[17] Barber, D.A., Ebert, M. and Evans, N.T.S. (1962) J. Exp. Bot. 13, 397-403. [18] Seiler, W., Holzapfel-Pschorn, A., Conrad, R. and Scharffe, D. (1984) J. Atm. Chem. 1,241-268. [19] HolzapfeI-Pschorn, A., Conrad, R. and Seiler, W. (1986) Plant Soil, in press. [20] Cicerone, R.J., Shetter, J.D. and Delwiche, C.C. (1983) J. Geophys. Res. 88, 11022-11024. [21] Holzapfel-Pschorn, A. and Seiler, W. (1986) J. Geophys. Res., in press. [22] De Bont, J.A.M. and Mulder, E.G. (1976) Appl. Environ. Microbiol. 31,640-647. [23] Winfrey, M.R. and Zeikus, J.G. (1977) Appi. Environ. Microbiol. 33, 275-281. [24] Abram, J.W. and Nedwell, D.B. (1978) Arch. Microbiol. 117, 89-92. [25] Lovley, D.R., Dwyer, D.F. and Klug, M.J. (1982) Appl. Environ. Microbioi. 43, 1373-1379. [26] Jacobsen, P., Patrick, Jr., W.H. and Williams, B.G. (1981) Soil Sci. 132, 279-287. [27] DeLaune, R.D., Smith, C.J. and Patrick, Jr., W.H. (1983) Teilus 35B, 8-15. [28] Garcia, J.L., Raimbault, M., Jacq, V., Rinaudo, G. and Roger, P. (1974) Rev. Ecol. Biol. Soil 11, 169-185. [29] Seiler, W. (1984) in Current Perspectives in Microbial Ecology (Khig, M.J. and Reddy, C.A., Eds.) pp. 468-477. ASM, Washington, DC.