- Email: [email protected]
Soil bacteria in land-drainage water
Water Research Pergamon Press 1973. Vol. 7, pp. 1295-1300. Printed in Great Britain
SOIL
BACTERIA
IN
LAND-DRAINAGE
WATER
M. R. EVANS and J. D. OWENS* Department of Microbiology, The West of Scotland Agricultural College, Auchincruive, Ayr, Scotland
(Receiced 15 February 1973) Abstract--The rate of discharge and the concentration of viable soil bacteria in the water from a subsurface field drain were monitored for 4 months during Winter. The concentration of bacteria in the water was related to the flow rate of the discharge by an equation of the form: log bacterial concentration = a + b log flow rate + c (log flow rate),'- where a, b and c are constants. The total number of viable bacteria in the drainage water discharged during the 4 months represented approximately 0-1 per cent of an estimate of the total number of viable bacteria present in the soil of the experimental plot. It was concluded that the numbers of bacteria lost from soil by wash-out in drainage water were an insignificant fraction of the probable annual production of bacteria in the soil.
INTRODUCTION
DURING a study into the factors affecting the concentration of faecal bacteria in subsurface drainage water from a pasture (EVANS and OWENS, 1972) it was noted that the water also contained relatively high numbers of bacteria presumed to be normal soil inhabitants. The presence of these bacteria in the drainage water raised the possibility that wash-out of natural soil bacteria in drainage water could represent a significant loss from the soil bacterial flora. In addition, it was of interest to see if the concentrations of normal soil bacteria in the drainage water varied with the flow rate of the drain discharge in a manner similar to that observed for faecal bacteria deposited onto the surface of the land. To investigate these questions the general viable count of bacteria in the discharge from the drain was monitored during Winter 1971/1972 and the total number of bacteria discharged was compared with an estimate of the number of bacteria in the soil of the drained area. No attempt was made to obtain maximum counts of bacteria. Only those bacteria able to produce colonies on plate count medium incubated at 20°C for 48 h were counted and it is assumed that the results for this group of bacteria are representative of those that would be obtained for other groups of soil bacteria.
METHODS
The experimental plot was the one used previously (EVANS and OWENS, 1972). However, during spring 197! it was ploughed and reseeded with clover. No animal excrement was applied to the plot between the time of reseeding and the completion of the experiments reported here. Records of the daily soil temperatures at depths o f I0 and 100 cm were obtained from Auchincruive meteorological station. The rate of discharge from the drain was monitored with a V notch weir gauge (EVANS and EDGAR, 1971). Because previous results showed that bacterial concentration in the water was greatly affected by the flow rate, water samples were collected * Present address: School of Biological Sciences, Universiti Sains Malaysia, Penang, Malaysia. 1295
M . R . EVANS and J. D. OWENS
1296
over a period of a few minutes rather than over 24 h. They were collected in sterile, evacuated Winchester bottles with an attached length of silcione rubber tubing, and the flow rate of the discharge at the time of sampling was noted. Usually two samples were collected each working day, and a total of 134 samples were obtained in the period of the experiment from 17 November 1971 to 24 March 1972. The numbers of viable bacteria in the water samples were determined by the membrane filtration technique, using m-plate count broth (Difco) and counting colonies after incubation at 20°C for 48 h. To obtain an estimate of the number of bacteria occurring in the soil and able to grow on plate count medium, soil samples were collected at three different depths from each of 5 areas of the experimental plot. The depths sampled were 0-10, 10-20 and 20-30 cm and each area sample was a composite sample from five sites within the area. Each composite sample was passed through a sieve (2 mm mesh) to remove gravel and then 20 g of soil and 180 ml of mineral base E (OweNs and KEDDIE, 1969) were blended in an Atomix blender (M.S.E. Ltd., 25-28 Buckingham Gate, London, S.W.I) at maximum speed for 2 rain. Appropriate dilutions of the blended material were spread over the surface of re-plate count broth medium solidified with agar, and after incubation at 20°C for 48 h colonies were counted. RESULTS
Relationship of concentration of bacteria in the drain discharge to f tow rate of the discharge The results are summarized in FIG. 1, from which it is evident that the concentration of bacteria in the drainage water increased as the flow rate of the discharge lO 9
:f
.,J
"..
i0 ~
in
,,Q D
o
107
8 "a
• ;..../..
o
10~
| a |... I ? .
°o
rn
0"0001
I o.ool
1 o.ol Flow to're,
I o.I
l i.o
L/s
FiG. 1. Curve s h o w i n g the relationship between concentration o f viable soil bacteria in the water and the flow rate o f the drain discharge.
Soil Bacteria in Land-Drainage Water
1297
increased. Regression analysis showed that log concentration of bacteria could be related to log flow rate by an equation of the form Y = a - - b X ÷ c X 2, where Y = log lo concentration of bacteria in the discharge, viable units 1-t ; .X"= log lo flow rate of the discharge, 1.s- 1 ; and a, b and c are constants. The equation giving the best fit was Y = 9.70 + 2.356(4- 0-187)X + 0.365(+ 0.040)X 2 the figures in parentheses being the standard errors of the regression coefficients. This equation explains 79 per cent of the variation observed in the concentration of viable bacteria in the discharge. N o improvement was obtained by including in the equation a time factor or a third degree term. The soil temperature showed some positive association with flow, but after allowing for this association, soil temperature showed no significant association with the concentration of viable bacteria in the drain discharge.
Comparison of total number of bacteria discharged in drainage water with total number in the soil of the experimental plot The total number of bacteria discharged in drainage water during the experimental period was calculated from the records of the flow rate of the drainage water using the regression equation given above. During the period of the experiment 3.6 x 105 I. of water containing 1"2 × l0 t`* viable bacteria were discharged from the drain. The results obtained on one occasion for the viable counts of bacteria in the soil are shown in TABLE 1. It can be seen that the numbers were similar in each of the different areas sampled and that the numbers at a depth of 20-30 cm were considerably lower than those at 0-10 cm. I f it is assumed that the numbers of bacteria in the soil are, in fact, uniform over the area of the experimental plot (0.7 ha) and that the numbers of bacteria at depths greater than 30 cm can be neglected, then the total number o f viable bacteria in the experimental plot is estimated to be of the order of 10 IT. TABLE 1. NUMBERS OF VIABLE BACTERIA IN THE SOIL OF DIFFERENT AREAS OF THE EXPERIMENTAL PLOT
Number of bacteria,* viable units x 10-~ g - ' in soil horizon Sampling area A B C D E Mean
0-10 cm
10-20 em
20--30 cm
23 22 23 22 33 24"6
23 18 14 12 /6 16-6
7-2 5"3 3-8 4-0 5-3 5"1
* B a c t e r i a g r o w i n g o n p l a t e c o u n t a g a r a t 2 0 ° C w i t h i n 48 h.
Thus the total number of viable bacteria in the soil able to grow on the medium used was about 1000 times the number discharged in the drainage water over a period of 4 months. W.R. 7/9--D
1298
M.R. EvA.xsand J. D. OwENs DISCUSSION
Relationship o f concentration o f bacteria to f l o w rate o [ the discharge
The concentration of viable bacteria, presumed to be normal soil inhabitants, in the drainage water was closely associated with the flow rate of the drain discharge. This may be interpreted as meaning that the rate at which bacteria adsorbed to soil particles become suspended in soil water is determined by the flow rate of water through the soil. Thus the higher the flow rate the higher the concentration of bacteria in suspension and the higher the concentration in subsurface drainage water. However, at flow rates below about 0.001 1. s- 1 the regression curve (FIG. 1) suggests that the concentration of bacteria in the water should increase with decreasing flow rate. Future work may confirm this suggestion, or may show it to be an artifact arising out of an insufficiency of data at low flow rates. In a previous study (EvANs and OWENS, 1972) it was noted that the concentration of faecal bacteria in drainage water was related not only to the flow rate of the discharge but also to time, and the occurrence of this time relationship was attributed to the numbers of viable faecal bacteria in the soil declining with passage of time. Evidently, in the present work the numbers of viable soil bacteria in the soil did not change significantly during the time of the experiment. Since they are presumed to be normal soil inhabitants this is, perhaps, as expected. The concentration of soil bacteria and the flow rate of the drain discharge were best related by an equation of the form Y -~ a -- b X ~, c X 2 ( Y = log bacterial concentration; X = log flow rate) whereas data on the concentration of faecal bacteria in drainage water best fitted an equation of the form Y = a ÷ bX, if the time relationship was ignored (EVANS and OWENS, 1972). These different equations probably do not reflect any real differences in the behaviour of the two groups of bacteria. It is possible that if each water sample used in the previous work had been collected in a few minutes instead of over 24 h and more samples examined, then the relationship of log concentration of faecal bacteria to log flow rate might also have been nonlinear. To facilitate comparison of the results on soil bacteria with those on faecal bacteria a linear log-log regression curve was fitted to the soil bacteria results. The equation Y = 8-01 -k 0.689(4- 0.043)Xexplains only 66 per cent of the variation in Y, but it is noteworthy, particularly since the faecal bacteria study was done 3 yr earlier than the soil bacteria study, that the regression coefficient on X is not significantly different from those observed for Escherichia coli (0.521) and enterococci (0-701). This would appear to support a previous suggestion (EVANSand OWENS, 1972) that the value of the regression coefficient is related to soil type. The fact that the soil bacteria data best fitted a curve described by a quadratic equation is interesting because BORMANN,LIKENSand EATON (1969) obtained a similar curve (FIG. 2) relating concentration of particulate matter (i.e. material passing through a 1 mm mesh net and retained on Millipore filters, pore size 0-45/zm) in a stream and the flow rate of the stream. To compare the curves, both have been plotted in FfG. 2 as concentration of particles against drainage rate (1. s- I h a - ~) rather than against flow rate (l. s- 1) because it seems likely that it is the rate of movement of water through soil that is important rather than the actual flow rate of the drain or stream discharge. As our results refer to viable soil bacteria in drainage water from 0.7 ha of farm land
Soil Bacteria in Land-Drainage Water
1299 -- 10 9
~
I0 e
10.0
o
/
°
.
g lot
I'0
2 c~
o w I 0 ~ ~.
0.1
I
0"01
0-0001
I
0"001
I
0"01
Dtoinocje
rate,
0'1
IO s
I I0
IO
l./s. ha
FiG. 2. Curves showing the relationship between drainage rate and the concentration of viable soil bacteria in land drainage water (continuous line; data from the present work) and the concentration of particulate matter in a stream (dashed line; data of BOR,~IANNet al., 1969). while the results of BORMANN et al. (1969) refer to particulate matter, presumably including inanimate material and viable bacteria, in a stream draining 13.23 ha of forest in New Hampshire, the degree of similarity in the shape of the curves is surprising. However, both curves relate to mainly subsurface drainage water. This raises the possibility that the relationship observed between the concentration of small particles in subsurface drainage water and flow rate may be a general one, at least over the range of drainage rates investigated. That this relationship is different from the linear log-log relationship generally observed between suspended solids load and flow rate of streams and rivers (HOAK and BRAMER, 1956; LEOPOLDand MILLER, 1956; LEOPOLD, WOLMAN and MILLER, 1964) can possibly be attributed to the different nature of the particles examined. River and stream suspended solids, particularly at high flow rates, include particles that will sediment rapidly under quiescent conditions whereas the particulate matter examined by BORMANN et al. (1969) and the bacterial cells studied in the present work sediment only very slowly. BORMANN et al. (1969) suggested that the shape of the curve in FIG. 2 (i.e. the values of coetficients b and c) reflects both the capacity of water to do work as its velocity increases, and the relative ease or difficulty with which moving water of a given velocity can remove material from an ecosystem. This latter factor they called the erodibility of the ecosystem, a term that could also be used to describe the relative ease or difficulty with which bacteria adsorbed to soil particles are removed by moving water. Thus the fact that the rate of change in concentration o f particulate matter with an increase in drainage rate is less than the rate of change in concentration of viable bacteria for a similar increase in drainage rate (FIG. 2), suggests that the bacteria were more erodible than the particulate matter. Apart from the influence of the type of particle, the erodibility of a component of an ecosystem would be affected by many other factors, including soil type, degree of aggregation of soil particles, and type and amount of vegetation (BORMANNet al., 1969).
1300
M.R. EvAys and J. D. OwE,',-s
Soil bacterial flora and losses in drainage water
The bacteria f o u n d in the drainage water are believed to be representatives o f natural soil bacteria rather than o f an intrinsic drain water flora or other habitat for two reasons. Firstly, it seems likely that some soil bacteria would find their way into drainage water since previous work (EVANS and OWEYS, 1972) showed that faecal bacteria deposited initially on the surface o f the land appeared in the soil drainage water. Secondly, the short time interval between the application o f large volumes o f animal excrement to the land and the appearance o f high concentrations o f faecal bacteria in the drainage water (EVANS and OWENS, 1972) indicates that the mean residence time o f the water in the drain is too short to allow much growth o f bacteria in the drain water. However, the possibility that some o f the bacteria in the drainage water originated f r o m the walls o f the drains or from vegetation on the surface o f the land c a n n o t be excluded. Nevertheless, if it is accepted that most o f the bacteria in the drainage water came from the soil, then the total n u m b e r o f viable soil bacteria in the drainage water discharged over a period o f 4 m o n t h s represented only about 0.1 per cent o f an estimate o f the total n u m b e r o f viable bacteria of the same type in the soil o f the experimental plot. Since water does not usually flow f r o m the drain for more than 6 months of the year, the m a x i m u m annual loss o f bacteria from the soil by this route would be only 0-15 per cent o f the total soil population. Even if the growth rate of bacteria in the soil was as low as the lowest estimates (GRAY and WILLIAMS, 1971) this loss is an insignificant fraction o f the annual production o f soil bacteria. However, this conclusion does not exclude the possibility that loss o f bacteria by wash-out in drainage water could be i m p o r t a n t in some soil types or climatic conditions. Acknowledgements--We thank MRS ELAINEADAIR for technical assistance and the Agricultural Research Council Unit of Statistics, University of Edinburgh for statistical analyses. This work forms part of an investigation into the treatment and disposal of farm wastes supported by the Agricultural Research Council.
REFERENCES ]3ORMANNF. H., LIKENSE. and EATONJ. S. (1969) Biotec regulation of particulate and solution losses from a forest ecosystem. Bioscience 19, 600-610. EVANS M. R. and EDGAR R. (1971) Automatic sampling and measurement of small liquid flows. Water Pollut. Con. 70, 111-113. EVANSM. R. and OWENSJ. D. (1972) Factors affecting the concentration of faecal bacteria in landdrainage water. J. gen. Microbiol. 71,477--485. GRAY T. R. G. and WtLLIAMSS. T. (1971) Microbiol productivity in soil. In: Microbes and Biological ProdactiviO,, (Edited by D. E. HUGHESand A. H. RosE), pp. 255-286. Society for General Microbiology, Cambridge. HOAK R. D. and BRAMERH. C. (1956) Natural sediment as a factor in stream pollution control. Sewage ind. Wastes 28, 311-322. LEOPOLDL. B. and MtLLERJ. P. (1956) Ephemeral streams: hydraulic factors and their relation to the drainage net. Geological Survey Professional Paper 282-A. U.S. Government Printing Office, Washington. LEOPOLDL. B., WOLMANM. G. and MILLERJ. P. (1964) Fluv. proc. in Geomorph. pp. 180--184. W. H. Freeman, San Francisco. OWENSJ. D. and KEDDtER. M. (1969) The nitrogen nutrition of soil and herbage coryneform bacteria. J. appl. Bacteriol. 32, 338-347.
SOIL
BACTERIA
IN
LAND-DRAINAGE
WATER
M. R. EVANS and J. D. OWENS* Department of Microbiology, The West of Scotland Agricultural College, Auchincruive, Ayr, Scotland
(Receiced 15 February 1973) Abstract--The rate of discharge and the concentration of viable soil bacteria in the water from a subsurface field drain were monitored for 4 months during Winter. The concentration of bacteria in the water was related to the flow rate of the discharge by an equation of the form: log bacterial concentration = a + b log flow rate + c (log flow rate),'- where a, b and c are constants. The total number of viable bacteria in the drainage water discharged during the 4 months represented approximately 0-1 per cent of an estimate of the total number of viable bacteria present in the soil of the experimental plot. It was concluded that the numbers of bacteria lost from soil by wash-out in drainage water were an insignificant fraction of the probable annual production of bacteria in the soil.
INTRODUCTION
DURING a study into the factors affecting the concentration of faecal bacteria in subsurface drainage water from a pasture (EVANS and OWENS, 1972) it was noted that the water also contained relatively high numbers of bacteria presumed to be normal soil inhabitants. The presence of these bacteria in the drainage water raised the possibility that wash-out of natural soil bacteria in drainage water could represent a significant loss from the soil bacterial flora. In addition, it was of interest to see if the concentrations of normal soil bacteria in the drainage water varied with the flow rate of the drain discharge in a manner similar to that observed for faecal bacteria deposited onto the surface of the land. To investigate these questions the general viable count of bacteria in the discharge from the drain was monitored during Winter 1971/1972 and the total number of bacteria discharged was compared with an estimate of the number of bacteria in the soil of the drained area. No attempt was made to obtain maximum counts of bacteria. Only those bacteria able to produce colonies on plate count medium incubated at 20°C for 48 h were counted and it is assumed that the results for this group of bacteria are representative of those that would be obtained for other groups of soil bacteria.
METHODS
The experimental plot was the one used previously (EVANS and OWENS, 1972). However, during spring 197! it was ploughed and reseeded with clover. No animal excrement was applied to the plot between the time of reseeding and the completion of the experiments reported here. Records of the daily soil temperatures at depths o f I0 and 100 cm were obtained from Auchincruive meteorological station. The rate of discharge from the drain was monitored with a V notch weir gauge (EVANS and EDGAR, 1971). Because previous results showed that bacterial concentration in the water was greatly affected by the flow rate, water samples were collected * Present address: School of Biological Sciences, Universiti Sains Malaysia, Penang, Malaysia. 1295
M . R . EVANS and J. D. OWENS
1296
over a period of a few minutes rather than over 24 h. They were collected in sterile, evacuated Winchester bottles with an attached length of silcione rubber tubing, and the flow rate of the discharge at the time of sampling was noted. Usually two samples were collected each working day, and a total of 134 samples were obtained in the period of the experiment from 17 November 1971 to 24 March 1972. The numbers of viable bacteria in the water samples were determined by the membrane filtration technique, using m-plate count broth (Difco) and counting colonies after incubation at 20°C for 48 h. To obtain an estimate of the number of bacteria occurring in the soil and able to grow on plate count medium, soil samples were collected at three different depths from each of 5 areas of the experimental plot. The depths sampled were 0-10, 10-20 and 20-30 cm and each area sample was a composite sample from five sites within the area. Each composite sample was passed through a sieve (2 mm mesh) to remove gravel and then 20 g of soil and 180 ml of mineral base E (OweNs and KEDDIE, 1969) were blended in an Atomix blender (M.S.E. Ltd., 25-28 Buckingham Gate, London, S.W.I) at maximum speed for 2 rain. Appropriate dilutions of the blended material were spread over the surface of re-plate count broth medium solidified with agar, and after incubation at 20°C for 48 h colonies were counted. RESULTS
Relationship of concentration of bacteria in the drain discharge to f tow rate of the discharge The results are summarized in FIG. 1, from which it is evident that the concentration of bacteria in the drainage water increased as the flow rate of the discharge lO 9
:f
.,J
"..
i0 ~
in
,,Q D
o
107
8 "a
• ;..../..
o
10~
| a |... I ? .
°o
rn
0"0001
I o.ool
1 o.ol Flow to're,
I o.I
l i.o
L/s
FiG. 1. Curve s h o w i n g the relationship between concentration o f viable soil bacteria in the water and the flow rate o f the drain discharge.
Soil Bacteria in Land-Drainage Water
1297
increased. Regression analysis showed that log concentration of bacteria could be related to log flow rate by an equation of the form Y = a - - b X ÷ c X 2, where Y = log lo concentration of bacteria in the discharge, viable units 1-t ; .X"= log lo flow rate of the discharge, 1.s- 1 ; and a, b and c are constants. The equation giving the best fit was Y = 9.70 + 2.356(4- 0-187)X + 0.365(+ 0.040)X 2 the figures in parentheses being the standard errors of the regression coefficients. This equation explains 79 per cent of the variation observed in the concentration of viable bacteria in the discharge. N o improvement was obtained by including in the equation a time factor or a third degree term. The soil temperature showed some positive association with flow, but after allowing for this association, soil temperature showed no significant association with the concentration of viable bacteria in the drain discharge.
Comparison of total number of bacteria discharged in drainage water with total number in the soil of the experimental plot The total number of bacteria discharged in drainage water during the experimental period was calculated from the records of the flow rate of the drainage water using the regression equation given above. During the period of the experiment 3.6 x 105 I. of water containing 1"2 × l0 t`* viable bacteria were discharged from the drain. The results obtained on one occasion for the viable counts of bacteria in the soil are shown in TABLE 1. It can be seen that the numbers were similar in each of the different areas sampled and that the numbers at a depth of 20-30 cm were considerably lower than those at 0-10 cm. I f it is assumed that the numbers of bacteria in the soil are, in fact, uniform over the area of the experimental plot (0.7 ha) and that the numbers of bacteria at depths greater than 30 cm can be neglected, then the total number o f viable bacteria in the experimental plot is estimated to be of the order of 10 IT. TABLE 1. NUMBERS OF VIABLE BACTERIA IN THE SOIL OF DIFFERENT AREAS OF THE EXPERIMENTAL PLOT
Number of bacteria,* viable units x 10-~ g - ' in soil horizon Sampling area A B C D E Mean
0-10 cm
10-20 em
20--30 cm
23 22 23 22 33 24"6
23 18 14 12 /6 16-6
7-2 5"3 3-8 4-0 5-3 5"1
* B a c t e r i a g r o w i n g o n p l a t e c o u n t a g a r a t 2 0 ° C w i t h i n 48 h.
Thus the total number of viable bacteria in the soil able to grow on the medium used was about 1000 times the number discharged in the drainage water over a period of 4 months. W.R. 7/9--D
1298
M.R. EvA.xsand J. D. OwENs DISCUSSION
Relationship o f concentration o f bacteria to f l o w rate o [ the discharge
The concentration of viable bacteria, presumed to be normal soil inhabitants, in the drainage water was closely associated with the flow rate of the drain discharge. This may be interpreted as meaning that the rate at which bacteria adsorbed to soil particles become suspended in soil water is determined by the flow rate of water through the soil. Thus the higher the flow rate the higher the concentration of bacteria in suspension and the higher the concentration in subsurface drainage water. However, at flow rates below about 0.001 1. s- 1 the regression curve (FIG. 1) suggests that the concentration of bacteria in the water should increase with decreasing flow rate. Future work may confirm this suggestion, or may show it to be an artifact arising out of an insufficiency of data at low flow rates. In a previous study (EvANs and OWENS, 1972) it was noted that the concentration of faecal bacteria in drainage water was related not only to the flow rate of the discharge but also to time, and the occurrence of this time relationship was attributed to the numbers of viable faecal bacteria in the soil declining with passage of time. Evidently, in the present work the numbers of viable soil bacteria in the soil did not change significantly during the time of the experiment. Since they are presumed to be normal soil inhabitants this is, perhaps, as expected. The concentration of soil bacteria and the flow rate of the drain discharge were best related by an equation of the form Y -~ a -- b X ~, c X 2 ( Y = log bacterial concentration; X = log flow rate) whereas data on the concentration of faecal bacteria in drainage water best fitted an equation of the form Y = a ÷ bX, if the time relationship was ignored (EVANS and OWENS, 1972). These different equations probably do not reflect any real differences in the behaviour of the two groups of bacteria. It is possible that if each water sample used in the previous work had been collected in a few minutes instead of over 24 h and more samples examined, then the relationship of log concentration of faecal bacteria to log flow rate might also have been nonlinear. To facilitate comparison of the results on soil bacteria with those on faecal bacteria a linear log-log regression curve was fitted to the soil bacteria results. The equation Y = 8-01 -k 0.689(4- 0.043)Xexplains only 66 per cent of the variation in Y, but it is noteworthy, particularly since the faecal bacteria study was done 3 yr earlier than the soil bacteria study, that the regression coefficient on X is not significantly different from those observed for Escherichia coli (0.521) and enterococci (0-701). This would appear to support a previous suggestion (EVANSand OWENS, 1972) that the value of the regression coefficient is related to soil type. The fact that the soil bacteria data best fitted a curve described by a quadratic equation is interesting because BORMANN,LIKENSand EATON (1969) obtained a similar curve (FIG. 2) relating concentration of particulate matter (i.e. material passing through a 1 mm mesh net and retained on Millipore filters, pore size 0-45/zm) in a stream and the flow rate of the stream. To compare the curves, both have been plotted in FfG. 2 as concentration of particles against drainage rate (1. s- I h a - ~) rather than against flow rate (l. s- 1) because it seems likely that it is the rate of movement of water through soil that is important rather than the actual flow rate of the drain or stream discharge. As our results refer to viable soil bacteria in drainage water from 0.7 ha of farm land
Soil Bacteria in Land-Drainage Water
1299 -- 10 9
~
I0 e
10.0
o
/
°
.
g lot
I'0
2 c~
o w I 0 ~ ~.
0.1
I
0"01
0-0001
I
0"001
I
0"01
Dtoinocje
rate,
0'1
IO s
I I0
IO
l./s. ha
FiG. 2. Curves showing the relationship between drainage rate and the concentration of viable soil bacteria in land drainage water (continuous line; data from the present work) and the concentration of particulate matter in a stream (dashed line; data of BOR,~IANNet al., 1969). while the results of BORMANN et al. (1969) refer to particulate matter, presumably including inanimate material and viable bacteria, in a stream draining 13.23 ha of forest in New Hampshire, the degree of similarity in the shape of the curves is surprising. However, both curves relate to mainly subsurface drainage water. This raises the possibility that the relationship observed between the concentration of small particles in subsurface drainage water and flow rate may be a general one, at least over the range of drainage rates investigated. That this relationship is different from the linear log-log relationship generally observed between suspended solids load and flow rate of streams and rivers (HOAK and BRAMER, 1956; LEOPOLDand MILLER, 1956; LEOPOLD, WOLMAN and MILLER, 1964) can possibly be attributed to the different nature of the particles examined. River and stream suspended solids, particularly at high flow rates, include particles that will sediment rapidly under quiescent conditions whereas the particulate matter examined by BORMANN et al. (1969) and the bacterial cells studied in the present work sediment only very slowly. BORMANN et al. (1969) suggested that the shape of the curve in FIG. 2 (i.e. the values of coetficients b and c) reflects both the capacity of water to do work as its velocity increases, and the relative ease or difficulty with which moving water of a given velocity can remove material from an ecosystem. This latter factor they called the erodibility of the ecosystem, a term that could also be used to describe the relative ease or difficulty with which bacteria adsorbed to soil particles are removed by moving water. Thus the fact that the rate of change in concentration o f particulate matter with an increase in drainage rate is less than the rate of change in concentration of viable bacteria for a similar increase in drainage rate (FIG. 2), suggests that the bacteria were more erodible than the particulate matter. Apart from the influence of the type of particle, the erodibility of a component of an ecosystem would be affected by many other factors, including soil type, degree of aggregation of soil particles, and type and amount of vegetation (BORMANNet al., 1969).
1300
M.R. EvAys and J. D. OwE,',-s
Soil bacterial flora and losses in drainage water
The bacteria f o u n d in the drainage water are believed to be representatives o f natural soil bacteria rather than o f an intrinsic drain water flora or other habitat for two reasons. Firstly, it seems likely that some soil bacteria would find their way into drainage water since previous work (EVANS and OWEYS, 1972) showed that faecal bacteria deposited initially on the surface o f the land appeared in the soil drainage water. Secondly, the short time interval between the application o f large volumes o f animal excrement to the land and the appearance o f high concentrations o f faecal bacteria in the drainage water (EVANS and OWENS, 1972) indicates that the mean residence time o f the water in the drain is too short to allow much growth o f bacteria in the drain water. However, the possibility that some o f the bacteria in the drainage water originated f r o m the walls o f the drains or from vegetation on the surface o f the land c a n n o t be excluded. Nevertheless, if it is accepted that most o f the bacteria in the drainage water came from the soil, then the total n u m b e r o f viable soil bacteria in the drainage water discharged over a period o f 4 m o n t h s represented only about 0.1 per cent o f an estimate o f the total n u m b e r o f viable bacteria of the same type in the soil o f the experimental plot. Since water does not usually flow f r o m the drain for more than 6 months of the year, the m a x i m u m annual loss o f bacteria from the soil by this route would be only 0-15 per cent o f the total soil population. Even if the growth rate of bacteria in the soil was as low as the lowest estimates (GRAY and WILLIAMS, 1971) this loss is an insignificant fraction o f the annual production o f soil bacteria. However, this conclusion does not exclude the possibility that loss o f bacteria by wash-out in drainage water could be i m p o r t a n t in some soil types or climatic conditions. Acknowledgements--We thank MRS ELAINEADAIR for technical assistance and the Agricultural Research Council Unit of Statistics, University of Edinburgh for statistical analyses. This work forms part of an investigation into the treatment and disposal of farm wastes supported by the Agricultural Research Council.
REFERENCES ]3ORMANNF. H., LIKENSE. and EATONJ. S. (1969) Biotec regulation of particulate and solution losses from a forest ecosystem. Bioscience 19, 600-610. EVANS M. R. and EDGAR R. (1971) Automatic sampling and measurement of small liquid flows. Water Pollut. Con. 70, 111-113. EVANSM. R. and OWENSJ. D. (1972) Factors affecting the concentration of faecal bacteria in landdrainage water. J. gen. Microbiol. 71,477--485. GRAY T. R. G. and WtLLIAMSS. T. (1971) Microbiol productivity in soil. In: Microbes and Biological ProdactiviO,, (Edited by D. E. HUGHESand A. H. RosE), pp. 255-286. Society for General Microbiology, Cambridge. HOAK R. D. and BRAMERH. C. (1956) Natural sediment as a factor in stream pollution control. Sewage ind. Wastes 28, 311-322. LEOPOLDL. B. and MtLLERJ. P. (1956) Ephemeral streams: hydraulic factors and their relation to the drainage net. Geological Survey Professional Paper 282-A. U.S. Government Printing Office, Washington. LEOPOLDL. B., WOLMANM. G. and MILLERJ. P. (1964) Fluv. proc. in Geomorph. pp. 180--184. W. H. Freeman, San Francisco. OWENSJ. D. and KEDDtER. M. (1969) The nitrogen nutrition of soil and herbage coryneform bacteria. J. appl. Bacteriol. 32, 338-347.