JOURNALOF BIOSCIENCEAND BIOENGINEERING Vol. 95, No. 6, 577-582. 2003

Competitive Growth of Gordonia and Acinetobacter in Continuous Flow Aerobic and Anaerobic/Aerobic Reactors

HYUNGJIN KIM’* AND KRISHNA R. PAGILLA2

Department of Environmental Engineering, &won Science College, Kyonggi 445-742, Korea’ and Department of Chemical and Environmental Engineering, Illinois Institute of Technology, Chicago, IL 60616, USA2

Received 18 October 2002IAccepted 4 February 2003

Gordonia amarae and Acinetobacter calcoaceticus pure culture experiments were conducted in continuous flow chemostats fed with an acetate/mineral salts medium under aerobic and anaero- bic-aerobic conditions. The growth parameters such as volatile suspended solids (VSS) suggest that Acinetobacter can successfully out-compete Gordonia under both aerobic and anaerobic- aerobic conditions when both are present in a dual culture. Gordonia was reduced to near detec- tion limits (~10 mg VSSIC), while Acinetobacter successfully grew to about 300 mg VSSII when both were grown as a dual culture in a single-stage aerobic chemostat with an initial concentra- tion of 200 mg VSSII each. In a two-stage anaerobic-aerobic chemostat with recycling, Gordonia was completely washed out in about 24 h, whereas, Acinetobacter successfully grew to over 170% of its initial concentration. In a single-stage aerobic chemostat, a mathematical model describing the competitive growth of the two species based on single culture kinetic parameters predicted Gordonia washout in ~30, and the same was observed from experiments. The model predicted an Acinetobacter concentration of 200% of its initial concentration in about 30 h, while experimen- tally this value was about 150%. Under anaerobic-aerobic conditions, the two-stage model pre- dicted Gordonia washout and an Acinetobacter concentration of about 170% of the initial concen- tration. The experimental results agreed with the prediction, but the decrease in the Gordonia and increase in the Acinetobacter concentrations were much more rapid than predicted by the model.

[Key words: Gordonia, Nocardia, Acinetobacter, enhanced biological phosphorus removal, chemostat, competition, modeling]

Phosphorus removal from wastewater is receiving in- creased attention from municipal wastewater treatment plants to protect recipient waters from nutrient overloading, leading to eutrophication. Many conventional secondary ac- tivated sludge plants are being retrofitted to include en- hanced nitrogen and phosphorus removal. In enhanced bio- logical phosphorus removal (EBPR), the activated sludge stores phosphate in excess of metabolic requirements, and can be removed from the wastewater by wasting the acti- vated sludge (1, 2). In such a process, solid-liquid separa- tion in the secondary clarifier becomes more critical be- cause any increase in effluent suspended solids results in in- creased effluent particulate phosphorus concentration. This paper describes the results from continuous flow chemostat experiments using two aerobic bacteria that are commonly found in activated sludge. One, a Gordonia sp. (formerly known as Nocardia amarae) that causes foaming problems due to its excessive growth in the activated sludge, and the other, an Acinetobacter sp. that can accumulate polyphos- phate and thereby remove phosphorus from wastewater (3, 4). The aim of the study is to determine steady-state biolog- ical conditions in a continuous flow activated sludge treat- ment system such that polyphosphate-accumulating Acine-

* Corresponding author. e-mail: kimhyunl@ssc.ac.kr phone: +82-3 I-350-2452 fax: +82-31-354-8987

tobacter sp. responsible for EBPR can out-compete tila- mentous Gordonia sp. that causes foaming problems.

Dual cultures competing for a single resource as the lim- iting substrate in continuous flow chemostat systems have been studied by several researchers (5-7). Veldkamp (5) found that two organisms could coexist in a chemostat sys- tem under steady-state conditions provided that there is a certain concentration at which the corresponding growth rates of both organisms are equal. If such a limiting sub- strate concentration can be achieved in the chemostat sys- tem, both organisms can coexist; otherwise the dominant organism with the higher specific growth rate will out-com- pete the species with the lower growth rate. Hao et al. (6) demonstrated that solid-liquid separation problems due to excessive growth of filamentous Sphaerotilus sp. in an acti- vated sludge wastewater treatment process at low dissolved oxygen (DO) conditions were due to the competitive advan- tage of Sphaerotilus sp. over floe-forming organisms. van Niekerk et al. (7) studied the competitive growth of the floe-forming bacterium Zoogloea ramigera and the filamen- tous bacterium Type 021N using acetate as the substrate under aerobic conditions. A mathematical model developed to describe dual species competition for a carbon substrate in a chemostat with an aerobic selector was able to predict dilution rates at which each species could dominate (8). Lau et al. (9) studied the competitive growth of the filamentous

577

578 KIM AND PAGILLA J. BIOSCI. BIOENG..

organism Sphaerotilus natans and the floe-forming orga- nism Citrobacter sp. using double Monod kinetics. Yoon and Blanch (10) stated that dual cultures competing for two limiting substrates such as organic carbon and oxygen can coexist if the microbial parameters such as specific growth rate and substrate affinity for competing organisms allow a steady state to be reached. Furthermore, there is a greater possibility of achieving steady state if the competing orga- nisms are limited by two growth-limiting substrates than if there is only one growth-limiting substrate.

In a properly functioning EBPR activated sludge system consisting of an anaerobic stage followed by an aerobic stage, excessive growth of filamentous bacteria in mixed- culture activated sludge can be controlled by Acinetobacter sp. due to its competitive edge in taking up carbon sub- strates under anaerobic conditions prior to the aerobic stage. It was found empirically in an activated sludge process that an anaerobic zone prior to the aeration basin selected out filamentous organisms responsible for solid-liquid separa- tion problems (11). Pitt and Jenkins (12) demonstrated that an anaerobic zone prior to an aerobic zone in the activated sludge system reduced the populations of Nocardia (Gor- donia) responsible for foaming. Pagilla et al. (13) found that Nocardia (Gordonia) control in activated sludge can be achieved by using an anaerobic selector prior to the aerobic zone. The mean cell residence time (MCRT), growth pH, anaerobic selector retention time, and presence of nitrates influence the functioning of EBPR and the Nocardia (Gor- donia) control. Hence, the objective of this research was to determine the competitive growth behavior of Gordonia and Acinetobacter under continuous flow anaerobic/aerobic se- quencing conditions, and develop a mathematical model to predict such competition.

MATERIALS AND METHODS

Culture and growth conditions Gordonia amarae strains isolated from foaming activated sludge (14), and Acinetobacter calcoaceticus strains obtained from the American Type Culture Collection (ATCC no. 3 10 12; Rockville, MD, USA) were grown in 1-Z Erlemneyer flasks provided with pure oxygen, and maintained in a temperature (25”C)-controlled room. The growth medium for all the experiments consisted of sodium acetate as the carbon source and a mineral salts medium as follows (per liter distilled water): NaCH,COO.3H,O, 5 g; KH,PO,, 0.15 g; NH,Cl, 0.5 g; MgS0,.7H,O, 0.05 g; FeCl,, 0.02 g; CaCl,, 0.02 g; and 20 ml of Hutner solution. One liter of the solution contains: nitrilotriacetic acid, log; (NH,),Mo,02,.4H,0, 0.01 g; FeS0,.7H,O, 0.1 g; and 50 ml of Metals 44 solution. One hundred ml of Metals 44 solution contains: EDTA, 250 mg; ZnS0,.7H,O, 1095 mg; FeS0,.7H,O, 500 mg; MnSO,.H,O, 154 mg; CuSO,. 5H,O, 39.2 mg; Ca(NO,),. 4H,O, 21.6 mg; and N$B,O,. lOH,O, 17.7 mg. The growth me- dium was sterilized in an autoclave at 121°C and 1.5 atm for 30 min.

After inoculation of 10 ml of culture inoculum into the flask containing the growth medium, the top of the flask was sealed with a cotton plug wrapped in cheese cloth, and then covered with aluminum foil. The inoculated flasks were placed on a magnetic stirrer set at 300 rpm (N,/O, supply rate = 1.5 Nmin) to keep the contents completely mixed and incubated at room temperature (25’C). Agar plates were streaked every 2 to 3 weeks and incu- bated at 35°C. The agar plate medium consisted of yeast extract

8 6

FIG. 1. Schematic diagram of the two-stage anaerobic-aerobic chemostat system: 1, feed medium bottle; 2, effluent bottle; 3, mag- netic stirrers; 4, anaerobic reactor; 5, aerobic reactor; 6, nitrogen cylin- der; 7, off-gas washing cylinder; 8, oxygen cylinder; 9, sampling sy- ringes; 10, peristaltic pumps.

(3 gll), peptone (5 g/l), and agar (15 g/l, maintained at pH 7.0. Continuous flow chemostat experiments Two different

chemostat systems with an inoculum culture of 10 ml were used in continuous flow pure culture experiments. A two-stage chemostat system consisting of two 0.5-Z Pyrex glass reactors with working volumes of 0.2 1 and 0.4 1 used as the anaerobic and aerobic stages, respectively, was employed. A l-l Pyrex glass reactor with a work- ing volume of 0.6 1 was used as the aerobic control chemostat. The general schematic layout of the two-stage chemostat system is pre- sented in Fig. 1. The anaerobic reactor of the two-stage system and the control were inoculated and operated in batch mode until sta- tionary phase biomass concentration was achieved. To minimize the effect of dissolved oxygen on the anaerobic reactor, the feed medium was not mixed and the recycle flow was taken from the subsurface of the culture in the aerobic reactor. A recycle flow of four times the feed rate (r = recycling flow rate/feed rate) was nec- essary to allow the biomass to grow under anaerobic/aerobic con- ditions.

Analytical methods Both Gordonia and Acinetobacter bio- mass concentrations were measured by a gravimetric method as volatile suspended solids (VSS), and carbon substrate was mea- sured as COD (15). To determine the relative mass fraction of Gordonia and Acinetobacter in dual culture competition experi- ments, a counting method developed by Pitt and Jenkins (12) was used. Samples of the cultures were Gram stained and counted under a microscope at 1000x magnification with oil immersion. The VSS of the Gordonia fraction in the dual culture was deter- mined by measuring the Gordonia filament count (12) and con- verting it to a mass concentration in terms of VSS. The relation- ship between Gordonia filament count and Gordonia mass for pure culture was found to be 5.37x lo4 intersectionsimg VSS. The linear relationship was developed with 5 replicate measurements for each sample over a range of 20 to 500 mg VSSIZ, and the correlation coefficient (R2) was 0.97, indicating a high degree of reliability. Acinetobacter VSS was determined from the difference between

COMPETlTlVE GROWTH OF GORDONIA AND ACINETOBACTER 579

Anaerobic Reactor Aerobic Reactor

v, xi, s

Effluent

Recycling Flow 1

FIG. 2. Variables involved in mass balance calculations for the two-stage chemostat system.

the total VSS of the dual culture and the Gordonia VSS. Acineto- batter was identified by microscopy using Niesser stained speci- mens. The pure cultures were microscopically examined periodi- cally for possible contamination using Gram staining and Niesser- stained specimens.

Mathematical model for a two-stage chemostat A mass balance on the anaerobic and aerobic reactors was conducted to es- tablish a mathematical formulation for competition in the two- stage chemostat system. The variables involved in mass balance are presented in Fig. 2.

Mass balance for the anaerobic reactor A mass balance on Acinetobacter biomass in the anaerobic reactor:

~=r.D,,.x,+Y”,.q,,.~~,-(l+r)D;~~,

Assuming no Acinetobacter growth in the anaerobic reactor:

Similarly, a mass balance on Gordonia biomass in the anaerobic reactor:

Assuming no Gordonia growth in the anaerobic reactor:

$$=r.D;.U,-(1 +r)D;X,,

A mass balance on the substrate in the anaerobic reactor assuming that only Acinetobacter can take up substrate under anaerobic con- ditions:

z =D;(S,+rS)-(1 +r)D;S,-qn,.X,,

Mass balance for the aerobic reactor A mass balance on Acinetobacter biomass in the aerobic reactor:

% =(l+r)D,[X,,-X,l+~,-k,,)X,

Assuming kd, is negligible and using the Monod model for growth:

$ =(I +r)D,[X”,-X,]+(e)X,

Similarly, a mass balance on Gordonia biomass in the aerobic re- actor:

A mass balance on substrate in the aerobic reactor for both species:

The above mathematical equations describe the transient behav- ior of two competing species in single and two-stage chemostats. The fourth order Runge-Kutta method (16) was used to numeri- cally solve the equations for each system after each time step. Since negative values of substrate and biomass concentrations are not acceptable, the numerical values of these concentrations should be checked after each time step. A time step of 0.005 d (7.2 min) was used in the numerical solution.

RESULTS AND DISCUSSION

Growth kinetics under aerobic and anaerobic-aerobic conditions The single-stage aerobic chemostat dilution rates ranged from 1.6 to 7.7 d-’ for Acinetobacter and 0.7 to 2.8 d-’ for Gordonia, and the two-stage anaerobic-aerobic chemostat dilution rates ranged from 1.3 to 3.5 d-’ for Acinetobacter and 0.7 to 1.9 d-’ for Gordonia. The steady- state biomass concentration and residual soluble substrate concentration at various dilution rates were determined after operation for three liquid retention times under steady-state conditions. Each data point obtained was the average of three sample measurements at steady state. The growth ki- netic parameters were determined by a Lineweaver-Burke plot of the Monod equation. The maximum specific growth rate @,,) and half saturation constant (KS) of Gordonia and Acinetobacter under different experimental conditions are summarized in Table 1. Acinetobacter has higher pmax (9.8 d-‘) and lower KS (5 10 mg COD/I) values compared with those of Gordonia (u,,,, 3.2 d-’ and KS, 820mgCODII) under aerobic conditions. These results indicate that Acine- tobacter can grow better than Gordonia under the same con- ditions due to its higher maximum specific growth rate, and higher substrate affinity as indicated by the lower KS value. Figure 3 shows the specific growth rate 01) as a function of the substrate concentration for Acinetobacter and Gordonia grown in aerobic chemostats. Each data point is an average of three measurements from three replicate experiments. It can be seen that Acinetobacter has a higher specific growth rate over the complete range of substrate concentration compared to that of Gordonia, and hence should washout Gordonia if both are competing for the same substrate in a dual or mixed culture. The aerobic yield coefficients were found to be 0.23 and 0.35 mg VSS/mg COD for Gordonia and Acinetobacter, respectively.

The Pnl,X values found during this work are comparable with those reported in the literature for both Gordonia (,u,,, range, 2.1 to 3 .O d-‘) (14, 17) and Acinetobacter (p,,, range,

TABLE 1. Growth kinetic parameters for Acinetohacter and Gordonia

Kinetic parameter Acinetobucter Gordoniu

Single-stage aerobic chemostat /(man cd-‘) 9.8 3.2 KS (mgll) 510 820

Two-stage anaerobic-aerobic chemostat P(,, (d-l) 4.6 2.8 KS (mdl) 680 890

580 KIM AND PAGILLA J. BIOSCI. BIOENG.,

0 300 600 900 1200 1500 Substrate concentration (mg COD/L)

FIG. 3. Comparison of specific growth rate of Gordoniu and Acinetobacter as a function of substrate concentration in a single-stage aerobic chemostat. Symbols: closed circles, Gordoniu; open circles, Acinetobacter.

10.0 to 16.7 d-l) (18, 19) grown with acetate as the carbon source. The K, value of Acinetobacter from this study (KS, 5 10 mg COD/r> was much higher than that reported by Hao and Chang (18) (KS, 65 mg COD/I). Similarly, the Gordoniu KS value of 820 mg COD/I was found to be over two orders of magnitude higher than that reported by Blackall et al. (15) (K,, 2.5 mg COD/I), and it was in the same range as that reported by Baumann et al. (17) (KS, 675 mg COD/Z). These high KS values may be due to the high substrate concentration levels (100-l 500 mg COD/Z) used during this study. Since the KS values of both Gordonia and Acineto- batter were found to be in the same range, the effect of higher values in competition models will be similar for both species.

Chudoba (20) suggested that floe-forming bacteria (such as Acinetobacter sp.) are generally pu,,-strategists, and fila- mentous bacteria (such as Gordonia sp.) are usually K-strat- egists. Although floe-forming bacteria can grow faster (due to a high ,u_) than tilamentous bacteria, a high substrate concentration (due to the high KS) is also necessary. On the other hand, filamentous bacteria can grow even at low sub- strate concentration (low KS), but only slowly (low ,u,,,,,). Based on this theory, Acinetobacter will be p,,-strategists at higher dilution rates, whereas Gordonia will not. Further- more, Gordonia cannot be a K-strategist due to its high KS values when growing on such a readily biodegradable sub- strate as acetate. Hence, when both Acinetobacter and Gor- doniu are present in a mixed culture, Acinetobacter is likely to dominate over Gordonia according to their kinetic pa- rameters obtained experimentally.

In two-stage anaerobic-aerobic chemostats, the ,u~, values were 4.6 and 2.8 d-l for Acinetobacter and Gordonia re- spectively, and the KS values were 680 and 890 mg COD/Z, respectively (Table 1). It can be seen that for both Acineto- batter and Gordonia growing in a two-stage anaerobic- aerobic chemostat system, the p_ values were lower and KS values were higher than those found for the same species grown in a single-stage aerobic chemostat. This suggests

o- I I 1 I I I 0 300 600 900 1200 1500

Substrate concentration (mg COD/Q

FIG. 4. Comparison of specific growth rate of Gordonia and Acinetobacter as a function of substrate concentration in a two-stage anaerobic-aerobic chemostat. Symbols: closed circles, Gordoniu; open circles. Acinetobacter.

that the anaerobic stage prior to the aerobic stage affects the growth and substrate uptake of both Gordonia and Acineto- batter. Furthermore, the kinetic parameters from the two- stage system indicate that Acinetobacter, due to its higher P max and lower KS values, can grow better than Gordonia under anaerobic-aerobic conditions. Figure 4 shows the spe- cific growth rate of Acinetobacter and Gordonia as a func- tion of substrate concentration under anaerobic-aerobic con- ditions in the two-stage chemostat. Each data point is an average of three measurements from three replicate experi- ments. It can be seen that Acinetobacter has a higher spe- cific growth rate than Gordonia over the entire substrate concentration range, suggesting that Gordonia will be out competed when both are present in a mixed culture under anaerobic-aerobic conditions.

Dual culture competition in a single-stage and two- stage hemostat Dual culture experiments were con- ducted with Acinetobacter and Gordonia in continuous flow single-stage aerobic and two-stage anaerobic-aerobic chemostat systems. An equal mass (equal to 50 mg VSS/l) of Gordonia and Acinetobacter taken from mono-cultures grown in aerobic batch reactors was introduced into the dual culture chemostat. The dual culture competition ex- periments were then initiated at dilution rates of 2.0 d-’ and 1.5 d-l in single-stage aerobic and two-stage anaerobic-aero- bic chemostats, respectively. The dilution rates were se- lected based on the maximum specific growth rates ob- tained in single-culture chemostat experiments of Gordoniu and Acinetobacter. In a single-stage chemostat (Fig. 5), the Gordonia concentration decreased exponentially with time to near detection limits (< 10 mg VSS/l) after 96 h from an initial Gordonia concentration of 200 mg VSSIl in the dual culture. Each data point is an average of three measure- ments from three replicate experiments. Acinetobacter grew successfully to steady levels of about 150% of the 200 mg VSSII initial concentration. On the other hand, Gordonia was completely washed out from the dual culture in the two-stage anaerobic-aerobic chemostat with recycling after

COMPETITIVE GROWTH OF GORDONIA AND ACINETOBACTER 58 1 VOL. 95,2003

I I , I I ! 200 a D __,,__. D . .._., .g . . . . . . . . . . 0 . . . . . . . . . . . . ??

,,~ (,,,,_ u ,,(,,,, ~ ,,___._ ??_........... . . . ..o........““.‘“’ ~~g?c---_ I “‘.‘...O . . . . . . . . .

1 I I 1 I I I

0 20 40 60 80 100

Time (h)

FIG. 5. Mathematical model predictions of Acinetobacter and Gordonia competition in a single-stage aerobic chemostat (a) and a two-stage anaerobic-aerobic chemostat (b). Symbols: closed circles, Gordon& experiment; open circles, Gordoniu, model; closed squares, Acinetobacter, experiment; open squares, Acinetobacter, model.

about 24 h addition of the initial inoculum in the same two- stage dual culture chemostat system. These results indicate that Acinetobacter was able to compete successfully with Gordonia under anaerobic/aerobic conditions in the two- stage chemostat, and in the single-stage chemostat under aerobic conditions.

Mathematical model results Model simulations of Acinetobacter and Gordonia competition were performed for both single-stage aerobic and two-stage anaerobic-aero- bic chemostat reactors. In the single-stage aerobic chemo- stat (Fig. 5), the model predictions for Gordonia concentra- tion in the dual culture are in close agreement with those ob- tained experimentally. However, the model over-predicted the Acinetobacter concentration by about 50% of the initial Acinetobacter concentration compared to the experimental results. The model trend pattern predictions for both Gor- donia and Acinetobacter are in good agreement with those obtained from the dual culture experiments. It is likely that the concentration of Acinetobacter with a very high pmaX will be under-predicted compared to the experimental results due to the sensitivity of the model for ,umaX. Any deviations in the experimental conditions in the dual culture competi- tion experiments and the experiments conducted to deter- mine kinetic parameters could amplify the differences in the P max values. It is likely that the effect is more pronounced for Acinetobacter due to its high ,u,, value. In the two-stage anaerobic-aerobic chemostat, Gordonia and Acinetobacter concentrations predicted by the model changed at a lower rate compared to the concentration patterns obtained experi- mentally (Fig. 5). The outcome of the competition in the dual culture was nearly the same based on the model predic- tions and the experimental results under aerobic and anaero- bic-aerobic conditions. Hence, some differences between

the model predictions and experimental data were evident, but the growth pattern trends are in good agreement. Further work with optimum growth conditions, both during kinetic parameter determination and dual culture competition ex- periments, could provide much better agreement between the model and experimental data.

D, : D, : KS, : KS, : k : 6’ : 4 : nl

4 : n2

; I

s, :

t : V” :

x,, :

x,, :

r,, :

Y : n? P : maxl

P : max2

1.

2.

3.

4.

5.

6.

7.

8.

NOMENCLATURE

dilution rate of aerobic reactor (d-‘) = Q/V dilution rate of anaerobic reactor (d-‘) = Q/V,, half saturation constant of Acinetobacter (mgll) half saturation constant of Gordonia (mgll) endogenous decay rate of Acinetobacter (d-l) influent flow rate (l/d) substrate uptake rate of Acinetobacter (mg COD/mg VSSld) substrate uptake rate of Gordonia (mg CODimg VSSld) recycling ratio (return flow rateiinfluent flow rate) effluent substrate concentration (mg COD/I) influent substrate concentration (mg COD/Z) time (d) anaerobic reactor volume (I> biomass concentration of Acinetobacter in the anaerobic reactor (mg VSSII) biomass concentration of Gordonia in the anaero- bic reactor (mg VSS/l> yield coefficient of Acinetobacter (mg VSSlmg COD) yield coefficient of Gordonia (mg VSS/mg COD) maximum specific growth rate of Acinetobacter (d-9 maximum specific growth rate of Gordonia (d-‘)

REFERENCES

Buchan, L.: Possible biological mechanism of phosphorus removal. Water Sci. Technol., 15, 87-103 (1983). Ohtake, H., Takahashi, K., Tsuzuki, Y., and Toda, K.: Up- take and release of phosphate by a pure culture of Acineto- batter calcoaceticus. Water Res., 19, 1587-1594 (1985). Kim, H. and Pagilla, K. R.: Competitive growth of Nocardia and Acinetobacter under anaerobic/aerobic batch operation. Water Res., 34, 2667-2674 (2000). Tandoi, V., Majone, V., May, J., and Ramadori, R.: The be- havior of polyphosphate accumulating Acinetobacter isolates in an anaerobic-aerobic chemostat. Water Res., 32, 29033 2912 (1998). Veldkamp, H.: Ecological studies with the chemostat. Adv. Microb. Ecol., 1, 59-94 (1971). Hao, 0. J., Richard, M. G., and Jenkins, D.: The half satu- ration coefficient for dissolved oxygen: a dynamic method for its determination and its effect on dual species competition. Biotechnol. Bioeng., 25,403-416 (1983). van Niekerk, A. M., Jenkins, D., and Richard, M. G.: The competitive growth of Zoogloea ramigeru and type 021N in activated sludge and pure culture - a mode1 for low F:M bulking. J. Water Pollut. Control Fed., 59, 262-273 (1987). van Niekerk, A. M., Jenkins, D., and Richard, M. G.: A mathematical model of the carbon-limited growth of filamen- tous and floe-forming organisms in low F/M sludge. J. Water Pollut. Control Fed., 60, 100-l 06 (1988).

582 KIM AND PAGILLA J. BIOSCI. BIOENG.,

9. Lau, A. O., Strom, P. F., and Jenkins, D.: The competitive growth of floe-forming and filamentous bacteria: a model for activated sludge bulking. J. Water Pollut. Control Fed., 56, 52-61 (1984).

10. Yoon, H. and Blanch, H. W.: Competition for double growth- limiting nutrients in continuous culture. J. Appl. Chem. Bio- technol., 27,260-268 (1977).

11. Jenkins, D., Richard, M. G., and Daigger, G. T.: Manual on the causes and control of activated sludge bulking and foam- ing, 2nd ed. Lewis Publishers, Chelsia, Michigan (1993).

12. Pitt, P. A. and Jenkins, D.: Causes and control of Nocardia in activated sludge. Res. J. Water Pollut. Control Fed., 62, 143-150 (1990).

13. Pagilla, K. R., Jenkins, D., and Kido, W. H.: Nocardia con- trol in activated sludge by classifying selectors. Water Envi- ron. Res., 68,235-239 (1994).

14. Blackall, L. L., Tandoi, V., and Jenkins, D.: Continuous culture studies with Nocardia amarae from activated sludge and their implications for foaming control. Res. J. Water Pol- lut. Control Fed., 63,44-50 (1991).

15. Greenberg, A. E.: Standard methods for the examination of water and wastewater. Am. Public Health Association, Wash- ing, D.C. (1998).

16. Gerald, C. F.: Applied numerical analysis, 6th ed. Addison- Wesley Publishing Company, California (1998).

17. Baumann, M., Lemmer, H., and Ries, H.: Scum actino- mycetes in sewage treatment plants - I. Growth kinetics of Nocardia amarae in chemostat culture. Water Res., 22, 755- 759 (1988).

18. Hao, 0. J. and Chang, C. H.: Kinetics of growth and phos- phate uptake in pure culture studies of Acinetobacter species. Biotechnol. Bioeng., 29, 8 19-83 1 (1987).

19. Deinema, M. H., van Loosdrecht, M., and Scholten, A.: Some physiological characteristics of Acinetobacter spp. ac- cumulating large amounts of phosphate. Water Sci. Technol., 17, 119-125 (1985).

20. Chudoba, J.: Control of activated sludge tilamentous bulk- ing. VI. Formulation of basic principles. Water Res., 19, 10 17- 1022 (1985).