• 沒有找到結果。

Optimization and performance evaluation of the continuous algal toxicity test

N/A
N/A
Protected

Academic year: 2021

Share "Optimization and performance evaluation of the continuous algal toxicity test"

Copied!
8
0
0

加載中.... (立即查看全文)

全文

(1)

1337

q 1997 SETAC Printed in the USA 0730-7268/97 $6.001 .00

Environmental Chemistry

OPTIMIZATION AND PERFORMANCE EVALUATION OF THE CONTINUOUS ALGAL

TOXICITY TEST

C

HUNG

-Y

UAN

C

HEN

* and K

UO

-C

HING

L

IN

Institute of Environmental Engineering, National Chiao Tung University, 75, Po-Ai Street, Hsinchu, Taiwan 300, Republic of China (Received 9 July 1996; Accepted 4 October 1996)

Abstract—Responses ofSelenastrum capricornutum to cadmium in a chemostat are evaluated under various test conditions defined

by the nitrogen to phosphorus ratio, phosphate strength, dilution rate, and ethylenediaminetetraacetic acid content. Based on the test results, these parameters were optimized to achieve superior test sensitivity. This continuous system was also modified to allow testing on a weekly basis. The performance of the modified testing technique was evaluated, employing six different metal toxicants. Compared with batch test results reported by other researchers and our own experiment, the continuous test achieved superior sensitivities and reproducibility. The test method presented in this study is believed to be an ideal technique both for research and regulatory purposes.

Keywords—Algae Chemostat EC50 Sensitivity Reproducibility

INTRODUCTION

In recent years, toxicity testing with microalgae has been used extensively in ecotoxicogical studies. Several standard algal test protocols were developed for the comparison of the relative toxicity of chemicals and/or waste discharges [1–4]. These methods, although employing different medium com-positions, are quite similar in some respects: they are all ba-sically batch tests and the growth mediums are all enriched with macronutrients including phosphate and nitrogen. In ad-dition, considerable amounts of chelating agents such as eth-ylenediaminetetraacetic acid (EDTA) were added to the growth mediums.

Previous research indicated that the sensitivity of algal tox-icity tests is governed by several factors. The composition of the medium was found to significantly affect the sensitivity of the test algae [5,6]. Furthermore, the type of limiting nutrients was also a possible factor affecting toxicity. Hall et al. [7] reported that algae were more sensitive under phosphorus lim-itation than they were under nitrogen limlim-itation. Stauber and Florence [6] suggested that chelators, which may alter metal toxicity through complexation, should not be included in the medium. Vasseur and Pandard [8] found that toxicity decreased when the quantity of algal inoculum increased. Nyholm and Kallqvist [9] indicated that excessively high biomass could lead to high pH levels due to poor gas-exchange conditions, which might eventually affect toxicity.

The choice of response variables (i.e., biomass or growth rate) also makes a difference in test sensitivity. Analysis of experimental results from the International Standards Organ-ization’s ring-tests [10,11] show that median effective con-centrations (EC50s) based on biomass were generally lower and could differ by a factor of two from EC50s based on growth rate. Nevertheless, some scientists prefer to use growth rate because it allows greater reproducibility of results between laboratories [12,13]. Considering the variation in test results,

* To whom correspondence may be addressed.

various authors [14,15] reported that the intralaboratory pre-cision of batch tests ranged from 20 to 32%. Interlaboratory comparisons, however, showed that EC50 values could vary by several orders of magnitude [16]. Nyholm and Kallqvist [9], in their critical review of batch-culture toxicity tests, in-dicated that results from these test methods varied extensively. Parallel to the development of batch testing techniques, continuous culture in a chemostat was also applied to algal toxicity testing [17,18]. The experimental results from che-mostats demonstrate that continuous tests generally exhibit much better sensitivities than do batch tests. Nevertheless, Kayser [19] reported that the difference in sensitivity obtained by the two testing methods was not highly significant. The fact that contradictions occur is not suprising as they may arise from differences in the growth medium composition or op-erating conditions (dilution rate, temperature, light intensity, etc.) of the different test systems. On the other hand, our pre-vious studies indicated that the sensitivity of the continuous test was correlated with nutrient concentration, which was de-termined by the dilution rate of the system [20].

The above discussions reveal that medium composition, chelator concentration, cell density, and dilution rate of con-tinuous cultures are also likely to be crucial factors in deter-mining the sensitivity of toxicity testing with a chemostat. Our main objective in this study was to evaluate the effects of these parameters in the chemostat and to optimize them for better sensitivities. Another objective was to standardize the contin-uous technique for algal toxicity testing so as to minimize any discrepancies in test results between different researchers. The evaluation consists of two parts, the optimization of the afore-mentioned parameters and the performance evaluation of the optimized test system.

MATERIALS AND METHODS

The alga Selenastrum capricornutum (UTEX 1648) was obtained from the University of Texas, Austin, Texas, USA. The alga was grown in a transparent glass reactor with the volume of the culture medium kept at 400 ml. The growth

(2)

Table 1. Composition of U.S. Environmental Protection Agency’s (EPA) medium and the revised algal growth medium

Compound Concentration (mg/L) Compound Concentration (mg/L) NaNO3 25.5a(12.75b) H3BO3 186 NaHCO3 15.0 MnCl2·4H2O 415.6 K2HPO4 1.04a(0.52b) ZnCl2 3.27 MgSO4·7H2O 14.7 CoCl2·6H2O 1.428 MgCl2·6H2O 12.16 CuCl2·2H2O 0.012 CaCl2·2H2O 4.41 Na2MoO4·2H2O 7.26 FeCl3·6H2O 160.0 Na2EDTA·2H2O 300a(30b) aThe original concentration in the EPA medium.

bConcentration revised by this study.

Table 2. Effects of nitrogen to phosphorus (N:P) ratios on a toxicity test (atD5 0.3/d and Cd 5 0.075

mg/L)a N:P NO3-N (mg/L) PO4-P (mg/L) Cell density (106cells/ ml) m/D 1 2 IR (%) 1 2 11.3 4.2 0.372 8.18 0.59 0.65 41 35 22.5 4.2 0.186 4.39 0.24 0.19 76 81 45.0 8.4 0.186 4.17 0.25 0.21 75 79

aD5 dilution rate: Cd 5 cadmium; NO

3-N and PO4-P5 nutrient concentrations in the influent; m 5 disturbed growth rate; 1 and 25 replicates; IR 5 inhibition rate; IR (%) 5 (1 2 m/D) 3 100.

medium was continuously supplied by variable-speed pumps. Air agitation was used to achieve adequate mixing. The che-mostat reactors were placed in a constant-temperature room at 24 6 18C. Light intensity was set at 4,304 lux (610%). A detailed description of the experimental setup can be found in the authors’ earlier works [20,21]. Samples for analysis were taken directly from the effluent pore. The population density of the algae was determined on a 24 h basis using an electronic particle counter (Culter Electronics, Inc., Luton, England). Growth rate was chosen as the response parameter. The system was first allowed to reach a steady state with no toxicant in-volved. Then toxicant was added both to the reactor and the medium influent until the required concentration level was reached. This addition created a nonsteady state in the reactor. In a chemostat at steady state, the specific growth rate (m) is equal to the dilution rate (D) and the population density re-mains constant. If a toxicant reduces the growth rate, cells wash out faster than they are replaced by growth (i.e.,m , D or m/D , 1) and the population density declines. The new growth ratem (disturbed growth rate) can be calculated from the rate of population decline according to the following equa-tion:

x2

m 5 ln

[

1 2

@

(t22 t ) 1 D1

]

x1

wherex1 andx2 are densities, in terms of cells/ml, at timet1

and t2, respectively. The steady state growth rate, which is

equal toD, can also be considered as the chemostat’s unper-turbed growth rate. Hence, the EC50 is calculated as the tox-icant concentration that reducesm to one-half of its unpertur-bed value. A more detailed discussion regarding the equation and the definition of the EC50 can be found in the first author’s previous work [20].

In order to develop a relatively simple and short-term

tech-nique, the exposure time for the nonsteady state test was set at only 24 h. Furthermore, after completing the first part of evaluation (the optimization of medium composition and test conditions), an 8-L chemostat reactor was run under the same conditions as the 400-ml test vessels. This larger reactor pro-vided a steady supply of algae for toxicity tests conducted in the 400-ml vessels.

For comparison purposes, batch tests were also conducted following the U.S. Environmental Protection Agency’s (EPA) method [2]. Again, algal growth rate was selected as the re-sponse parameter. For each toxicant concentration and the con-trol, three flasks were used for the batch tests. The initial inoculum cell density was approximately 15,000 cells/ml. In our past experiments, inconsistent test results were obtained when inoculum cell density was below 10,000 cells/ml.

All glassware was thoroughly cleaned with phosphorus-free detergent and rinsed with tap water. This was followed by an acid rinse (10% HCl), after which the glassware was neutral-ized with a saturated solution of sodium bicarbonate, and fi-nally rinsed with double-distilled deionized water. Growth me-dium was filter-sterilized through a 0.45-mm membrane filter. Orthophosphate content was analyzed according to the stan-nous chloride method [22]. The NO3-N concentration was also

analyzed by ion chromatography. All chemicals used were of reagent grade.

RESULTS

The EPA medium was selected as the initial basis for this study, simply because it was used in our previous work. Table 1 lists the original composition of the medium. In chemostats, cell density is controlled by both dilution rate and medium strength. Thus, the parameters to be optimized were the nitro-gen to phosphorus (N:P) ratio, dilution rate (D), medium strength, and EDTA concentration. Optimization was governed by two major criteria: the growth condition during the steady state and the sensitivity to toxicants during the second stage of nonsteady state tests.

Optimization of the test system

Table 2 illustrates the three different N:P ratios ranging from 11.3 to 45.0 that were tested atD5 0.3/d. The nutrient concentrations in the influent are NO3-N and PO4-P. Tests for

different conditions were carried out in duplicate. Cell density was measured at steady state. In these tests only nitrogen and phosphate contents were changed. It is clear that the phosphate content was directly proportional to cell density, indicating that phosphorus was the limiting factor for algal growth. Ni-trogen concentration appeared to have no significant effect on algal growth. After establishing steady states, a 1-d nonsteady state toxicity test was conducted with cadmium (cd) (0.075

(3)

Table 3. Effects of phosphate strength on toxicity test (atD5 0.5/d and Cd 5 0.075 mg/L)a Phosphate strength (%) Cell density (106cells/ ml) m/D 1 2 IR (%) 1 2 100 3.56 0.30 0.33 70 67 50 1.70 0.21 0.22 79 78 20 0.72 0.03 0.04 97 96

aD5 dilution rate; Cd is cadmium; m 5 disturbed growth rate; 1 and 2 5 replicates; IR 5 inhibition rate; IR (%)5 (1 2 m/D) 3 100.

Table 4. Effects of dilution rate and phosphate strength on algal growth and toxicity test (Cd5 0.075 mg/L)a

Phosphate strength (%) Cell density (106cells/ ml) pH PO4-P (mg/L) NO3-N (mg/L) Growth condition mave/D Mean inhibition rate (%) Sensitivity D5 0.3/d 100 4.06 8.30 0.002 1.992 Fair 0.22 78 Good 50 1.73 7.78 0.002 1.000 Fair 0.16 84 Good 20 — — — — Poor — — — D5 0.5/d 100 3.56 8.24 0.007 2.020 Good 0.31 69 Good 50 1.70 7.66 0.005 1.008 Good 0.21 79 Good 20 0.72 7.52 0.005 0.453 Poor 0.03 97 Good D5 0.7/d 100 3.34 8.15 0.011 2.222 Good 0.60 40 Poor 50 1.56 7.60 0.011 1.173 Good 0.48 52 Fair D5 0.9/d 100 3.20 8.23 0.016 2.279 Good 0.73 27 Poor 50 1.51 7.58 0.015 1.218 Good 0.53 47 Poor aCd5 cadmium; PO

4-P and NO3-P5 equilibrium concentrations in the reactor; mave5 average disturbed growth rate (day21);D5 dilution rate (day21).

mg/L), and results in relative growth rates (m/D) and the cor-responding inhibition rates were obtained. The choice of 0.075 mg Cd/L was made because significant inhibition was observed in our past experiment. Comparisons were made to determine whether one test condition consistently achieved better sen-sitivities than others. It was found that N:P ratios of 22.5 and 45.0 did not significantly differ from each other. Algae in medium with an N:P ratio of 11.3 were the least sensitive, with a mean inhibition rate of only 38% (Table 2). Because increasing nitrogen content did not improve the test sensitivity, for the following experiments the N:P ratio was kept the same as that in the original EPA medium (i.e., 22.5).

Table 2 indicates that phosphorus content in the medium directly controls cell density. Hence, evaluating the effects of medium strength (or cell density) on toxicity requires altering only the phosphate concentration in the medium. The primary concern here is to prevent other nutrients or trace elements from becoming limited while the strength of the entire medium is being reduced. Therefore, the term ‘‘medium strength’’ should be more appropriately replaced by ‘‘phosphate strength.’’ However, the nitrogen concentration must be ad-justed with the phosphate content to maintain the desired N: P ratio. Table 3 presents the effects of phosphate strength on toxicity tests atD5 0.5/d and Cd 5 0.075mg/L. One hundred percent phosphate strength represents the original EPA me-dium and 50% strength implies that the nitrogen and phos-phorus contents were reduced to one-half of that in the EPA medium, while the amounts of other nutrients remained the same. Cell density was generally proportional to phosphate strength. In terms of sensitivity, we observed that inhibition

rate increased with a decrease in cell density. Such a phenom-enon can be explained as follows: higher cell density means lower toxicant load per individual cell and, thus, a lower degree of inhibition. Vasseur and Pandard [8] made a similar obser-vation that greater EC50 values are associated with a higher density of inoculum.

Table 4 presents a summary of a series of comparisons with respect to different dilution rates. Orthophosphate and nitrate contents were the equilibrium concentrations in the reactor and were determined when the system reached steady state. AtD 5 0.5/d, no significant difference was observed for the equilibrium phosphorus concentration between the three phos-phate strengths listed. This phenomenon suggests that the med-ium’s phosphate strength does not affect the phosphorus con-centration in the chemostat under the same dilution rate. The pH values were all maintained within the range of 7.5 to 8.5. The rating of the growth conditions was based primarily on the frequency of obvious microbial contamination, which was manifested in a change of the color of the culture from a blue-green hue to a slightly yellowish one. Test systems that easily or occasionally induced the above condition were rated as poor or fair, respectively. Good growth conditions were those where no obvious contamination occurred and where a steady state could be easily maintained for long periods (e.g., 3–6 months). Significant microbial contamination generally occurred under low-strength and low-dilution-rate conditions. Previous studies have indicated that bacterial contamination is unavoidable for chemostates under long-term operation. However, the bacterial biomass is relatively small compared to that of the total culture, and its effect on the algal culture is almost undetectable [23].

(4)

Fig. 1. Growth conditions ofSelenastrum capricornutum under three different concentrations of ethylenediaminetetraacetic acid in a chemostat

with dilution rate (D)5 0.5/d.

Table 5. Effects of concentration of ethylenediaminetetraacetic acid (EDTA) on the algal growth and toxicity test (atD 5 0.5/d, 50%

phosphate strength, and Cd5 0.075 mg/L)a Na2 EDTA· 2H2O (mg/L) Cell density (3106 cells/ml) Inhibition rate (%)b 1 2 300 1.676 0.10 75.3 80.3 150 1.616 0.07 90.6 87.4 30 1.706 0.06 97.0 93.6

aD5 dilution rate; Cd 5 cadmium. b1 and 2 are replicates.

For the case ofD 5 0.3/d and 20% phosphate strength, after several trials, steady state was never attained. Toxicity testing with 0.075 mg/L of Cd indicated that the inhibition rate in-creased with decreasing phosphate strength for the four dif-ferent dilution rates (see Table 4). In addition, the degree of inhibition increased as the dilution rate diminished. This is in accordance with our previous report [20], which concluded that tolerance in algal cells is reduced as nutrients become more limiting. The differences in tolerance of various dilution rates can also be compared (see Table 4). At 100% phosphate strength, cell densities underD 5 0.5/d and D 5 0.7/d were not significantly different. However, the inhibition rate for the former was 69%, whereas that for the latter was only 40%. This indicates that the average tolerance of algal cells

de-creases at lower dilution rates. It can be concluded that the sensitivity of algal toxicity tests is mainly determined by two factors: dilution rate and cell density. The dilution rate deter-mines the general tolerance in a microbial culture. Under a specific dilution rate, the cell density of an algal culture, which is regulated by the phosphate strength, determines the indi-vidual toxicant load of the cell. The case ofD 5 0.5/d and 50% phosphate strength, which provided both good growth condition and sensitivity, was thus selected for further opti-mization evaluation. The test condition at 100% strength (D

5 0.5/d) was not selected because of the apparent lower

in-hibition rate.

Based on the above test conditions and medium compo-sition, the next task was to determine the appropriate concen-tration for chelators. Figure 1 depicts the growth conditions, from initial inoculation to reaching a steady state, employing three different Na2EDTA·2H2O concentrations. Medium with

150mg/L yielded the lowest biopopulation (1.61 3 106cells/

ml) at a steady state, but it did not differ significantly from the other two cases. Comparisons based on a student’st test at the a 5 0.05 level also showed no differences between Na2EDTA·2H2O5 150 and 30 mg/L. Reducing Na2EDTA·2H2O

content to 30mg/L apparently did not adversely affect algal growth. Table 5 lists algal responses to 0.075 mg/L of Cd. As expected, data from duplicate tests consistently showed a better sensitivity at lower EDTA concentrations. Hence, 30mg/L of Na2EDTA·2H2O was considered adequate for the chemostat

(5)

Fig. 2. Control chart based on cell density in the algal incubator.

system. Table 1 also lists the revised algal growth medium, which can be compared to the original EPA medium.

Obviously, our selection of the optimum test condition was not entirely based on the test’s sensitivity. The final choice, in fact, was to sacrifice part of the sensitivity in exchange for a more stable system. Such a sacrifice is aimed at ensuring that a steady state can be easily maintained for long-term op-erations. The significance of such a choice is revealed in the operation of an algal incubator, which will be discussed below.

Performance evaluation

Using the revised algal medium given in Table 1, an 8-L algal incubator was run at D 5 0.5/d. Records of the first month of operation were collected in order to establish quality assurance criteria for toxicity testing. Figure 2 is the control chart for cell densities in the chemostat, which shows that the variation of the biopopulation was within 1.7 3 106 6 6%

cells/ml. The pH values were 7.65 6 0.15 at steady state. Nutrient concentrations were periodically checked and were quite similar to the corresponding values given in Table 4. If cell density and pH were within the normal range for three consecutive days, toxicity testing was conducted by dispensing algal liquor into six 400-ml chemostat reactors (one control and five treatments) for the nonsteady state test. A preliminary 2-d comparison showed that no significant differences were observed for cell densities, pH, and nutrient concentrations, within the 400-ml vessels after being transferred from the larg-er reactor. The purpose of plarg-erforming a control run was to determine whether any unidentified variations occurred that

might significantly affect algal growth. The 24-h toxicity test was performed in exactly the same manner as mentioned ear-lier: at the beginning of the test, a toxicant was added to the reactor and the medium influent until the required concentra-tion level was reached. For each test run, 2.4 L of algal liquor was removed from the 8-L incubator (30% volume). A pre-vious study indicated that a sample withdrawal up to 30% of the chemostat volume does not have a significant impact on the algal population [23].

Six metal toxicants (Cd, lead [Pb], zinc [Zn], copper [Cu], nickel [Ni], and mercury [Hg]) were tested to evaluate the general sensitivity of the optimized continuous system. Table 6 lists the test results in terms of EC50, standard deviation (SD) of the EC50, and the coefficient of variation (CV). The CV values for various metals varied from 8.5 to 22.6%. In addition, for comparative purposes, Table 6 lists data derived from the batch method as reported in the literature [24–28]. These data were mainly calculated on biomass unless other-wise specified. Obviously, for Cd, Zn, Pb, and Hg, the literature EC50 values were two to four times greater than our values, which were calculated on growth rate. Similar sensitivities were found for Cu. In addition, based on a growth rate cal-culation, the literature value for Cu was 20 times greater than that from the continuous test. On the other hand, the data for Ni indicated that the batch test was more sensitive than the continuous technique. Chiaudani and Vighi [27] used an in-oculum density of 4,000 cells/ml to obtain such a low EC50 value. According to the findings of Vasseur and Pandard [8], such superior sensitivity could be related to the low cell density

(6)

Table 6. Sensitivity and reproducibility of the continuous algal toxicity testa Metal toxicant EC50 (mg/L) n SD CV (%) Reportedb EC50 (mg/L) Cd 0.013 5 0.0011 8.5 0.04 [23] Cu 0.021 3 0.0031 14.5 0.04 [23] 0.4 [24]c Zn 0.015 3 0.0017 11.3 0.06 [23] Pb 0.256 3 0.0580 22.6 0.5 [25] Ni 0.125 3 0.0125 10.2 0.012 [26] Hg 0.027 3 0.0053 19.2 0.06 [27]

aEC505 median effective concentration; n 5 numbers of replicate tests; SD 5 standard deviation; CV 5 coefficient of variation; Cd5 cadmium; Cu 5 copper; Zn 5 zinc; Pb 5 lead; Ni 5 nickel; Hg5 mercury.

bReference numbers are given in brackets. cEC50 based on growth rate.

Table 7. Reproducibility of median effective concentrations (EC50s) (for copper) obtained withSelenastrum capricornutum using various

methodsa Method n EC50 (mg/L) CV (%) Source

Conventional 6 13.8 32.1 Van Coillie et al. [15] Conventional 3 29.9 22.6 Blaise et al. [14]

MP 3 54 20.4 Blaise et al. [14]

MP 3 60 25.0 Blaise et al. [14]

Chemostat 3 21.3 14.5 Our data [this study] aConventional method 5 the algal bottle test; MP method 5 the

microplate algal test;n 5 number of replicates; CV 5 coefficient

of variation.

Table 8. Median effective concentration (EC50) values and 95% confidence intervals (CIs) for continuous and batch toxicity tests

Metal toxicant Continuous test EC50 (mg/L) 95% CI (mg/L) Batch test EC50 (mg/L) 95% CI (mg/L) Ratio of batch EC50 to continuous EC50 Cd 0.013 (0.0040b) 0.012–0.017 0.341c(0.238) 0.110–1.672 26.3 Cu 0.021 (0.0029) 0.018–0.031 0.038 (—)d 0.031–0.054 1.83 Zn 0.015 (0.013) 0.011–0.026 0.178 (0.132) 0.110–0.271 11.9 Pb 0.256 (0.0044) 0.285–0.356 2.655 (0.0163) 1.983–4.539 10.4 Ni 0.125 (0.113) 0.106–0.173 0.233 (0.173) 0.174–0.311 1.91 Hg 0.027 (4.673 10212) 0.011–0.085 0.009 (1.503 10212) 0.002–0.034 0.33 aCd5 cadmium; Cu 5 copper; Zn 5 zinc; Pb 5 lead; Ni 5 nickel; Hg 5 mercury.

bThe EC50 was estimated based on free metal concentration. cValue also presented in Lin and Chen [21].

dNinety-nine percent of the Cu was complexed with ethylenediaminetetraacetic acid.

in the inoculum. However, the latest ASTM method [4] rec-ommends the use of an inoculum density of 23 104cells/ml.

Table 7 compares the test reproducibilities of different test methods using Cu. The CV values from chemostat tests were much lower than those obtained by batch methods (i.e., con-ventional bottle test and microplate techniques). One possible explanation is that, for chemostat testing, algae were always maintained at the steady state in similar physiological con-ditions, whereas for batch tests, the variously prepared initial inoculums could result in different physiologic states.

To create a uniform basis for comparison, batch tests were conducted following the EPA method [2]. The EC50s calcu-lated based on growth rate were compared with the result from the continuous tests, as given in Table 8. The continuous test was found to be more sensitive than the batch test for Cd, Cu, Zn, Pb, and Ni. Ratios of EC50s between the two types of tests ranged from 1.83 to 26.3. The 95% confidence intervals for the two sets of EC50s did not overlap. Hence, we may assert that the differences are obvious at a 5% level of sig-nificance. The continuous method was found to be less sen-sitive to Hg, as its EC50 was three times greater than that of the batch test. The difference, however, was not significant at the 5% level. The lower sensitivity of the continuous test to Hg may be due to the formation of more volatile forms of Hg from reactions between Hg and organic compounds released by algae. Aeration may cause a rapid decrease in the amount of Hg in solution [29]. Comparisons were also made according to the free ionic forms of the metals, which were considered

to be the most toxic. Free metal concentrations in standard EPA medium and in our revised medium were calculated using the MINEQL program [30]. The EC50s based on free ion concentrations are calculated and listed in Table 8. Similarly, the continuous test resulted in lower EC50 values for most cases but not for Hg and Cu. The literature has indicated that certain metal–EDTA complexes (CuEDTA and ZnEDTA) are also quite toxic [31]. Hence, free Cu ions may be the most toxic form, but certainly are not the major toxic form of Cu. Therefore, irrespective of whether relating to total concentra-tion analysis or to ionic form concentraconcentra-tion analysis, EC50s obtained from the continuous test are generally lower than those obtained from the batch test. This means that the con-tinuous test is more sensitive than is the batch test. Figure 3 depicts the concentration–response curves based on the total Cd concentrations from both batch and continuous tests. The slope of the concentration–response curve is very steep for the continuous test, which implies that the tolerances for algae are distributed within a narrow range. In comparison, the moderate slope of the batch test indicates that the tolerance distribution is wider. Our interpretation of such a difference in slopes is as follows: the algae cultured in the chemostat probably have more uniform tolerances because they are kept by the steady state at similar physiological conditions. We believe such uni-formity may be one of the reasons for the good reproducibil-ities shown in Tables 6 and 7.

To identify any inconspicuous changes that might have oc-curred in the algal culture, Cd was selected as the reference toxicant. As shown in Table 6, the five tests for Cd were conducted within a period of 6 months. The CV value was only 8.5%, which indicates that the general tolerance for the algal culture was very consistent during long-term operations.

(7)

Fig. 3. The concentration–response relationships forSelenastrum capricornutum in batch and continuous toxicity tests.

Additional quality assurance criteria, including cell density and pH for the 8-L incubator, were discussed above. The contin-uous tests in our performance evaluation were conducted out replicates, primarily because of the maximum 30% with-drawal of the chemostat volume. However, as revealed in Ta-bles 2, 3, and 5, the continuous tests produced minute varia-tions among replicates. This is not always found in batch test results. For such a stable system, tests can be run with no replication because any abnormal results can still be deter-mined from the relative inhibition rates under different toxicant concentrations. On the other hand, the volume of the test vessel could also be reduced to 200 ml to allow for duplicate tests. The continuous toxicity test system developed in this study poses the following advantages over batch tests: the test du-ration is only 24 h instead of the 96 h required by traditional tests and this method is also considerably more sensitive and has superior reproducibility. Previously, the major disadvan-tage of the chemostat system, thereby limiting its application in algal toxicity assessment, was that it is quite time-consum-ing for the system to reach the steady state. Subsequent toxicity testing often severely inhibits culture growth and causes the recovery to take a longer time. Our modification of adding the 8-L incubator, however, minimized the disturbance of the steady state brought about by toxicity testing and enables us to conduct tests on a weekly basis. This is almost equivalent to the total run time for the batch test (including the time for preparing the inoculum). In addition, according to our own experience, the batch test is not necessarily less labor-intensive than the continuous one. We believe the modified test system is ideal for laboratories or regulatory agencies that are rou-tinely involved in algal toxicity testing.

CONCLUSIONS

This study describes the effects of the N:P ratio, dilution rate, phosphate strength, and EDTA concentrations on algal

growth and algal sensitivity to metal toxicants in a chemostat. The N:P ratio for the original EPA medium, (22.5) was found to be adequate. A low phosphate strength resulted in low cell density and, consequently, better algal sensitivity. Likewise, a low dilution rate for the chemostat improved test sensitivity. The EDTA content was reduced to 30mg/L without any ad-verse effect on algal growth. Based on these results, we have optimized the continuous algal testing system with respect to the above parameters. This optimized system can be considered as the basis for standardizing future toxicity testing employing the continuous system, to minimize discrepancies arising be-tween laboratories. We modified the continuous technique by introducing an algal incubator so that testing can be conducted more frequently. In addition, several quality assurance criteria were developed to insure better laboratory precision of the test. The performance of the continuous testing technique de-veloped by this study was evaluated with six different metal toxicants. Compared to batch test results reported in the lit-erature as well as our own experiment, the continuous test achieved superior sensitivities and reproducibility compared to the batch test. We believe the test method discussed herein is an ideal technique for both research and regulatory purposes.

Acknowledgement—This research was supported in part by grant NSC

84-2211-E-009-007 from the National Science Council, Taiwan, Re-public of China.

REFERENCES

1. Organization for Economic Cooperation and Development. 1984. Guideline for testing chemicals. 201. Alga growth inhibi-tion test. Paris, France.

2. U.S. Environmental Protection Agency. 1985. Freshwater algae acute toxicity test.Fed. Reg. 50:39, 323–39, 333.

3. International Standards Organization. 1987. Water quality— Algal growth inhibition test. Draft International Standard ISO/ DIS 8692. Geneva, Switzerland.

(8)

guide for conducting static 96h toxicity tests with microalgae. E1218-90. InAnnual Book of ASTM Standards. Philadelphia, PA.

5. Millington, L.A., K.H. Goulding and N. Adams. 1988. The influence of growth medium composition on the toxicity of chem-icals to algae.Water Res. 22:1593–1597.

6. Stauber, J.L. and T.M. Florence. 1989. The effect of culture medium on metal toxicity to the marine diatomNitzschia clos-terium and the freshwater green alga Chlorella pyrenoidosa. Wa-ter Res. 23:907–911.

7. Hall, J., F.P. Healey and G.G.C. Robison. 1989. The interaction of chronic copper toxicity with nutrient limitation in two chlo-rophytes in batch culture.Aquat. Toxicol. 14:1–13.

8. Vasseur, P. and P. Pandard. 1988. Influence of some experi-mental factors on metal toxicity toSelenastrum capricornutum. Toxic. Assess. 3:331–343.

9. Nyholm, N. and T. Kallqvist. 1989. Methods of growth inhibition toxicity tests with freshwater algae.Environ. Toxicol. Chem. 8:

689–703.

10. Hanstveit, A.O. and H. Oldersma. 1981. Evaluation of the re-sults of the second ISO-interlaboratory study with an algal tox-icity test. ISO/TC147/SC5/WG5. International Standards Orga-nization, Geneva, Switzerland.

11. Hanstveit, A.O. 1982. Evaluation of the results of the third ISO-interlaboratory study with an algal toxicity test. ISO/TC147/SC5/ WG5. International Standards Organization, Geneva, Switzer-land.

12. Nyholm, N. 1985. Response variable in algal growth inhibition toxicity tests—Biomass or growth rate?Water Res. 19:273–279.

13. Nyholm, N. 1990. Expression of results from growth inhibition toxicity tests with algae.Arch. Environ. Contam. Toxicol. 19:

518–522.

14. Blaise, C., R. Legault, N. Bermingham, R. Van Coillie and P.

Vasseur. 1986. A simple microplate algal assay technique for

aquatic toxicity assessment.Toxic. Assess. 1:261–281.

15. Van Coillie, R., P. Couture, R. Schoenert and C. Thellen. 1982. Mise au point d’une ’evaluation rapide de la toxicit’e originale des effluents et de leurs composantes a l’aide d’algues. Environ-nement Canada, Service de la Protection de l’EnvironEnviron-nement, Quebec, PQ, Canada.

16. Hanstveit, A.O. 1980. Evaluation of the results of the European ISO-test programme with algal toxicity tests. ISO/TC147/SC5/ WG5N16. International Standards Organization, Geneva, Swit-zerland.

17. Hall, J., F.P. Healey and G.G.C. Robison. 1989. The interaction of chronic copper toxicity with nutrient limitation in chemostat cultures ofChlorella. Aquat. Toxicol. 14:15–26.

18. Wong, P.T.S., Y.K. Chau and D. Patel. 1983. The use of algal batch and continuous culture techniques in metal toxicity study. In J.O. Nriagu, ed.,Aquatic Toxicology. Wiley, New York, NY,

USA, pp. 449–466.

19. Kayser, H. 1976. Waste-water assay with continuous algal cul-tures: The effect of mercuric acetate on the growth of some marine dinoflagellates.Mar. Biol. 36:61–72.

20. Chen, C.-Y. 1994. Theoretical evaluation of the inhibitory effects of mercury on algal growth at various orthophosphate levels.

Water Res. 28:931–937.

21. Lin, K.-C. and C.-Y. Chen. 1996. The effect of limiting nutrient on metal toxicity to Selenastrum capricornutum. Toxicol. En-viron. Chem. 56:47–61.

22. American Public Health Association, American Water Works

Association and Water Pollution Control Federation. 1992.

Standard Methods for the Examination of Water and Wastewater,

18th ed. American Public Health Association, Washington, DC. 23. Healey, F.P. and L.L. Hendzel. 1975. Effect of phosphorus de-ficiency on two algae growing in chemostats.J. Phycol. 11:303–

309.

24. Toussaint, M.W., T.R. Shedd, W.H. Van Der Schalie and G.R.

Leather. 1995. A comparison of standard acute toxicity tests with

rapid-screening toxicity tests.Environ. Toxicol. Chem. 14:907–

915.

25. Blaylock, B.G., M.L. Frank and J.F. McCarty. 1985. Com-parative toxicity of copper and acridine to fish, Daphnia and

algae.Environ. Toxicol. Chem. 4:63–71.

26. Monahan, T.J. 1976. Lead inhibition of chlorophycean microal-gae.J. Phycol. 12:358–362.

27. Chiaudani, G. and M. Vighi. 1978. The use ofSelenastrum capricornutum batch cultures in toxicity studies. Mitt. Int. Ver. Limnol. 21:316–329.

28. Beaubien, A., L. Lapierre, A. Bouchard and C. Jolicoeur. 1986. A multispecies toxicity assessment procedure using microcalor-imetry: Comparison with other toxicity evaluation methods. Tox-ic. Assess. 1:187–200.

29. Ben-Bassat, D. and A.M. Mayer. 1975. Volatilization of mecury by algae.Physiol. Plant. 33:128–132.

30. Westall, J.C., J.L. Zachary and F.M.M. Morel. 1976. MINEQL, a computer program for the calculation of chemical equilibrium composition of aqueous systems. Technical Note 18. Department of Civil Engineering, Massachusetts Institute of Technology, Cambridge, MA, USA.

31. Starodub, M.E., P.T.S. Wong, C.I. Mayfield and Y.C. Chau. 1987. Influence of complexation and pH on individual and com-bined heavy metal toxicity to fresh green alga. Can. J. Fish. Aquat. Sci. 44:1173–1180.

數據

Table 1. Composition of U.S. Environmental Protection Agency’s (EPA) medium and the revised algal growth medium
Table 4. Effects of dilution rate and phosphate strength on algal growth and toxicity test (Cd 5 0.075 mg/L) a
Fig. 1. Growth conditions of Selenastrum capricornutum under three different concentrations of ethylenediaminetetraacetic acid in a chemostat
Fig. 2. Control chart based on cell density in the algal incubator.
+3

參考文獻

相關文件

Wang, Solving pseudomonotone variational inequalities and pseudocon- vex optimization problems using the projection neural network, IEEE Transactions on Neural Networks 17

Define instead the imaginary.. potential, magnetic field, lattice…) Dirac-BdG Hamiltonian:. with small, and matrix

The Model-Driven Simulation (MDS) derives performance information based on the application model by analyzing the data flow, working set, cache utilization, work- load, degree

Microphone and 600 ohm line conduits shall be mechanically and electrically connected to receptacle boxes and electrically grounded to the audio system ground point.. Lines in

Without making any improvement in the surface treatment of the solar cell, due to the refractive index of incident medium refractive index and different, which led to the

What are the factors of strengths, weaknesses, opportunities and threats for instant coffee producers in Indonesia?... What is the competitive strategy that should

The analysis indicated that, community commitment, dependence, and reputation positively influence sense of virtual community.. The findings also reveal that reciprocity

are the largest manufacturers, but the operating performance that the Asia Cement Co’s average of four indicators are leading the same industry and named in the third