Hydrobiologia vol . 56,
1,
pag . 39-47, 1977
TOXICITY OF INTERMITTENT CHLORINATION TO FRESHWATER FISH : INFLUENCE OF TEMPERATURE AND CHLORINE FORM Alan G. HEATH Department of Biology, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061 U .S .A . Received December 22, 1976 Keywords : Toxicity, Chlorine, Chloramine, Monochloramine, Fish, Temperature
Abstract Fingerling size
Salmo gairdneri, Oncorhynchus kisutch, Notemigonus crysoleucas, Cyprinus carpio, and Ictalurus punctatus
were exposed in the laboratory three times daily for up to seven days to pulses of either free chlorine or monochloramine . This regime simulated conditions often encountered in the outfall of steam electric generating plants which chlorinate intermittently . LCso's, LT50's, and response isopleths giving various percentage mortalities, were computed from the bioassays . S. gairdneri, O . kisutch, and I. punctatus were the most sensitive to both types of chlorine . C. carpio were most resistant and the N. crysoleucas were intermediate in sensitivity . Temperature had relatively little effect on the toxicity of intermittent chlorine to the species tested . In this type of test regime, free chlorine was three to fourteen-fold more toxic (depending on the species) than monochloramine . Water quality criteria for the protection of fish should, in the future, take this differential toxicity into consideration .
Introduction Chlorine is used extensively as a disinfectant for municipal water supplies and waste water and as a method of anti-fouling in cooling water systems . About 9o percent of all steam electric generating plants in the U .S .A . chlorinate their cooling water on a programmed basis (White, 1975) and with the construction of more plants planned, the potential for environmental impact is noteworthy . There is currently a considerable body of experimental data on the toxicity to fish of continuously dosed chlorine which has relevance to disinfection systems (Brooks &
Dr. W. Junk b. v . Publishers - The Hague, The Netherlands
Seegert, 1975 ; Brungs, 1976) . However, in anti-fouling operations, the most common practice in once-through systems is to use intermittent chlorine doses of about 30 minutes two or three times in each 24 hours (White, 1975) . Cooling tower `blowdown operations' also produce intermittent exposures of non-target organisms to chlorinated effluent . According to recent reviews there is a paucity of information on which to base water quality criteria for intermittent chlorination practices (Brooks & Seegert, 1975 ; Brungs, 1976) . Furthermore, little is known of the effects temperature has on chlorine toxicity (Brooks & Seegert, 1975) . The chemistry of chlorine in natural waters is complex (White, 1972) but for most practical purposes so far as toxicity to non-target organisms is concerned, it exists, in two forms : HOCi or OC1 - , commonly called free chlorine, and NH2C1, monochloramine or combined chlorine . Total residual chlorine (TRC) refers to the sum of the free and combined chlorine concentrations . The relative proportion of free and combined chlorine present following a chlorination dose depends primarily on the concentration of ammonia present in the receiving water to combine with the chlorine . There is considerable disagreement in the literature regarding the relative toxicity to fish of the two chlorine forms (Brooks & Seegert, 1975 ; Dickson et al., 1977 ; Merkens, 1958 ; Rosenberger, 1971 ; Tompkins & Tsai, 1976) . After reviewing the available literature, Brungs (1973) concluded there was little difference between their toxicities to fish except that free chlorine seemed to act faster . This conclusion was based primarily from data on continuous exposure studies . Preliminary tests in this laboratory using intermittent exposure to chlorine suggested a 39
marked toxicity difference between free and combined chlorine so this study was initiated using several species of fish at different temperatures .
Materials and methods Fingerling size fish were used for the toxicity bioassays . The species and their sources were as follows : Rainbow Trout (Salmo gairdneri), U .S . National Hatchery, Wytheville, Virginia ; Coho Salmon (Oncorhynchus kisutch), Platte River State Hatchery, Michigan ; Carp (Cyprinus carpio), private hatchery, Windsor, Virginia ; Golden Shiner (Notemigonus crysoleucas), private hatchery, Windsor, Virginia ; Channel Catfish (Ictaluruspunctatus), private hatchery, Windsor, Virginia . The fish were maintained in the laboratory in 19o L or 38o L aquaria which received a steady flow of dechlorinated, municipal water, Water for the holding and bioassay tanks was dechlorinated by passage through a column of activated charcoal . Some water quality characteristics were : average hardness 45 mg/L, conductivity 150 µMHOS, dissolved oxygen near saturation, copper 0 .05 mg/ L or less and zinc less than 0 .02 mg/ L . In some cases, oxytetracycline was used to control disease . Hemoglobin concentration from tail blood was measured by the cyanmethemoglobin method routinely in all but the shiners . These tests indicated an anemic condition in two batches of fish and this is so indicated in the Results . All other fish were apparently healthy . Feeding was daily ad libitum using Purina Trout Chow . All the fish ate well and grew in the laboratory . Test fish were acclimated to the experimental temperatures at least two weeks before being bioassayed . The laboratory system for the bioassays was specially built for this study . It consisted of ten 25 liter aquaria, each receiving a carefully regulated flow of water at 36 liters per hour from a temperature controlled head box . The aquaria were partially submerged in two large fiberglass tanks for temperature control. Water entered one end of each aquarium and exited through a stand pipe which drew water from the bottom at the other end . In order to supply intermittent chlorination (pulses) to the test aquaria, metering pumps were used to draw from a concentrated calcium hypochlorite solution which was injected through a `T' into the water inflow of each aquarium . The metering pumps were controlled by a timer that turned them on for 45 minutes three times each 24 hrs . This caused the concentration of chlorine to rise and 40
fall in the aquaria as illustrated in Fig . i . The shape of the `pulse' was similar to field measurements taken in the receiving waters of some steam electric generating plants (Dickson et al., 1974 ; Draley, 1973) . In order to get monochloramine pulses, an ammonium chloride solution was added to the inflowing water at the same point and at the same time as the hypochlorite solution The amount of NH4C1 added was determined experimentally each run so that the highest flow of hypochlorite solution was converted entirely to monochloramine . This ammonia solution was also added to the control tanks as well . In order to test for ammonia mortality, the concentration used in the control tanks was generally higher than that in the chlorine tanks . In no case was there any death in the ammonia controls . For each fish bioassay test, four concentrations of intermittent chlorine were used in duplicate aquaria . There were generally to fish per aquarium . In addition, two aquaria served as controls (i .e . no chlorine exposure) . Dead fish were removed and weighed approximately hourly for the first 12 hours and subsequently, every 4 hours except between the hours of 2300 and o800. The amplitude of the chlorine pulse was checked in each aquarium twice daily . Chlorine measurements were made using a Wallace & Tiernan amperometric titrator and both free and combined chlorine were determined each time . Although some workers (Barch & Truchan, 1976) have reported difficulty in measuring free chlorine, we encountered no problems in this regard . Periodically, analyses were performed for pH and ammonia nitrogen
Fig. i . Sample chlorine concentration changes in a bioassay tank during two successive chlorination pulses . This illustrates the shape of the pulse . All concentrations given in this paper were measured at the peak of the pulses .
(Chaney & Marbach, 1962) . The pH did not change significantly during the chlorination pulses . Ammonia nitrogen samples were taken at the chlorination pulse peak . At the end of each run, which generally lasted more more than 7 days, all remaining fish were removed, measured, and weighed and the aquaria allowed to flush with chlorine-free water for two or more days . A total of 3o bioassay runs were made . Statistics on water quality and fish size are given in Table i . Over the limited size range of fish used, there was no obvious relationship of mortality to fish size observed . Median lethal concentrations (LC5o's) were determined by probit analysis methods (Finney, 1971) . Computations were performed using an IBM 37o and the statistical analysis system (SAS) probit analysis procedure combined with a program to compute 95 percent confidence intervals on the obtained value . Careful observation of the bioassay tanks made it possible to note when fish first lost equilibrium . LC5o's and LT5o's were also computed using equilibrium loss as the end point rather than death . In all cases, the numbers were well within the confidence intervals (see Table 2 for example) for those computed on death so only the latter are reported in the other tables for each species .
Results The data presentations in this paper are of three types :
Table 1 . Water Qu,hly and Phh Smlisolm
Avg NH, -N
BioesuyT-Conditions
Ayg. pH
A g
nig/L
Avg. length
N
I-)
fi
(i) Median lethal concentration (LC50) . This concentration is that computed for the peak of the chlorine pulses given three times daily for the time interval indicated . (2) Median lethal time (LT5o) . This is the median time to death for a given concentration of chlorine pulse . At the higher concentrations of free chlorine there was a tendency for mortality to occur during or immediately after a chlorination pulse . At lower concentrations, death occurred more randomly . (3) Response isopleth (in cases where data were extensive) . These data points were computer derived from probability curves and indicate predictions of varying amounts of mortality ranging from no death to complete loss of the population at various exposure time intervals . The zero percent plot on the response isopleths represents the highest chlorine concentrations measured where no death occurred over the exposure time indicated . It does not necessarily represent a `safe' concentration but may approach it depending on how `safe' is defined . Since all the dilution water had some ammonia present, it was not possible to obtain pure solutions of free chlorine . In the assay using free chlorine, generally, the higher the test concentration, the higher the percentage of free chlorine . Consequently, for all data tables and graphs of free chlorine there was also some monochloramine (combined chlorine) present and, therefore, the total residual was higher . An indication of the relative amounts of the two forms in the test tanks can be gained by examination of LT50 data in the tables . In bioassays using monochloramine, there was never any free chlorine present . Throughout the data tables, the 95%o confidence intervals are in parentheses . In some cases the data were insufficient to yield an interval with my computer program and, thus, some blanks are present .
uainnr»~ no~~e rtee .m~rme 5°C 17° C 17°C Chloramlne 11 °
<01 <01 D47 4g
6 .57 6 .57 6 .8
6 .7 9 .6
90 92 0 .2 10 .7
200 190 100 00
Rainbow Trout (Salmo gairdneri)
CO" S41-1 i'ree ChWnne 6"C 12°C r,mi e 6°c Iz°c
<0) 02
1 . 86 Llo
50 750
sB io
250 1-I6
so loo
Carp Ite, cM1lorine 6`C 1a°c a,iornmi11e 6°c
<01 mos
775
10 .5
100
1 .17 »
7.6a 7.6
to
50 so
50
Golden Shiner s°C 1a°c moramme 5°C 27"C
<01
17.11 21 757
1 .67 z .6s
700 loo
1 .61 70 1.
100 1110
Channel (<07049 rrers nmd~e °c 2d° C Chioramine 5°C 1h°C
0,02 O .OZ
103 l03
loo l00
027 o .2
I0 0 ias
10< loo
Table 2 presents the lethality data for this species . It is evident that of the three temperatures tested, the trout showed the greatest sensitivity to free chlorine at 12°C (i .e . LC5o's were significantly lower for some of the time intervals) . With exposure times longer than 72 hours, temperature had little if any effect on the lethal peak concentration . The monochloramine tests were run for longer times than the free chlorine tests in order to reach an asymptote on the mortality curve . By comparing LC5o's at similar exposure times, it is seen that free chlorine is 8-14 fold more toxic than the combined form (Table 2) . 41
Table 2 . Tontdty of i.-ft- chlorine exposure to rainbow trout (Salmo gird-d) All cananteations were meuured at peak of chlorine pulse . Parentheses endow 95% wnfidence interval . Median Lethal Concentration of Free CNodne (mg/Ls
Hours 24 32 48 56 72 96 104 (20 ( 28 140 166
5°
12°
0 .294 ( .273- .317) 0 .231 ( .207- .262) 0 .162 ( .142.180) 0 .179 ( .I35-.312) 0 .103 6084-426) 0 .082( .064-.1011 0 .0626064- .100) 0 .074 ( .038-.091) 0 .074 ( .056-0.911 0 .069 ( .052-.087)
0.258 (.231-.297) 0308 (.178-.259) 0.090 . ( 082.103) 0.078 (.069-.089) 0.069 076) 0.062(.055-.IOU) 0.060(.052-267) 0.052 (.045-.060) 0.059 (.050-.070) 0.059 (.050-.070) 0.054 (.045-.065)
Medun Lethal Concentration of (7NOramine at 12°C (mg)L)
17°
Houn
0 .263 ( .217-.319) 0 .157 184) 0 .124 ( .107-.146) 0 .100 ( .085-.116) 0 .074 ( .062-.091) 0 .095( .062-345)? 0 .089( .059-147) 0 .089 ( .059-.24 /) 0 .071 ( .051-1 .44) 0 .066 ( .049-.109) 0 .066 ( .049-.109)
120 130 145 170 191 214 240 288
Death
Equilibrium lose
0.75 (.70-1 .06) 0.71 (67-.82) 0.69 ( .65-76) 0.66 ( .63-.70) 0.65 ( .62-.69) 0.62659- .66) 0 .59( .56-.63) 0 .58 ( .55-.62)
0.72 (.68- .871 0.70 (.67- .78) 0.67 ( .64- .72) 0.66 (.61- .68) 0.64 (.61- .68) 0.60657- .64) 0.59(.55- .63) 0.58 (.54- .62)
LL
Median Lethal Time For the Peak Connenhatfone Indicated ce.meeoatpn5°C (TomdFree) His 0 .22/0.11 0 .30/0.20 0 .40/0.30 0 .40/0.31 0 .50/0.38 0 .60/0.46
Cnnc .12°C (Total/Free)
Coot.Irc (Total/Free)
Hn
73 (68-76) 0.20/0.07 48 .4(45 .6-51 .2)0.20/0.08 20 (1640) 0.28/0.13 29 .3 (27.330.)) 0.36(0.29 11.7 (9.8-13 .5) 10.5 (9.4-12 .2)
\
Hrs
73 (67-76) 0.41/ .03 60 (546-65 .110 .30/.03 32.3 (30.7-34 .3) 0 .38/.12 26.8 (24.9.281) 0 .30/:17 0 .40/.25 0 .20/.11
Concentration
374 (221-1290) 206.5)171-313) 42 .4 (38.8-46 .4) 36 .6 (34.438 .4) 25 .7 (24.226 .8) 25 .8 (23 .7-27 .8)
I
0 .5
I
I
I
I
I
o-5°C 0-12 - 17
0 .6
z
I
I
I
endeat8 302(282-339) 237 (240.288) 118 (138-155)
I
d
I
Chbramine
I I
I 20
I
I
1
40
I GO
1
I
Free Chlorine o
0 O
4
.2
0
1
1 1 1 1
0
3
.4
•
I
1
1
1
1
1
1
I
140
1 160
At the shorter time intervals, overall sensitivity of the salmon fingerlings (Table 3) to free chlorine pulses was comparable to that of the rainbow trout . But with longer exposure times (i .e ., 48 hrs and longer), the salmon seem more resistant than the trout . This is reflected in both
6°C
12°C 0 .297(170-.341) 0 .262 6207-194) 0 .157 .(146-.169) 0 .139( .127-.151)
0./98(.145-.220) 0.181 (.146-.256)
0 .119 (.108-.129) 0 .093(.082-.111) 0 .093 (.080 ..113) 0 .084 (.074-.095)
.2
0
1 120
Coho Salmon (Oncorhynchus kisutch)
24 32 48 56 64 72 144 (53 168
A 0 i
I
Fig. 3 illustrates the free chlorine response isopleth for rainbow trout . These data suggest a rather high sensitivity of trout to free chlorine so that in a week of exposure to 0 .04 mg/ L pulses, approximately io percent of a population would be killed . Pulses of 0.15 mg/ L would kill all the exposed fish. It should be noted, however, that when no free chlorine was present, no mortality occurred in total concentrations as high as 0 .4 mg/ L .
MedunLethalConcentratioeFreeCNOriee(ms/L)
0
1
100
Fig. 3 . Response isopleth for rainbow trout at 5°C exposed to chlorination pulses every eight hours . The lines connect points of similar predicted percent mortality for various concentrations of free chlorine pulses and exposure times .
Hours
A
1
80
Table 3. Tontcity of intermittent chlorine expomre to coho vlmon ( Oein hynchue kirutch). All concentrations were measured at peak of 0NOrine pulse . Parentheses enclose 95% confidence interval .
p 0 .4
1
20 40 60 100 200 MEDIAN LETHAL TIME (HRS)
1
1
1 1
400
Fig. 2 . Median lethal times of rainbow trout exposed to chlorination pulses every eight hours, these containing different proportions of free chlorine (vertical axis) . The numbers above the individual points indicate the total chlorine concentration. 42
I 0
0
0
0 10
O
TIME (HRS)
a
J 2 U w 0 .3 w X U- 0 .2 J 0 .1
0 .1-
HO
0 .4/0 0.5/0 0.6/0 0.7/0
LC50's computed on equilibrium loss are also presented in Table 2. These show a negligible difference compared to the death LC5o's . In Fig. 2 the median lethal time is plotted against free chlorine where various ratios of free to combined chlorine (monochloramine) were pulsed through the test tanks . It is evident that when there is any free chlorine present at all in the pulse, the lethality can be most accurately predicted from this concentration . For example, there were several widely separated median lethal times associated with 0,4 mg/L total chlorine, but at the progressively higher concentrations of free chlorine, the survival time was shortened logarithmically . Total chlorine was not an accurate predictor of lethality unless the proportion of free chlorine in the mixture was constant .
0 .7
w w CC
Median Lethal Time For Peek Cancennatioe Indicated Concen (To%VFeea-) 0.38/0.33 0.31/036 0.30/033 0.39/0.30 0.2410.19 01110.15 0.29/0.22 0.28/0.20 0.18/0.13 0.20/0.12 0.14/0.10 0.19/0.09
R" (6*)
Hrs (12°) 20 (16 .7-22.4) 26 (2530) 28(26-29)
35(32-37) 41 (3914) 50 (47-53) 57 (52.62) 60 (55.66) 52(49-56) 98(90-105) 125 (118 .135) 159 (145-191)
MedianLethalConcentntknChloramine(mg/L) Houn
6°C
80 96 104 120 (44 152 166 176 192 215 248 287 336
0.677 ( .621-.856) 0.6)9(577-.672) 0.537(188-.615) 0528 (.476-.600) 0.488(.437-.533)
12°C 0 .683 0 .640 ( .535 1 .9) 0 .609 ( .501-.886) 0 .573( .471-.700) 0 .5536436-.685) 0 .516 (.416-.618) 0 .505 0.409 0.463 0.445 0 441 0.551
Median Lethal Tun . For Peak Concentration Indicated
Concenmation 0 .4 0 .5 0 .48 0 .59 0 .69
Hn(6°)
Hn(12°) 602 (412-
1
301 (200-1027) 130 (121-136) 98(94-103)
173 (164-183) 119(113-125) 84 (79-89)
Table 4. Toxicity of Intermittent Chlottne Exposure to CNnnel Catfish (lcm/umr facvudt) . AU -w-t- were mearrd at the peak of the dtotine purse . par-h- endow 95% confidence intern . Medun Lethal Conanvatbn---
Median Lethal Conanvation Free Chlorine (mg/L) 5°C Hours 48 56 72 96 120 144 152
5`C
24 °C Hop"
0100 (0.149) 0.1521 .105 . 1 0.120 ( .MB-.17q 0.082 ( .027-.137) 0.050 ( .044-067) 0.033( .062) 0-033 ( -062)
70 72 96 120 t44 152 168
0 .093 ( 0 .064 ( 0 .051 ( 0 .0321 0.030) 0 .025 (
.022- .143) .026 .145) .011- .082) .012.043) .019- -054) .021- .042)
80 96 120 144 168
24°C Hours
Hoare 0 .143 (0 .112-0.169)
( .&/L)
) 0 .3131Q1410.275 ( .2061 0.234 f .185 .330) 0.217 ( 769.253) 0.208 ( .165 :450)
48 72 96 120 44 168
1411 (0.3541 0.328 ) .206+541) 0160 ( .112- .461) 0246 ( .11&.735) 0 246 1 .1 15-.735) 0 241 f .072-. 90)
Median Lethal TiConcentrarien (Total/Ftee)
Hoore
C---i. n (TotafFrea)
0.14/0.04 134(126 .14510 .17/001 0.17/0.07 106 (101-112) 0 .22/0.05 023/0.12 70 (66-75) 026/072 018/0.17 51 (47-541 0 .31/077
Home
CWoremine C.- (-IL)
LT 50 (h)
Chloramine LT 50 COrn (mg/L) (h)
213(176-351(0Ilon truth 0.16 325(215-1479) 1231116-130) 0 .15 224 (190-2398) 024 180 (148-254) 43 (39481 010 169 (162251) 0.34 67 ( 68-71) 22 (19-25) 0 .29 91 (8695) 0.44 51 ( 47-54)
Channel Catfish (Ictalurus lacustris)
0
140 180 20 40 00 80 100 HOURS EXPOSURE
220
200
Fig . 4 . Response isopleth for coho salmon at 12°C exposed to chlorination pulses every eight hours . The lines connect points of similar predicted percent mortality for various concentrations of chlorine pulses and exposure times .
the LC50's (Tables 2 and 3) and in the zero percent isopleths . At 14o hrs the highest values for no death were around 0.03 and 0 .05 mg/ L for the trout and salmon, respectively (Figs. 3 and 4) . There was a slight, but statistically significant, effect of a 6'C difference in temperature on the free chlorine toxicity to the salmon (Table 3) . The 72 hr LC5o at 6° was o .181 and at 12° it was o . i 19 mg/ L . A six-fold difference in toxicity to coho salmon is evident between the free and combined forms of chlorine . For example, the 144 hr LC5o's at 12'C were 0 .093 and 0 .553 mg/ L, respectively (Table 3) . Chloramine pulses of 0 .38 mg/L for one week resulted in no death . This was well above the concentration of free chlorine causing loo% lethality (Fig . 4) . One batch of cohos obtained from the same source were four times as sensitive to chlorine as those reported on here . The hemoglobin analyses indicated a prevalence of anemic fish in this batch .
Table 4 presents the LC5o and LT5o data for free chlorine and monochloramine toxity to the channel catfish . Two widely separated temperatures were tested (5 and 24'C) . At the higher temperature the catfish may have been slightly more sensitive than at 5'C to free chlorine pulses, however, the LC5o's are not significantly different . At higher concentrations of free chlorine the median lethal times (LT5o's) were statistically shorter at 24'C indicating a more rapid toxic action . Overall, the median lethal concentrations for the channel catfish were remarkably similar to those of the rainbow trout (Table 2) . However, the free chlorine response isopleth of the channel catfish is narrower than that of the trout suggesting a rather narrow lethality threshold (Fig . 5) . Moreover, the concentrations resulting in no catfish mortality are lower than those of the trout so only a trace of free chlorine in the chlorination pulses may cause some mortality to this species of catfish . There is virtually no effect of temperature on monochloramine toxicity to the catfish (Table 4) . Chloramine was less toxic than the free chlorine, and the response isopleth was comparatively wide . The concentration resulting in no mortality after loo hrs was approximately 0 .12 mg/ L compared to the trout figure of 0 .4 mg/ L chloramine and the 120 hr LC5o's were 0 .25 and 0.75 mg/L for the catfish and trout, respectively . Thus, the channel catfish fingerlings were more sensitive to chloramine than were the rainbow trout .
Carp (Cyprinus carpio) The data presented in Table 5 indicate that carp are quite resistant to chlorine compared with the other species tested in this study . The 166 hr LC50 for free chlorine at 6°C was 0 .245 and for chloramine 1 .19 mg/1, an almost 43
0.7
Table 5 . Toetooty of Interminem Chlorine 0,00, .!. to Carp (Cypinus rerpfo) .All concentrations were meamtId at the peck of the III-- pulse. Parenthesis enclote 95% confidence interval.
Median Lethal ConcentmHOo (mn/L)
0 .6
6°C Hoors
24°C
FmeChlwine
Chloramine
0 .538(.399-_983) 0 .400(.339627) 0 .3311.293 .402) 0 .283( .250 .321) 0 .245 (,214 .273)
16 120 144 152 166
H .Marine I4.e
1 .72 1 .60 1 .40
0 .403( .346-.504) 0 .278 (.238-126) (3 .219 C185 210) 0 .2191.185-.260) 0 .219 (-185 :260) 0.219 (.185-.260)
48 72 96 120 166
1.19
No death i 0.4/0. .1 No death in 1 .1 chlaramine Note : Apptoxhnat,ly 30% of tlm 24°C Carp weae anemic .
0
Median Lethal Time (hours)
0 .2
Concenttation(TOmUFree6°C) 0 .38/0.24 0.45/0.31 0.52/0.35 o death at 0.29/0.15 1 .44/0 1 .44/0 1 .74/0
0 .1 0 .25
I
1
Hours 153 (145-166) 1411: ' 11 -159) 141 (5-149)
135 (141-176) 194 (909-10)) 94(90 100)
',, 0 .20 E Lid z
0 .15
it 0 J
u
0 .10
w 0.05
0
20
40
60 80 100 120 HOURS EXPOSURE
140 160
Fig . 5 . Response isopleth for channel catfish at 24°C exposed to chlorination pulses every eight hours. The lines connect points of similar predicted percent mortality for various concentrations of chlorine pulses and exposure times .
Interestingly, the temperature effect was reversed in chloramine so that toxicity was greatest at the lower temperature and, moreover, the difference between temperatures grew larger with longer time intervals (Table 6) . Examination of the response isopleths in Figs . 6 and 7 indicate a greater difference in toxicity between free and combined chlorine at 24°C than at 5°C . It is noteworthy that the highest concentration of free chlorine resulting in no death was the same at both temperatures (0 .06 mg/ L) but differed by o . i mg/ L for chloramine at the two temperatures .
Discussion 5-fold difference between the two chlorine forms (Table 5) . Even at concentrations of 1 .74 mg/ L chloramine it required 96 hours of exposure before 50% mortality occurred . No death was seen in 0.15 mg/ L peak free chlorine . Data from the 24'C acclimated fish tests are included in Table 5 but these figures may have an important limitation . It was discovered after the tests that approximately 30% of the control fish were anemic (not true of the 6°C fish) . They seemed healthy otherwise but the low hemoglobin levels may have contributed to the greater 240 C sensitivity .
The results of these bioassays show that temperature has relatively little effect on the lethal concentrations of either free chlorine or monochloramine to the five fish species tested . (This conclusion should not necessarily be extended to conditions where a constant chlorine concentration is used .) Even though there were some statistically significant differences attributable to temperature, these
Table 6. Toxicity of Intermittent Thloriim Exposure to Golden Shiners (NOrcmiyonus e,ymfeucas). All concentrations were Measured It the Peak of th,CM1btlne Pals, . Parentheu, enclose 95% confidence inter .!, Median Lethal C000,ot,atio, Ft., Chlorine
Media, LethalConcentratioyChloramine
Hoors 5°C 24°C Honor
Golden Shiner Notemigonus crysoleucas) Golden shiners were intermediate in sensitivity to intermittent chlorine compared to the other species tested . For the shorter time intervals in free chlorine (less than 12o hrs) there was a significant temperature effect in that the higher temperature caused a lower LC50 (Table 6) . 44
30 48 56 72 96 20 144 168
0,84 1 .628-1 .158) 0,550 ( .502- .639) 0 .302 /467 .561) 0 .388 (.355 . .428) 0 .269 ( .237- .299) 0 .205 ( .170- .233) 0 .181 ( .145- .209) 0 .162 (.129- .188)
0.257 (127 .290) 0122 ( .192 . .252) 0.212 ( .180- .245) 0.217 (180-,245) 0.193 ( .164-128) 0.182 ( .154-.214) 0.177 ( .150,209) 0 .177 ( .150.209)
48 72 96 120 144 168
3°C 24°C 0.993(.928-1 .077) 0.871 (.816- .927) 0.724 (.639- .790) 0,763 (.605- 735) 0.644 6542- .814) 0.533 6298 .61))
Median Lethal Time Conce (Total/Free) n 02110.16 0.30/0.26 0.39/0.35 0.50/0.46 0.61/056
Hre 138(130155) 95 (9099) 70(6575) 64 (5970) 48(4431)
Can (Total/Ft,e) n 0.16/0.08 0.24/0.15 0.34/026 0.44/0.75
1 .0941.985-1 .521) 0 .9791.905-1 .119) 0 .9701.869-1 .014) 0.921 006) 0.921 (.858-I 0061 0.921 (.858-1 .006)
Madian Lethal Time
Hts. 1%dea1h190(it, 256 (190440) 34 .5(294-00.0) 22 .3 (19 .4.24.8)
Con«nation (Total/F-) Hs 0 .51/0 0 .72/0 0 .91/0 1 .12/0
177(158225) 110 (102-120) 63(5668) 33 (29-36)
Con t,ehnn (Tam!/Fur) 0 .61/0 0 .80/0 1 .01/0
teen 2884 353 (2351030) 67(6076)
i
w z Cr
0 J U -~
1 .2 1 .1 1 .0 0 .9 0 .8 0 .7 0 .6 0 .5 0 .4 0 .3 0 .2 0 .1 0
1 .0-
0-0-0
o70 0/.
C
E d 0
o-o--0 20
0 .8-
z
U \_a1_0-0----00
"I
01~ -0-0100% 9-6-e 70
m IV 4)
. *_ ~0-0
S L_
o
20 LL t *0 U 0 I I I I I I I I I I 0 20 40 60 80 100 140 180 HOURS EXPOSURE •
0 20 406080 100 140 HOURS EXPOSURE
180
Fig. 6 . Response isopleth for golden shiners at 5°C exposed to chlorination pulses every eight hours . The lines connect points of similar predicted percent mortality for various concentrations of chlorine pulses and exposure times .
were never more than a doubling of the LC5o with a decrease in temperature and were usually of less magnitude than this . Small temperature effects on intermittent chlorine toxicity were also observed with bluegills (Bass & Heath, 1977) and rainbow trout (Brooks & Seegert, in press) . Brooks & Seegert (in press) exposed their fish to either a single 30 minute chlorine dose, or a series of three 5 minute doses separated by 3 hours each, whereas, Bass & Heath (1976) used the same regime as was used in my study . The yellow perch, however, exhibited a marked temperature schlorine interaction (Brooks & Seegert, 1977 . Dickson et al. (1977) exposed goldfish intermittently for 24 hours to free chlorine and combined at various frequencies of dosing, different temperatures, and various lengths of dose time . There was a relatively small effect of temperature, although the range of temperatures tested was small . After reviewing several chlorine investigations, including their own, Brooks & Seegert (1975) concluded that for fish there is a species specific range of temperatures over which chlorine toxicity is little affected . But outside this range, elevated test temperatures or thermal shocks may alter the resistance to chlorine (Stober & Hanson, 1974) . Some temperatures (e .g . > 20'C for trout) are harmful alone, so in combination with a toxicant they may be synergistic (Thatcher et al., in press) .
0-0
Fig. 7 . Response isopleth for golden shiners at 24°C exposed to chlorination pulses every eight hours . The lines connect points of similar predicted percent mortality for various concentrations of chlorine pulses and exposure times .
A major finding of this study was a much greater toxicity of free chlorine compared to monochloramine . It is difficult to compare other studies in this regard because of the tendency of most workers to give only total residual chlorine values . Also most of the other investigators have used a constant concentration rather than a pulse regime . In the most directly comparable study, Dickson et al. (1977) concluded that chloramine was slightly more toxic than free chlorine . However, they apparently had 25% dichloramine (or trichloramine) present which is more toxic than either free chlorine or monochloramine (Heath, unpublished observations ; Holland et al., 1960) . This same phenomenon (i .e . presence of higher forms of chloramine) may help explain the contradictory results in the literature on the relative toxicities of free and combined chlorine. Another potentially important factor is the receiving water pH . As pH is raised above 7.5, OCI predominates more over HOCI . The former has less disinfection ability (White, 1972) and, therefore, may be less toxic to non-target organisms, although this has not been systematically investigated . It is noteworthy that Dickson et al. (1977) used pH 7 .85 water while my dilution waters were generally below pH 7 .6 . The differences in toxicity between free and combined chlorine that I observed range from three to fourteenfold, depending on fish species and exposure time . This is, for the most part, a greater difference than has been observed in continuous exposure studies or single short45
term exposures (Brooks & Seegert, 1975 ; Holland et al.,
in that the differences in concentrations between that
1960 ; Rosenberger, 1971 ; Merkens, 1958 ; Tompkins &
causing no death and that resulting in 100% mortality are
Tsai, 1976) . At present I do not have a good explanation
fairly small, particularly for free chlorine . This difference
for this . A critical factor may be the opportunity for
becomes even smaller with longer exposure times so if
some or complete recovery between chlorination pulses
fish stay in the effluent area, it can be predicted that should
which would not be true for continuous exposure . Since a variety of fish species were used in this study,
a few fish get killed by chlorine, a small increase in the
their sensitivities to chlorine can be compared . For free
lation . Conversely, a small decrease in concentration
chlorine, rainbow trout and channel catfish were quite similar with the trout slightly more sensitive to the higher
would be very beneficial . The isopleths and LC50 data indicate that intermittent
concentrations . In order of decreasing sensitivity, the
doses of monochloramine would require very high con-
others were coho salmon, golden shiner, bluegill (Bass &
centrations to cause much mortality if the fish stay in the
Heath, 1977) and carp . Their relative sensitivities to chlo-
area for short intervals of time, whereas free chlorine
ramine were different, however . In order of decreasing
doses would result in considerable mortality at much
sensitivity, they were channel catfish, coho salmon,
lower chlorine concentrations . Therefore, one approach
golden shiner, rainbow trout, and carp .
to reduce impact that deserves consideration would be
The high sensitivity of channel catfish to chlorine was surprising since it is a warm water species and these are
ously with the chlorination dose in situations where the
chlorine concentration may eliminate the resident popu-
the perfusion of ammonia into the effluent simultane-
not generally as sensitive as the salmonids (Brungs, 1973) . So far as could be seen, the catfish were healthy ; they also
effluent had a high percentage of free chlorine . According
fed well and had normal hemoglobin levels . In a separate
almost instantaneous and if the ammonia is present in
investigation in this laboratory using catfish from the
excess (i.e . more than equimolar) the formation of di-
same stock, they were found to be intermediate in sensi-
chloramine (which is highly toxic) is suppressed .
to White (1972) the formation of monochloramine is
tivity (compared with trout, golden shiners, goldfish, and
Calculating median lethal concentrations and response
bluegill) to copper and cyanide, but most sensitive to
isopleths on the basis of free chlorine, as was done in this
chromate . Therefore, they were not a grossly abnormal
paper, may seem to overemphasize the toxicity of chlo-
group of fish . It points up the marked species variations
rine and is, therefore, a conservative approach . The total
that cannot be predicted from their ecology . It also casts
residual was actually 10-50% higher in the test tanks .
into question the efficacy of proposed water quality crite-
However, where there was any free chlorine present, pre-
ria for chlorine which apply differentially to cold water
diction of lethality on the basis of total residual was poor
and warm water species (Brungs, 1973) . Brungs (1976)
unless the chlorine peak was 0 .5 mg/L or higher . There-
later modified this proposal in the light of the high sensi-
fore, when proposing water quality criteria for inter-
tivity of some warm-water minnows to chlorine .
mittent chlorination regimes, it would seem advisable to
The data incorporated in the response isopleths may be useful in assessing potential impact of intermittent chlo-
have separate ones for free chlorine and monochloramine .
rination . For example, if it can be established how long certain species normally stay in a discharge canal, the percent mortality, if any, of the population may be predicted
Acknowledgements
from the isopleth . At present there is mostly anecdotal evidence regarding residence time (Basch & Truchan,
This study was supported by a grant from the American
1976 ; Brooks & Seegert, 1975) and a great need for syste-
Electric Power Service Corporation . Mr . Kim Koller
matic studies although these would be difficult to carry
provided excellent technical help with this project . I thank
out .
the Wytheville National Fish Hatchery and the Platte
In lake situations where fish will probably not receive more than two or three chlorine exposures at most before moving out of the area, the isopleth would indicate fairly high concentrations of either free (< o .1 mg/ L) chlorine or monochloramine (< 0 .2 mg/ L) would be tolerable . The response isopleths suggest a fairly sharp threshold 46
River State (Michigan) Anadromous Fish Hatchery for providing fish .
References
Basch, R. E . & Truchan, J . G . 1976 . Toxicity of chlorinated power plant condensor cooling waters to fish . U . S. Environmental Protection Agency Ecological Res . Ser. EPA-600/3-76-009 115 P . Bass, M . L. & Heath, A . G . 1976 . Toxicity of intermittent chlorination to bluegills : interaction with temperature . Bull . Env . Contam . and Tox . 17 : 416-423 . Brooks, A. S . & Seegert, G . L . 1975 . The toxicity of chlorine to freshwater organisms under varying environmental conditions . In R . L . Jolley (ed) . Proc . of Conf. on Environm . Impact of Water Chlorination . Oak Ridge Nat . Lab . CONF-751096 pp . 277-298 . Brooks, A . S . & Seegert, G . L . 1977 . The effects of intermittent chlorination on selected great lakes fishes ; effects of 5- and 30minute exposures on rainbow trout and yellow perch . Trans . Amer . Fish . Soc. 1o6 : 278-286 . Brungs, W . A . 1973 . Effects of residual chlorine on aquatic life . J . Wat . Poll . Contr . Fed. 45 : 2180-2193 . Brungs, W . A . 1976 . Effects of wastewater and cooling water chlorination on aquatic life . U . S . Environmental Protection Agency Ecological Res . Ser . EPA-600/3-76-o98 .45 P • Chaney, A . L . & Marbach, E . 1962 . Modified reagents for determination of urea and ammonia . Clin. Chem . 8 : 130-132 . Dickson, K . L ., Hendricks, A . C ., Crossman, J . S . & Cairns, J . C . 1974. Effects of intermittently chlorinated cooling tower blowdown on fish and invertebrates . Environ . Sci . and Technol . 8 :
White, G . C . 1972 . Handbook of Chlorination . Van Nostrand Reinhold, New York . White, G . C . 1975 . Current chlorination and dechlorination practices in the treatment of potable water, wastewater, and cooling water . In R . L. Jolley (ed) . Proc . of Conf. on Environm. Impact of Water Chlorination . Oak Ridge Nat . Lab . CONF751096, PP-7 -24 .
845-849 . Dickson, K . L ., Cairns, J . C., Gregg, B . C., Messenger, D . I ., Plafkin, J . L . & Van der Schalie, W . 1977 . Effects of intermittent chlorination on aquatic organisms and communities . January J. Wat . Poll . Contr. Fed . P . 35 - 4 4 . Draley, J . E . 1973 . Chlorination experiments at the John E. Amos plant of the Appalachian Power Co . April 9-to, 1973 . Argonne Nat . Lab . Rept . No . ANL/ES-23 . Environmental and Earth Sciences, Argonne, Illinois . Finney, D . J . 1971 . Probit Analysis . 3rd Ed . Cambridge Univ. Press, New York . 333 P . Holland, G . A ., Lasater, J . E ., Newman, E . D . & Eldrige, W. E. 1960 . Toxic effects of organic and inorganic pollutants on young salmon and trout, p . 198-214. Res . Bull . No . 5, Dept . of Fisheries, State of Washington . Merkens, J . C. 1958 . Studies on the toxicity of chlorine and chloramines to the rainbow trout . Water Waste Treatment J . 7 :150-151 . Rosenberger,'D . R. 1971 . The calculation of acute toxicity of free chlorine and chloramines to coho salmon by multiple regression analysis. Thesis, Michigan State Univ ., East Lansing, Mich . 33 P. Stober, Q . J. & Hanson, C . H . 1974. Toxicity of chlorine and heat to pink (Oncorhynchus gorbuscha) and chinook (0. tshawytcha) salmon . Trans . Amer . Fish . Soc . 1o3 : 569-576. Thatcher, T . 0 ., Schneider, M . J . & Wolf, E . Q . (in press) . Bioassays on the combined effects of chlorine, heavy metals, and tempearture on fishes and fish food organisms . Part I . Effects of chlorine and temperature on juvenile brook trout (Salvelinus fontinalis) . Bull . Env . Cont . and Tox . Tompkins, J . A. & Tsai, Chu-Fa . 1976 . Survival time and lethal exposure time for the blacknose dace exposed to free chlorine and chloramines . Trans . Amer . Fish . Soc. 10 5 : 313-321 . 47