This is Part II of a study investigating the disinfection of swimming pools.
 Part I appeared in March, 1970. The results show that the sanitation
 of swimming pools can be improved, and that free available chlorine
 is a better disinfecting agent than combined chlorine. Other
 findings are presented.


A. P. Black, Ph.D.; M. A. Keirn, M.S.; J. J. Smith, Jr., B.S.; G. M. Dykes, Jr., B.S.C.E.; and
F. E. Harlan, M.S.P.E.

 THERE has been a rapid increase in             of 1968-four of the authors traveled
    the number of public swimming pools         over 2,200 miles sampling 193 pools,
 constructed in this country in recent          ranging in size from 11,000 gallons to
 years, and their proper disinfection and       over 600,000 gallons, located along the
 operation is becoming an increasingly          Atlantic and Gulf coasts of Florida.
 important public health problem. The           They included the two Hall of Fame
 results of a recent study of 193 typically     pools at Fort Lauderdale, and all read-
 operated public pools indicate the need        ers of this paper will be glad to know
 for the promulgation and enforcement           that both met all of the standards used
 by state or local health departments of        for the study.
 adequate codes of practice covering all
 aspects of pool operation, and, hopefully,     Distribution by Pool Size
 of minimum qualifications for their
operators. In the preparation or revi-             Of the 193 pools, 7 had a capacity
sion of these codes, it is extremely im-        under 20,000 gallons and 7 were above
portant that critical terms be exactly          300,000 gallons; 52 were between
defined. Specifically, for each disin-          20,000 and 40,000 gallons; 52 between
fectant, the type of residual required          40,000 and 60,000 gallons; 23 between
should be clearly stated, and also the          60,000 and 80,000 gallons; 13 between
method or methods approved for its              80,000 and 100,000 gallons; and 39
determination.                                  were between 100,000 and 300,000
   The objective of this study was to           gallons.
sample many different types of public           Distribution by Pool Use
swimming pools in order to get a sta-
tistically sound evaluation of the safety,         Of the 193 pools, 108 were located at
operational practices, and maintenance          hotels or motels along Florida's "Gold
problems encountered in typically oper-         Coast," including a great many of the
ated pools. During the study, which             most beautiful establishments in this
was conducted in two parts-one in Au-           highly tourist-oriented area; 41 were
gust of 1967 and the other in August            located at large apartments or condo-

740                                                               VOL. 60, NO. 4, A.J.P.H.
                                                      DISINFECTION OF SWIMMING POOLS

miniums; 26 were municipal pools; 11                 1. Very slight haziness observable in deep
wvere YMCA or YWCA pools; and 7                    water
                                                     2. Main drain not clearly visible
were located at country clubs.                       3. Main drain visible only as a dark spot,
                                                   water milky
Distribution by Method of Disinfection               4. Unable to see main drain.
  Of the 193 pools, 88 were disinfected               A Beckman Model N battery-oper-
with chlorine gas; 49 with sodium                  ated pH meter was used to measure pH.
hypochlorite; 26 with calcium hypo-                Free available chlorine and total chlorine
chlorite; and 28 with cyanuric acid-               residuals were measured by am-
stabilized chlorine. In one pool, no treat-        perometric titration and also with an
ment of any kind was being used.                   orthotolidine test kit, both at pool water
                                                   temperature and, in the 1968 study, on
Distribution by Filter Type                        samples chilled to approximately 10 C.
                                                   Additional parameters used in the 1968
   Of the 193 pools, 56 had sand fil-              study were alkalinity and cyanuric acid
ters; 49 pressure diatomite filters; 76            level. Cyanuric acid was determined
vacuum diatomite filters; 8 had cart-              turbidimetrically using a commercial*
ridge filters; 2 had anthracite filters            test kit. Total alkalinity was measured
and 2 were flow-through pools.                     with a commercial test kit employing a
   The only significant difference be-             25 ml sample.
tween the two studies was that fewer
motel pools and more apartment pools
were used on the 1968 trip. Relative dis-          Bacteriological Parameters Used
tribution of capacity and type of disin-              Bacteriological parameters used were
fection were the same. Two anthrafilt              coliform density-both by the multiple
pools were tested in 1967 and two flow-            tube technique and membrane filtra-
through type pools in 1968.                        tion-fecal streptococci, total staphylo-
                                                   cocci, Pseudomonas aeruginosa, and 48-
Materials and Methods                              hour standard plate count. Two samples
  The following parameters were eval-              were taken from each pool. A 100 ml
uated or recorded for the 193 pools:               sample was taken in a bottle containing
                                                   sodium thiosulfate and sent to a board
  1. Indicator organisms and bacteriological       of health laboratory for coliform de-
quality                                            termination by membrane filtration.
  2. Level and forms of chlorine
  3. Precision and accuracy of the orthotoli-      This provided a check on results ob-
dine test                                          tained by the Earle B. Phelps Sanitary
  4. pH and water temperature                      Engineering Laboratory of the Univer-
  5. Turbidity (see below)                         sity of Florida, which made the other
  6. Type of filter
  7. Type of pool surface                          bacteriological determinations.
  8. Type of algae. if present                        During the 1967 study, a 300 ml
  9. Pool capacity, circulation and bather load    composite sample from the shallow end
 10. Pool operation in general.                    of each pool was collected for the bulk
  In addition to the above, a record               of the determinations in a Roux bottle
wvas made of weather conditions at the             sterilized between samples in a strong
time each sample was collected.                    hypochlorite solution, then rinsed
  Turbidity of pool water was rated                thoroughly in pool water. This bottle
on the following arbitrary scale:                  was attached to a special rod and pulled
  0. No visible turbidity, pool water sparkling,     * Manufactured by the Hach Chemical Co.,

main drain outline sharp                           P. 0. Box 907, Ames, Iowa 50011

APRIL, 1970                                                                                741
around the shallow end of the pool at a
depth of approximately 18 inches. A
                                                                                                c"     cr,   ON "                      T
300 ml portion was decanted into a                                                              cl     Cl4   cl cli                    r-

sterile BOD bottle containing thiosul-                0
                                                      0                        Al

fate. This procedure was simplified in                04-                                       csi    c     cq            I       t       -

the 1968 study by the use of sterilized               6

BOD bottles pulled through the water                                               biD
in the above manner. Samples were sent
to the university laboratory each night        S
by bus and were processed within 24            CL     a)
hours.                                         0     ,--
                                                                                                cl     LO        c0                    H
                                                                                                                 -                     z
   The confirmed multiple tube test for        0
                                               S                        ._~
coliforms and MPN for fecal strepto-           0                       O>4--
                                                                                                 ocq in                  o
cocci were run by techniques from the          0

12th edition of Standard Methods for           0
the Examination of Water and Waste-            0                       0t

water, and 48-hour plate counts were           0
run on tryptone glucose yeast extract          S
agar at 350 C. Staphylococci were              S
                                               0      04
enumerated by membrane filtration of                  0
                                                                                                C)     C-
                                                                                                       -         h-
                                                                                                                 -         co          o
100 ml of sample and incubation on              0     6

m-Staphylococcus broth (Difco). Pseudo-         CL
monas aeruginosa was determined by              0                                                                                                    w.
the most probable number method, de-                                                                                                                 c

scribed by Favero, Drake, and Randall,'         0
using asparagine enrichment broth fol-          S                                                                                                    a

lowed by confirmation in acetamide              S

broth.                                          S
                                                                   r. ;'- t                                                                    I     t
                                                                                                co     co        cO        Cl

Results                                         0
                                                                :-4      0         0-

Bacteriological Quality of Pool Water           0
                                                S                        o
   Table 1 summarizes the bacteriolog-                          tn
                                                                                                U.n              Cl ,      C
ical results from all pools. Die-off studies    S               0

showed that in many instances plate            C-
                                                0               -0-4                             't,
                                                                                                             ~        ~~       ~   ~   ~

counts would tend to increase and num-                           c

bers of staphylococci and fecal strepto-        E           4
                                                                                   0                                    CC_                          E
cocci to decrease, on standing for 24           2                                                LrD
hours. Pseudomonas aeruginosa and              Cl)
Escherichia coli did not show significant
fluctuations when tested. Therefore, plate      0                                              (10

couInt, staphylococci, and fecal strepto-       S

cocci data can only be used for com-                                                                                           0                     -
parison between groups in this study.                                              C:4     r.

Coliforms and membrane filter results,                                             J                   AlV
however, appear to be valid as maximum
storage time allowable for membrane                                                            IL-
                                                                                          _0 ce-
filter tests is 30 hours, and samples were
run well within this time limit.                                                   .7
   With respect to the presence of coli-

742                                                                                      VOL. 60, NO. 4, A.J.P.H.
                                                       DISINFECTION OF SWIMMING POOLS

forms and to plate couInts, the APHA                filters, even though the latter group con-
recommends the following:2                          tained 0.3 mg/l free available chlorine.
   1. Not more than 15 per cent of samples             A comparison of pools having 0.3
examiinled shall showv the presence of con-         mg/l free available chlorine, measured
firmiied coliforms in any of five 10 ml portions.   by amperometric titration, with those
   2. Not more than 15 per cent of the samples      having 0.3 mg/l total chlorine but less
examiiined shall contain 1.0 coliform organisms     than 0.3 mg/l of free available chlorinie,
per 50 ml when the memhibrane filter test is
used.                                               showed that free available chlorine is
  3. Not more than 15 per cent of the samples       much more effective than combined
examiiined shall contain over 200 bacteria per      chlorine. Fewer isolates of all indicator
milliliter.                                         organisms were found in pools in which
                                                    the principal disinfectant species was
   Of the 193 pools reported on, 123 or             free available chlorine than in pools
65 per cent contained no coliforms by               where the major portion of the total
multiple tube and were satisfactory by              chlorine residual was combined. High
membrane filtration. Plate count data               plate counts also occurred more fre-
were not judged sufficiently reliable to            quently in the latter. Pools having 0.3
be considered as a criterion for compari-           mg/I of total chlorine but less than 0.3
son with other studies. However, com-               mg/l free available chlorine, however,
parison of plate count data among cate-             showed better bacteriology than those
gories within this study would be valid.            pools with 0.3 mg/l free available
Of the 55 pools containing at least 0.3             chlorine but less than 0.3 mg/l free
mg/l of free available chlorine, 45 or              chlorine at pool pH.
82 per cent contained no coliforms                     The last three groups of pools-hav-
whereas, of the 44 pools containing 0.3             ing low chlorine residuals, no residuals,
mg/I or less of total chlorine, only 21             or stabilized chlorine residuals with no
or 48 per cent were coliform-free. Of               free available chlorine-showed, as ex-
the 28 pools containing chlorine re-                pected, a high frequency of occurrence
siduals stabilized with cyanuric acid, 18           of coliforms and Pseudomonas aerugi-
or 64 per cent were coliform-free. The              nosa. Except for the low numbers of
explanation of the difference between               staphylococci found in stabilized chlo-
1967 and 1968 bacteriological data for              rine-disinfected pools, these organisms
the latter group is believed to be that             were much more numerous in the pools
only one pool in the 1967 study had                 with low residuals than where free
greater than 0.3 mg/I free available                available chlorine was present. High
chlorinie wlhile 8 of the 15 studied in             plate counits were also encountered more
1968 had at least this much free avail-             frequently in these three groups than
able chlorine, and all but one of these             in the pools with 0.3 mg/l free avail-
containe(d 1.0 mg/l or more.                        able chlorine.
   A calculatioin of free chlorine at. pool
p/I. to be definied later, was made for             Physical Parameters
the 55 pools containingr 0.3 mg/I or                   An attempt was made to relate the
-reater free a-ailable chloriine. Forty-            presence or absence of algae to the type
five of these lhad 0.3 mg/l or greater              of pool surface and to the type of fil-
free chlorinie at pool pIh. Of these 45,            ter being used. No significant differ-
only 6 had either colifornms or unsatis-            ence was found in the incidence of
factory membrane filters, wlhereas 4 of             algal growths as between pools with
the 10 pools Inot having 0.3 mg/l of                paiinted surfaces and those coated witl
free chlorine at pool pHi slhowed either            marbleite plaster. Although the imlci-
coliforms or unsatisfactorv membrane                deuice of algal growths in pools equipped

APRIL, 1970                                                                               743
with vacuum diatomite filters appeared          1. Free Chlorine in Aqueous Solu-
to be somewhat less than in pools using      tion-The reaction of chlorine gas with
either pressure diatomite or sand filters,   water to form free chlorine can be repre-
the relationship was not significant at      sented as follows:
the 95 per cent probability level.                   Cl2 +H2OHOCl +H + C1-
   An attempt was made to relate clarity     That form of chlorine represented as
of pool water to the type of filter being    HOCl (hypochlorous acid) is free chlo-
used. While the percentage of pools with     rine as found in swimming pools, and
brilliant, turbidity-free water was a        is more powerful both chemically and
little greater in the case of pools with     as a bactericide and algicide than any
vacuum diatomite filters than those          other form of the element.
having either sand or pressure diatomite        2. Free Available Chlorine at Pool
filters, the data do not definitely indi-    pII-The dissociation of HOCI at values
cate that one type of filter outperforms     above pH 5.0 in aqueous solution estab-
another as far as water clarity is con-      lishes an equilibrium with hydrogen and
cerned. It is believed that individual at-   hypochlorite ions so that two species of
tention by careful and conscientious         chlorine, HOCI and OC1-, are present.
operators is the most important single       The relative amount of each species is
factor (and the one most frequently          dependent upon pH, the absence of am-
lacking) in maintaining a clear, clean       monia, and of other environmental de-
pool.                                        bris. The following equation represents
Pool pH                                      the dissociation:
                                                          HOCIHI+   0+C1-
   The Florida Code requires that pool           Free chlorine       Available chlorine
pH be maintained within the range pH         The meaning of the word "available"
7.2-8.0. The APHA2 recommends that           must be made clear. The hypochlorite
the upper limit be extended to pH 8.2.       ion, OC1-, has little or no bactericidal
The results of our study seem to indi-       activity but it serves as a reservoir or
cate clearly that most pool operators are    bank from which more free chlorine,
either not aware of these limiting and       HOCI, may be formed as that initially
permissible values or that they make         present is removed by pool water
little or no attempt to follow them. Of      demand.
the 193 pools studied, 51 per cent con-         3. Combined or "Stabilized" Chlorine
tained water with pH values either be-
low 7.2 or above 8.2. Specifically, 83       -This is the form which chlorine as-
had pH values below pH 7.2 and 16            sumes in the presence of such materials
had pH values above pH 8.2.                  as ammonia or its salts, cyanuric acid,
                                             sulfamic acid or urea. The relative
Orfhotolidine Measurements                   amount of combined or "stabilized"
Misleading                                   chlorine and the forms it can take are
   The standard OT test as performed         dependent on pH, temperature, and the
by pool operators gives results which        concentrations of both HOCI and the
are both erroneous and misleading.           stabilizing material. The reaction with
Orthotolidine produces a yellow color        NH3 is as follows:
when oxidized at pH values lower than                 NH3 +HOCL:±NH2Cl +H20
1.8 by chlorine, chloramine, nitrites,       The product is called monochloramine,
and the higher oxidation states of iron      which is probably the most common
and manganese. Before we discuss the         form of combined or stabilized chlorine
reactions involved, perhaps some defini-     found in pools. However, its kill time is
tion of terms would be helpful.              only about one-thirtieth that of hypo-

744                                                              VOL. 60, NO. 4, A.J.P.H.
                                                 DISINFECTION OF SWIMMING POOLS

chlorous acid3 which is much too slow         free available chlorine is determined by
for swimming pool disinfection.               titration with phenylarseneoxide at pH
   According to Zettler and Solsor,4 free     7.0. Platinum electrodes measure the
chlorine is produced in water by the          drop in current as the reaction proceeds.
general reaction:                             The end point is reached when no fur-
          ACI+H20.AH+HOC1                     ther decrease in current occurs upon
                                              addition of a small amount of phenyl-
where A represents the negative part of       arseneoxide. After the free available
the molecule to which the positive chlo-      chlorine is titrated, the pH is lowered
rine is originally attached (in the above     to 4.0 and excess iodide added. Com-
case NH2). This may be Cl--Cl+ as in          bined chlorine present will oxidize
chlorine gas, NAO-Cl+, chloramide             iodide to iodine and the free halogen
 (cyanuric acid) or chlorimide. The           is titrated with phenylarseneoxide,
amount of free chlorine released when         thereby measuring the amount of com-
ACI is added to water depends on the          bined or "stabilized" chlorine present.
rate and degree of hydrolysis of ACI             Table 2 and Figure 1 show the per-
and the amount of AH present. In the          centage of HOCI (free chlorine) and
case of chloramine and other N-chloro-        OC1- (hypochlorite ion or available
inorganic compounds the rate of hydrol-       chlorine) present at pH values from 4-9,
ysis is so low that very little HOCI is       for a total chlorine residual of 1.0 ppm.
formed, hence the weak disinfecting              Table 2 gives the data for the usual
powers exhibited. The chloramides, in         pool pH values. It shows that, at a pool
rate and degree of hydrolysis, fall be-       pH of 7, about three-fourths of the re-
tween the complete hydrolysis of chlo-        sidual present is "free chlorine" or
rine gas and hypochlorite, and the inac-      HOCI. However, at pH 8, less than one-
tivity of chloramine.                         fourth is "free chlorine" and a little
   The pool operator is primarily in-         more than three-fourths is relatively in-
terested in the level of free available       active "available chlorine." No method
chlorine in the pool. The method of dif-      is available for measuring free chlorine
ferentiating free available chlorine and      at pool pH which is actually what we
combined chlorine by the standard             wish to know and which determines the
orthotolidine test is to cool the test sam-   rate of bacterial kill.
ple to 10 C, add orthotolidine solution,         As stated earlier, free available chlo-
and read the flash color within five sec-     rine is determined amperometrically at
onds as free available chlorine. A sec-       pH 7.0. The following equation repre-
ond sample is treated at pool tempera-        sents the species of chlorine at this pH:
ture and allowed to react five minutes                     HOCI±H++OC1l
after orthotolidine is added, then read.              72.5%             27.5%
The difference between the two deter-
minations is combined chlorine. During
this survey, however, we found no pool        Table 2-Effect of pH on forms of
operators and only one sanitarian who           chlorine5
measured chlorine residuals in the above        pH       % C12    % HOCI      %oC1-
manner. Most merely read the flash and
five minute residuals at pool water              4        0.5      99.5            0
temperature, which leads to inaccurate           5          0      99.5          0.5
results.                                         6          0      96.5          3.5
   A second, more accurate method of             7          0      72.5         27.5
                                                 8          0      21.5         78.5
measuring free available chlorine is by          9          0         1.0       99.0
means of an amperometric titrator. The

APRIL, 1970                                                                            745
Figure 1-Effect of pH on per cent HOCI                        from the pool pH and the residual de-
  for any free available chlorine residual                    termined by titration; however, most
                                                              operators are not fortunate enough to
                                                              have a titrator and must rely on the
                                                              orthotolidine test kit. The same free
                                                              available chlorine is, supposedly, meas-
   7                                                          ured by orthotolidine, but one signifi-
                                                              cant difference exists, namely, the pH
                                                              of the sample is lowered to below 1.8
                                                              when orthotolidine is added. At this
                                                              pH, and at pool temperature, most of
                                                              the combined or "stabilized" chlorine
       0   10   20   30   40   50   60   70   80   90   100
                                                              will be measured as free available chlo-
                          % HOCI
                                                              rine, indicating a higher than actual
                                                              level of active disinfectant. The rate of
   As one titrates the HOCI, the equi-                        the reaction with orthotolidine is, there-
librium shifts to the left and continues                      fore, a function of temperature, pH, and
to do so until all the HOCI and OCl-                          concentration.
are titrated. The final reading is the                           In order to evaluate the discrepan-
sum of both the powerful disinfectant,                        cies between residuals determined with
HOCI and OC1-, which is bactericidally                        orthotolidine and those determined
ineffective. This measurment is free                          amperometrically, both tests were car-
available chlorine.                                           ried out at each pool. Of the 110 pools
   Let us assume that a pool whose pH                         disinfected with chlorine or hypochlo-
is 7.5 shows a free available chlorine                        rite and which had measurable residual
residual of 1.0 ppm. Reference to Table                       by both methods, residuals measured by
3 shows that actually only 0.47 ppm of                        orthotolidine were almost uniformly
this is free chlorine and the other 0.53                      significantly higher than those measured
ppm is available chlorine, and the pool                       by amperometric titration. This can be
with a measured 1.0 ppm residual                              seen when the deviations are plotted in
barely meets the 0.4 ppm APHA2 cri-                           Figure 2 as orthotolidine residual minus
terion.                                                       amperometric residual versus the com-
   Therefore, to convert any measured
free available chlorine residual to the                       Table 3-Amount of free available chlo-
equivalent amount of free chlorine at                           rine required to provide 0.40 ppm free
pool pH, one merely multiplies the                              chlorine at pool pH
measured residual by the percentage
shown in column 2 (Table 3) for the                                 Percentage                  Free avail-
pool pH and divides the result by 100.                                of free                  able chlorine
                                                                     available                   required
For example, to convert 0.45 ppm of                           pH of chlorine                    to provide
free available chlorine, as determined                         pool   present       Conversion 0.4 ppm
either by the amperometric titrator or                        water   as   HOCI       factor  free chlorine
the OT test performed at about 10 C,                            7.2        62.3        1.60        0.64
pool pH 7.6:                                                    7.3        57.2        1.77        0.70
         4                                                      7.4        52.1        1.94        0.77
 0.45 x      =0.19 ppm truie free chlorine at                   7.5        47.0        2.14        0.85
  pH 7.6.                                                       7.6        41.9        2.37        0.95
  The pool operator could easily de-                            7.7        36.8        2.70        1.09
termine his free residual 1v calculation                        7.8        31.7        3.14        1.26

746                                                                               VOL. 60, NO. 4, A.J.P.H.
                                                                      DISINFECTION OF SWIMMING POOLS

Figure     2-Relationships of (OT7-                                three cases, no orthotolidine residual
     AMP±r) in mg/p to combined chlorine                           was found when total chlorine was
                                                                   greater than 0.25 ppm (one as high as
      2.0-                                                         0.45 ppm), when measured by ampero-
                                                                   metric titration. Figure 3 shows the dis-
                                                                   tribution of the 103 total chlorine re-
       1.5j1                                                       siduals measured by both methods. If the
                                                                   two methods of measurement were
                                                                   equivalent, the points plotted would have
                                                                   fallen along a line with a slope of 1.0.
E                                                                  It is apparent that they do not but were
                                                                   found to lie about a line with slope of
tL                                                                 0.83±0.07 with an intercept at 0.06
                   ,*   -
                                                                   ppm. This shows that the orthotolidine
z     rn v .
       e       -        ,.                                         residual becomes slightly less sensitive
                                   1.0   COMBINED CHLORINE
                                                                   at high residuals than at low residuals,
                                                                   and that it fails to pick up residuals
                                                                   less than 0.06 ppm, which is in the
                        RECOMENDATION 0.1 ppm)                     range of the consistent failure at resid-
                                                                   uals less than 0.15 ppm. A frequency
     -I. c-
                                                                   distribution of the deviation of the
                                                                   orthotolidine test and amperometric ti-
                                                                   tration method for total chlorine is given
bined chlorine residual obtained by                                in Table 4. Only 31 per cent were within
amperometric titration. The discrepancy                            the +0.1 ppm agreement recommended
became strongly positive in cases where                            by the APHA.
greater than 0.40 ppm of combined                                     In 27 of the 28 pools employing
chlorine was present. It was felt, after                           cyanuric acid stabilized chlorine, ortho-
the 1967 survey, that chilling the sam-                            tolidine residuals were compared with
ple to 340 F might reduce the strong                               residuals measured by amperometric
interference in the orthotolidine test by                          titration. Eleven of these 22 pools
concentrations of combined chlorine.                               showed no measurable free available
However, tests conducted during the
1968 study show that chilling does not
bring any improvement. Furthermore,                                Figure 3-Comparison of chlorine resid-
                                                                     uals determined with orthotolidine and
discrepancy did not appear to be in-                                 by amperometric titration
fluenced by either pH or alkalinity.
   The parallel lines of Figure 2 enclose
the area within which are found all
measured residuals which were within
 +0.1 ppm of amperometric values. In
this study 52.8 per cent of the measured
residuals fell outside this satisfactory
   The orthotolidine test also has short-
comings as a measure of total chlorine
residual. The test kit usually failed to
show any residual in cases where total
chlorine by amperometric titration was                                                  1.0
less than approximately 0.15 ppm. In                                               AMP. TITR. (mg/t)

APRIL. 1970                                                                                              747
chlorine residual by amperometric ti-          ing agent than combined chlorine. The
tration. Two had free available chlorine       over-all total of 35 pools showing the
residuals less than 0.3 ppm and nine           presence of coliforms represents 18 per
showed values greater than 0.3 ppm. It         cent of the total pools and is slightly
is noteworthy that each of the 22 pools        higher than the 15 per cent maximum
showed a high free available chlorine          isolation rate recommended by APHA.
residual with orthotolidine "flash" test.      However, the 56 pools yielding unsatis-
    Behavior of the titrator was sluggish      factory membrane filters is much higher
in measuring free available chlorine at        than the maximum. Most of these iso-
9 of the 28 pools, with significant free       lates were found in pools with little or
available chlorine residuals. In titrating     no free chlorine at pool pH.
free available chlorine in the absence             In this study, fecal streptococci were
 of cyanuric acid, the microammeter            found much less frequently than coli-
needle drop is sharp with incremental           forms, even though Streptococcus
 addition of phenylarseneoxide, and the         fecalis is more resistant to chlorine than
 end point in sharp. When cyanuric acid         E. coli.3'9 Additionally, fecal strepto-
 was present, both in the field study and       cocci were usually associated with the
 in the laboratory, needle response is          presence of other indicator bacteria.
 slow and irregular and the end point              Pseudomonas aeruginosa, a potential
 difficult to define due to needle drift.       pathogen and a causative agent of ear
 This opens the question of whether the         infection among swimmers,10 was found
 cyanuric acid stabilized chlorine so           in alarming frequency in pools where
 measured is really free as HOCI and            less than 0.3 mg/l free chlorine at pool
 OC1-, or whether it is merely bound            pH was present. This organism was not
 loosely enough to react with the reduc-        found more frequently than coliforms
 ing agent added. The cyanuric acid             in any disinfectant group studied.
 level was tested in the 15 pools encoun-          Staphylococci were found in a very,
 tered in the 1968 study. In pools show-        large proportion of the pools, but were
 ing free available chlorine, the average       found in large numbers only in pools
 cyanuric acid level was 58 mg/l as op-         with low chlorine residuals. These data
 posed to 80 mg/l in those without meas-        do not agree well with previous work
  urable free available chlorine residuals.      reported in the literature,l a fact
 This agrees with findings of other in-         probably due to die-off in shipment. It
  vestigators 67'8 who found that hydrol-       has been shown that staphylococci are
  ysis was greatly suppressed by levels of      present in high numbers in heavily used
  cyanuric acid greater than approxi-           pools in the absence of coliform organ-
  mately 50 mg/l.
     These results conclusively indicate the
  obsolescence of the orthotolidine test for   Table     4-Frequency         distribution     of
  measurement of chlorine in swimming             (Ampr-OTr) in 103 pools
  pools for two reasons. First, it does not            Deviation           Nuimber of pools
  differentiate between free and combined
  chlorine, but reports a substantial per-                  0                      7
  centage of combined chlorine as "free                ± 0.01-0.05                 9
  available." Second, it does not meet that            ± 0.06-0.10                16
  criterion of APHA which recommends                   + 0.11-0.15                12
  an accuracy of +0.1 ppm.                             ± 0.16-0.20                17
                                                       ± 0.21-0.25                10
 Discussion                                            ± 0.26-0.30                12
    The results show again that free avail-              >0.30                    20
 able chlorine is a much better disinfect-

748                                                                  VOL. 60. NO. 4, A.J.P.H.
                                                 DISINFECTION OF SWIMMING POOLS

isms, and that staphylococci are more         tion, and orthotolidine, coliform den-
resistant to disinfecting agents than E.      sity by both multiple tube and mem-
coh.                                          brane filtration, fecal streptococci, total
   Staphylococci are derived from the         staphylococci, and Pseudomonas aerugi-
skin and mucous membranes of human            nosa.
beings; their presence in pool water in-         2. The study indicates that the opera-
dicates the possibility of transmittance      tion of public swimming pools is not as
of staphylococcal infections, and also        good as it should be. One hundred and
that other pathogenic organisms from          five of the 193 pools failed to have the
the skin or mucous membranes may be           residual required by the state health
present. The high numbers found in the        code, and 106 failed to have pH values
absence of adequate or marginal chlori-       within the acceptable limits. Eighteen
nation provide further that large popu-       per cent of the pools studied contained
lations of these organisms may build up.      coliforms or showed unsatisfactory
   A standard of 100 staphylococci per        membrane filtration results. In many
100 ml in swimming pools has been pro-        instances, pool operators knew little
posed by Favero, et al1 This allows a         more than the simple details of me-
tremendous number of staphylococci to         chanical operation. All too many of
contact the skin of a bather immersed         them had little or no conception of the
in a pool. This number was only rarely        significance of pH or chlorine residual,
exceeded both in this study and in a          or how to maintain a residual within
previous study in which more heavily          acceptable limits. Many monthly reports,
used pools were sampled. Therefore, it        to be submitted to the health depart-
is felt that this criterion offers no great   ment, were found filled out days in
degree of safety. A more conservative         advance.
figure of perhaps 30 staphylococci/100           3. The standard orthotolidine test for
ml of pool water would be more ra-            chlorine residual, as performed by most
tional. Robinton, et al.,11 showed that       pool operators, yields misleading and
in pools with greater than 1.0 ppm free       erroneous results. At pool water tem-
available chlorine no more than four          perature, the flash reading of free avail-
staphylococci per 100 ml were present,        able chlorine residual also includes a
even where loads were reasonably              portion of the combined chlorine re-
heavy. Keirn and Putnam3 found ap-            sidual, the amount depending upon the
proximately 20 staphylococci/100 cc in        ratio of free and combined chlorine.
pools containing free available chlorine,     This interference is especially bad at
but in these pools residuals averaged         low free available chlorine residuals. At
much lower than in those investigated         present, the only foolproof way of ob-
by Robinton.                                  taining the free available chlorine re-
                                              sidual is by amperometric titration, a
Summary                                       method requiring expensive laboratory
   1. A field study was made of the              4. Coliforms were present in 18 per
chemical and bacteriological quality of       cent of the pools studied, mostly in pools
the water of 193 below-ground public          with inadequate chlorine residuals.
pools in Florida. These pools ranged in       Fecal streptococci were found much less
size from 11,000 to 624,000 gallons.          frequently than coliforms. Staphylococci
Parameters investigated included tur-         were isloated in a very large proportion
bidity, type of filter, presence or ab-       of the pools and were present in high
sence of algae, pH, determination of          numbers in pools with low chlorine re-
chlorine residual by amperometric titra-       siduals. Pseudomonas aeruginosa was

APRIL, 1970                                                                          749
isolated with about the same frequency                 Covering Public Swvimming Pools. Pre-
as coliforms.                                          pared by the Joint Committee on Swim-
                                                       ming Pools of the APHA in cooperation
   5. The bacteriological data clearly                 wvith USPHS, 1964.
show the importance of pH to the bac-             3.   Keirn, M. A., and Putnam, H. D. The
tericidal effectiveness of chlorine. Pools             Resistance of Staphylococci to Halogens
in which most of the free available                    as Related to a Swimming Pool Environ-
chlorine was calculated to be 0C1-                     ment. Health Lab. Sc. 5:180, 1968.
                                                  4.   Zettler, T. T., and Solsor, J. Q. Swimming
showed much worse bacteriology than                    Pool Sanitizers. Soap & Chemical Special-
those pools where the predominant spe-                 ties (Mar.) 1966, p. 51.
cies was HOCI.                                    5.   Black, A. P.; Kinman, R. N.: Thomas,
   6. The authors strongly recommend                   W. C., Jr.; Freund, G.; and Bird, E. D.
that free chlorine residuals at pool pH                Use of Iodine for Disinfection. J. Am.
                                                       Waterworks A. 57:1401, 1965.
be included in any swimming pool code             6.   Andersen, J. R. A Study of the Influence
and, further, that more effective opera-               of Cyanuric Acid on the Effectiveness of
tion of public pools could be obtained                 Chlorine. Presented at NSPI Convention,
by requiring licensing of operators and                Chicago, 1964.
complete automatic control of chlorina-           7.   Fitzgerald, G. P., and Der Vartanian, M.
                                                       E. Factors Influencing the Effectiveness of
tion and pH adjustment.                                 Swimming Pool Bactericides. Applied
  ACKNOWLEDGMENTS - The authors wish to                Microbiol. 15:504, 1967.
acknowledge the valuable assistance of Miss V.    8.    Robinton, E. D., and Mood, E. W. An
A. Morgan and Mrs. Albertha Pringley in                 Evaluation of the Inhibitory Influence of
making the bacteriological analyses in the              Cyanuric Acid Upon Swimming Pool Dis-
Earle B. Phelps Laboratory in the Bioenviron-          infection. A.J.P.H. 57:301, 1967.
mental Engineering Department, University of      9.    Stuart, R. S.. and Ortenzio, R. F. Swim-
Florida. Appreciation is also expressed for             ming Pool Chlorine Stabilizers. Presented
bacteriological analyses made by laboratories           at the 50th Midyear Meeting, Chemical
of the Florida State Board of Health at Jack-           Specialties Manufacturing Association,
sonville and Miami, and the County Health               Chicago (May 19), 1964.
Department Laboratory at Sarasota.               10.    Cothran, Walter W., and Hatlan. Jack B.
                                                        A Study of an Outdoor Swimming Pool
REFERENCES                                              Using Iodine for Water Disinfection. Stu-
                                                        dent Med. 10,4:493, 1962.
 1. Favero, M. S.; Drake, C. H.; and Randall,    11.    Robinton, E. D.; Mood, E. W.: and
    G. B. Use of Staphylococci as Indicators            Elliot, L. R. A Study of the Bacterial
    of Swimming Pool Pollution. Pub. Health             Flora in Swimming Pool Water Treated
    Rep. 79:61, 1964.                                   with High Free Residual Chlorine.
 2. Suggested Ordinance and Regulations                 A.J.P.H. 47:1101, 1957.

           Dr. Black is Research Professor Emeritus of Chemistry and Bioenvironmental
        Engineering, University of Florida, and Project Director, Iodine Demonstration
        Grant, City of Gainesville (P.O. Box 1329), Gainesville, Fla. 32601. Mr. Keim
        and Mr. Smith are Research Associates, Iodine Demonstration Grant, Gaines-
        ville. Mr. Dykes is Sanitary Engineer, Division of Water Supply, Florida State
        Board of Health, Jacksonville. Mr. Harlan is Assistant Professor of Physical
        Education and Swimming Coach, University of Florida, Gainesville.
           Work on this project was supported by Demonstration Grant WPD-19-05-67,
        WPD-19-06-68 awarded by the FWPCA of the Department of the Interior.
           The paper was presented before a Joint Session of the Conference of Local
        Environmental Health Administrators, the Conference of State Sanitary Engi-
        neers, the National Association of Sanitarians, and the Engineering and
        Sanitation and Laboratory Sections of the American Public Health Association
        at the Ninety-Sixth Annual Meeting in Detroit, Mich., November 14, 1968.

750                                                                   VOL. 60, NO. 4, A.J.P.H.

To top