DISINFECTION OF SEA WATER BY
C. B. Kelly
Second Reproduction in the
SALT WATER FISHERIES DIVISION
FLORIDA BOARD OF CONSERVATION
Data fr,)m experi.m.nt,; in artificria: .y nrd natturc.aly
pulluated sea water indicate that ultraviolet radiation
effectively destroys coliform organ:isms. The unit
developed for this purpose is described and the results
obtained are analyzed.
DISINFECTION OF SEA WATER BY
C. B. Kelly
The use of ultraviolet radiation for the disinfection of water has
been investigated primarily for its applicablity to fresh waters, although
in recent years investigations with sea water have been conducted in Japan
and in England, particularly in connection with water to be used for the
purification of shellfish Studies on fresh water have been directed
primarily toward the sterilization of drinking water and swimming pools,
but the method has also been evaluated for the disinfection of small
containers used for dispensaries and surgeries and for biochemical products
when heat treatment would have an adverse effect.
Our interest in the ultraviolet treatment of sea water arose from
the need for a continuously flowing supply, having a uniformly low bacte-
rial content, for study of the accumulation and elimination of enteric
bacteria in oysters.
Luckiesh and Holladay present fundamental data-on the radiation
requirements for disinfection of fresh waters of specified physical and
chemical characteristics, The factors which influence the quantity of
radiation required were found to be those which affect the penetration of
ultraviolet energy through the water. Turbidity, color, and dissolved
iron salts were found to be among the nsturelly occurring qualities which
had the grte;e*A t f-ec on the absorption coefficient of the nater.
Gillcreas and De Lalla determined the efficiency ofacommercial device in
treating Albany tap water from a surface supply and studied the inter-
ference of added color, turbidity, and colloidal iron compounds. They
found that with turbidity of 85 ppm, iron 3 ppm, color (inorganic) 90 ppm,
or with turbidity of 15 ppm, iron, 5 ppm, and color 150 ppm, effective
destruction of coliform organisms was accomplished at the rated capacity
of the equipment.
As to the relative efficiency of ultraviolet light in destruction
of micro-organisms other than those of the coliform group, Chaumeau lists
the following comparative exposure times required for the destruction of
some common microo.organisms: ctaphylococci, 5-10( seconds chol'ra, 1.0-15;
coli, 15-201 typhoid, 10G.-a2 dysentery, 10O30; pnanmoba.cilli 15l-30
subtilis, 30--401 tetanus, 20-b60 Buttolph, iHay~es, and Matelzky give the
following radiation requirements, in terms of microwatts per squaree centi-
meter to kill 63.2 per cent in one minute (1 leth): Alpha streptococci
and Staphylococcus aureus, 1.3; Beta hemolytic streptococci and Eacherihhia
coli, 5,3; molds, 10; and mold spores typical, 28, Kawabata and Harada
reported on the relative resistance of four groups of typical micro-
organisms in aqueous suspension in a unit utilizing a 15-watt germicidal
lamp delivering an ultraviolet intensity of 90 microwatts per square
Gram-negative bacteria were found to be the least resistant. The time
required to kill 99.9 per cent for gram-negative strains was found to be ac
follows: Proteus sp., 42 sec; Shigella sp., 47; Eberthella typhosa, 49j E,
coli, 60. Gram-positive organisms had the following requirements: Group A
Streptococcus, 83 seconds; Staph. aureus, 103; S. fecalis, 165; Bac.
subtilis sawamura 240; Bac. subtilis spores 369. The resistance of yeasts
was slightly greater, Saccharomyes sake requiring approximately 217 seconds
and Willia anomala 420 seconds.
Pilot plant studies were conducted on sea water by Public
Health and Fisheries Authorities of Hiroshima in an apparatus designed
to investigate methods for the purification of the Pacific oyster,
Crassostrea gigas, using three germicidal units similar in design to that
of Kawabata and Harada, with rather high rates of flow however. Water
from a tank containing oysters was recirculated to provide a replacement
of about one-half of the total volume of the tank per hour. A reduction
of approximately 90 per cent in coliform density was accomplished in one
passage through the ultraviolet unit. Pilot scale studies on the
efficiency of ultraviolet radiation in the sterilization of sea water were
also conducted by Kamimura and Suzuki at Matoya Bay, Japan, utilizing a
unit provided with a total of six 15-watt germicidal lamps. Flow through
the unit was at the rate of 12 gallons per minute. Approximately 90 per
cent of the bacteria were killed by a 15-second exposure.
In all of the semicommercial size units designed for the purifi-
cation of shellfish, including one recently developed in England, there
appears a similar principle. The ultraviolet units are of such capacity
that by recirculation of the water there will be accomplished a destruction
of bacteria concurrent with the use of the water for purification of shell-
fish. The water is not completely sterilized in the initial passage
through the treatment system. The rate of destruction, although slow,
accomplishes purification of the sea water to a low acceptable coliform
level by continuous recirculation during the 24-hour period of treatment,
and the water presented to the oyster some four to six hours prior to the
completion of the treatment period is practically zero in coliform content.
The principle is not acceptable in American practice. The
Public Health Service Manual of R commended Practice for the Sanitary
Control of the Shellfish Industry stipulates that water used for puri-
fication of shellfish, ifl created, should meet the Public Health Service
Drinking Water Standards. Thus, the coliform density of the water must
be such that not more than 10 per cent of the 10 milliliter portions
tested shall show the presence of organisms of the coliform group.
Occasionally three or more of the 10 milliliter portions may be positive
for coliform organisms, provided that this density shall not occur in
more than 5 per cent of the samples when 20 or more of the samples are
examined per month. To meet this bacteriologic requirement, the ultra-
violet treatment unit would need to be of much greater killing power than
those used in commercial or semicommercial establishments in foreign
- 2 -
During the course of laboratory experiments on shellfish con-
ducted at Pensacola, Fla., and Purdy, Wash,, we have had occai.cn to
design and to evaluate ultraviolet treatment units of a high degree of
efficiency in the sterilization of sea water. The units were designed
primarily for use in laboratory experiments to determine the rate of
bacterial accumulation by shellfish. In these experiments, oysters in
laboratory tanks were furnished flowing sea water artificially polluted by
the continuous addition of suspensions of coliform organisms. Samples of
water and oysters were examined periodically for coliform density and from
these tests the rate and extent of accumulation of bacteria by the oysters
were determined. Obviously, a requirement for the proper operation of
these studies was a constant load of coliform organisms in the water.
Since the sea water proved to be extremely variable in coliform content it
became necessary to treat the water to reduce the native coliform content
to a constant level. Reduction to zero density was considered to be the
easiest and most practical.
Preliminary experiments indicated that chlorination resulted in
the production of a water not acceptable to the oysters. Studies at
Pensacola demonstrated that the feeding activity of oysters was adversely
affected by chlorination and dechlorination of sea water. Therefore, other
methods of removal or destruction of coliform organisms were considered,
In an effort to avoid the harmful effects of chemicals on the activity of
the oysters, it was decided to evaluate ultraviolet radiation. After hav-
ing first demonstrated that ultraviolet treatment of sea water did not pro-
duce the same objectionable effect as chlorination and dechlorination, the
bactericidal efficiency of the treatment units were determined.
MATERIALS AND METHODS
Description of Ultraviolet Treatment Units
The ultraviolet treatment units used in these studies were of
two different designs. The unit used in the studies at Pensacola is a
modification of an apparatus suggested by Buttolph It consists of a
bottom section, essentially an open trough through which the water passes,
and a top section, a reflector unit housing the ultraviolet lamps. The
trough is 36 inches long, 8 inches wide and 6 inches high, with a longi-
tudinal baffle 2 1/2 inches high placed along the center line of the trough.
Intake and discharge pipes, 2 inches in diameter, are placed on the long
sides of the tank at diagonally opposite corners.
The lamp housing, fitting on the trough, is a semicylindrical
archshaped reflector, lined with specular finidsaluminum material. The
unit is provided with two 30-watt ultraviolet germicidal lamps, placed so
that the lamps are approximately one-half the distance from the lower edges
to the center line of the reflector. When the housing is in place on the
trough, the distance from the center of the lamps to the water level is
6 1/8 inches.
Preliminary experiments with the ultraviolet treatment unit indi-
cated that at least 150 liters (39.6 gallons) per minute could be treated
with a high degree of destruction of coliform organisms. At this rate of
flow the theoretical detention time in the treatment trough is 20 seconds,
and the actual detention time as determined by dye tests is 15 seconds.
The amount of ultraviolet energy then reaching the water at this rate of
flow is 960 microwatts per square centimeter per minute. We realize that
this is considerably in excess of the design data furnished by some of the
manufacturers as well as the intensities prevailing in other installations.
However, as previously mentioned we were interested in accomplishing as
near 100 per cent kill as possible, with an extremely high degree.of
Influence of Chlorination of Sea Water on Feeding Activity of Oysters.
Experiments at Pensacola demonstrated that chlorinated and
dechlorinated sea water had an adverse effect on feeding activity of the
eastern oyster, Crassostrea virginica. Similar studies conducted at Purdy
with the Olympia oyster, Ostrea lurida, demonstrated the same effect. Com-
parative activity of Olympia oysters furnished untreated and chlorinated,
dechlorinated sea water is presented in Table 1. Chlorine dose was
sufficient to produce a residual of 0.2 to 0.5 ppm after five minutes
contact and the water was dechlorinated with an excess of sodium thiosul-
fate. Activity was determined by the presence of fecal deposits. The
adverse effect of chlorination can be seen by the distinct difference
between per cent activity observed in the "treated" and "untreated" flats.
Similar experiments, using ultraviolet treated water, demon?
strated that this method of treatment did not produce the same effect.
The results of a series of four experiments are presented in Table 2.
It can be seen that there was little if any difference in the per cent
activity between oysters receiving treated and untreated water,
Germicidal Efficiency of UV Treatment Units,
The efficiency of the units at Pensacola was first determined
in preliminary experiments with artificially polluted sea water. Sea
water was seeded with sufficient cotton-filtered raw sewage to produce
coliform loads up to MPN 7,900 per 100 ml. The sewage was introduced
continuously into the sea water by means of a proportionating pump. The
polluted sea water was then allowed to flow through the ultraviolet treat-
ment units at the rate of 40 liters per minute. Twenty-five sets of
samples were examined during the course of experiments conducted on three
days. The units consistently reduced the coliform density of the water
to less than ..8 organisms per 100 ml (negative in five 10 ml volumes
tested). The raw water coliform MPN's ranged from 79 to 7,900, with a
median of 2,200 per 100 ml.
Table 1--Influence of Chlorinated and Dechlorinated Sea Water on Feeding
Activity of Olympia Oysters
Treatment of Water Test Period Water Per cent activity
for Test Tank Hours Temp. 0 C Treated Untreated
water flowing through 12 9 21 57
11 8 29 50
31 8.5 30 57
23 8 40 71
(both tanks supplied
untreated water) 21 8 59 60
23 8.5 66 65
Table 2--Influence of UV Treated Sea
Water on Feeding Activity of Olympia
Test Period Water Per cent activity.
Hours Temp. OC Treated Untreated
23 10 91 88
22 10 90 87
22 9 90 88
34 9 94 96
The apparatus used at Purdy was also evaluated .prior to operating
the unit in the routine experiments. The results of these tests are pre-
sented in Table 3. Five series of tests were conducted. As in the experi-
ments at Pensacola, the sea water was seeded with cotton filtered raw
sewage to produce the coliform densities shown. The Purdy experiments,
however, were augmented to allow the collection of more information
concerning the efficiency of this method of sterilization. In the coliform
determinations, combinations of 100, 10, and 1 ml portions, five replicated
per dilution, were planted in lauryl sulfate broth. In the Pensacola
experiments the largest portion planted was 10 ml. In addition to the
coliform tests, plate counts at 35cC and at 20C were determined on the raw
water and the treated water.
The 350C plate counts
were conducted on tryptone glucose -
extract agar, incubated for 24 hours
The 2OC plat counts were conduct!
on McLeods basal medium incu-
bated for five days. Water flow O O 0 ao aCo : o
0 0 0 0 0 C +
through the unit was maintained at 3 (D
60 liters per minute in two of the
experiments and at 100 liters per H
minute in three. The experiments % oO co eO 0
conducted with eight lamps at a m
flow of 100 liters per minute show- M
ed a lower per cent removal of H
plate- count, although the destruc- c "
ion of coliform organisms appeared p D
to be not affected by the higher t 'U W
rate of flow. In experiment 5, 0 0 0 0 o o
the UV radiation was increased by o
the addition of five lamps and the
results Indicate that this added L c
radiation produced exceedingly s I o o
high reductions in 35e and 200 O 0 j 0 o
plate counts as well as in coliform H
organisms. ,to : 0 L '
o 0 0 0 0 o
O O O O O 0 +
In an attempt to deter-
mine the bacterial types that were 0
resistant to the ultraviolet radia- ca
tion, identification was made of
the types of organisms present in A A A A6 A 01
0 0 0 0 4 c (D
colonies on the tryptone glucose P P 0 *3 O
agar and the MacLeod's agar plates co 0 c c co P E
inoculated with ultraviolet treated Z
water. Thirty such cultures were
examined. Nineteen were found to o v I
be Pseudoranas, two were Achromo- S\ cT
bacteria, one Flavobacterium, one
Micrococcus, and one Cytophaga. -0
Although cultures from the untreat- o Z
ed water were found to contain
Bacilli and Vibrio in addition to
the above, the latter two bacterial Y 0
types were not isolated from the
ultraviolet treated water. V V V V V (o
o \o 0o o \o o 0
Results of Pilot Plant Studies. *
\0 '0 D0 '0 I z 03
p- o o -o Vo j o r I
Information on the effi- z
ciency of UV treatment under pilot
plant conditions was obtained during 1 co
the course of experiments on the 0 C
accumulation of bacteria by oysters f- 1 0
In the Pensacola experiments,the OD \ N O OS O
performance of the unit was tested. O
periodically by the examination of
samples of water, as discharged
from the treatment unitand water
and oysters in a tank, used as a
control,which was fed UV treated
water. A summary of the results obtained in 18 experiments, conducted
during a period of nine months, is presented in Table 4.
Of the 176 samplesof UV treated effluent, 98.3 per cent showed
the absence of coliform organisms from three 100 ml portions, equivalent
to MPN less than 0,41, Two, or 1.2 per cent, of the samples showed one
of the three 100 ml portions to be positive for coliform organisms and
one showed an MPN of more than 0.41 per 100 ml. Of the 257 samples of
water examined from control flat C, 95 per cent showed the absence of
coliform organisms from five 10 ml volumes, 3.1 per cent showed coliform
MPN's between 2.0 and 4,5 (one or two 10 ml portions positive) and 1.9
per cent showed coliform MPN's in excess of 7.8. Oysters in the control
tank were also examined. More than threequarters of 257 samples tested
showed the absence of coliform organisms from 1 gm quantities, equivalent
to MPH less than 18, and 95 per cent showed coliform densities of less
At Purdy, the comparative raw water and UV treated water MPN's
were determined during the course of eight experiments on the rate of
accumulation and elimination of bacteria by oysters conducted from
November 1959, to June, 1960. The results are shown in Table 5. A total
of 95 samples of raw water were tested for coliform density. The coliform
MPN of the raw water ranged from less than 1.8 to 2,400 and the geometric
mean was 32. All but four of the 95 samples of the effluent from the
ultraviolet treatment system collected simultaneously showed the absence of
coliform organisms from five 100 ml portions, equivalent to an MPN of less
than 0.18. Each of the four samples showing positive coliform results
were found to have one of the five 100 ml portions positive, equivalent to
an MPN of 0.2 per 100 ml.
During the same experiments, 78 samples of water were taken from
the control tank containing oysters and receiving ultraviolet treated water.
The distribution of coliform MPN's is shown in Table 5, Seventy-three, or
approximately 94 per cent, showed the absence of coliform organisms fror
100 ml volumes. Of the remaining five samples, two were at MPN 0.2 and one
each at MPN 0.68, 2.3 and greater than 13.
The turbidity of 91 raw water samples was determined by comparison
with sta nard turbidity suspensions prepared as described in Standard
Methods. To test the effect of turbidity on the efficiency of ultraviolet
radiation, the samples were sorted into four turbidity ranges; i.e., zero
to trace, 1 to 5, 6 to 10, and 11 to 20. From these plots the average and
90 per cent probability range of occurrence were derived.
The ranges of coliforms MPN of the raw water and the geometric
averages at three turbidities are presented in Table 6. A. progressive
increase in coliform MPN, parallel with the increase in turbidity, can be
seen. The results obtained on the samples of treated water are also pre-
sented in Table 5 in terms of percentages of samples in each turbidity
group showing coliform MPNts less than 0.18, 0.2, and greater than 0,2.
Table 4-"Goliform MPN's of ultraviolet Treated Sea Water and of Water and
Oysters from Control Flat C.
Water, UV Effluent Water, Flat C Oyster, Flat C
MPN MPN MPN
0.41 <1.8 2.0-4.5 >7.8 <18 20-45 78- =49 4o
No.* 173 2 1 244 8 5 200 45 11 2
% + 98.3 1.2 0.5 95 3.1 1.9 77.5 17.5 4.3 0.8
*Number of samples showing stated coliform MPN's or less.
+Per cent of total samples of each type examined.
Table 5--Coliform MPN of UV Treatment Tank Effluent and Water in Control
Number of Samples Per Cent Cumulative Per cent
MPN UV UV UV
per 100 M1 Treated Tank C Treated Tank C Treated Tank C
<0,18 91 73 95.8 93.6 95.8 93.6
0.20 4 2 4.2 2.56 100.0 96.1
0,68 0 1 1.28 97.4
2.3 0 1 1.28 98.7
>13 0 1 1.28 100.0
samples 95 78
The effect of turbidity on the germicidal efficiency of the
treatment unit can be seen only from the increase in percentage occurrence
of MPN 0.2 with increases in turbidity. It should be noted that, although
coliform organisms were recovered from treated water samples, the highest
density determined was 0.2 per 100 ml, even in the presence of turbidities
up to 20 ppm.
The use of chlorine for the sterilization of sea water in shell-
fish purification plants is contraindicated, since it has been established
that chlorine has an adverse effect on the feeding activity of both the
Eastern oyster and the Olympia oyster. Further, the experiments have
Table 6--Effect of Turbidity on Efficiency of UV Treatment Units
Raw Water UV Treated V:cr
Turbidity Number of Coliform MPN/100 Ml Per Cent ot S:prl. .Showing
Range Samples Average Range Colifc.-m MPN
<0.18 0,2 0.2+
O-trace 40 17 <1.8- 330 100.0 0,0 0
1-5 24 17 <1.8- 240 92,0 8,0 0
6-10 18 61 7.8- 920 94.4 5.6 0
11-20 9 160 23.0-2400 87.5 12.5 0
(All samples)91 32 <1.8-2400 95.8 4.2 0
that dechlorination does not remove the factor which influences the feeding
activity, On the other hand, it has been demonstrated that ultraviolet
trea.ed water does not have a similar effect and there is experimental
evidence to demonstrate that oysters will purify themselves in such water.
Turbi,-ities of up to 20 ppm did not seriously interfere with efficiency
of tLe Purdy ultraviolet treatment unit. Obviously, a water containing
gros3 particles of particulate matter should be clarified by settling to be
acceptable for ultraviolet treatment.
There are at the present time means for measuring the
efficiency of operation of the ultraviolet lamps. A germicidal watt meter
is available to measure the direct ultraviolet output of the lamps.
Ins'tru.ments for measuring ultraviolet intensity at short distances from
bare tubes such as might be desired in ultraviolet treatment units are
available and more sensitive meters can be obtained to measure the intensity
at geatoer distances. All can be combined with suitable electronic re-
cording devices to produce a continuous record of the UV energy applied
to the water.
Data from results of experiments on both artificially
polluted sea water, and on sea water containing coliform organisms up to a
density of 2,400 occurring naturally in the water, indicate that ultra-
violet radiation is an effective means of destruction of coliform organisms.
The unit developed for the treatment of sea water in the laboratory
experiments on oysters at Purdy, Wash., is of a design and capacity
sufficient to evaluate the applicability of ultraviolet radiation in at
least small scale commercial installations. The unit accomplished more
than 99,96 per cent reduction of coliform organisms from a sea water seeded
with sewage and having an average coliform MPN of as high as 580 per 100 ml.
Percentage reduction in 350 and 200 plate counts were of the same order.
Pilot scale studies conducted over extended periods have demon-
strated that, with proper maintenance, the unit will operate at a high
degree of dependability. With raw water ranging in coliform MPNzc ,vs to
2,400 per 100 ml, and turbidity up to 20 ppm, only 5 per cent cf 91 amples
of UV treated water examined during a period of eight months oho(:ed the
presence of coliform organisms in 100 ml volumes tested, The mezirxiin MPN
of the treated water was 0.2 per 100 ml. The maintenance requirement of
these units are of the simple housekeeping nature. For proper operation
the units should be flushed periodically and the lamps and reflectors
cleaned to remove accumulated deposits*
1. Luckiesch, M., and Holladay, L. L. Disinfecting Water by Means of
Germicidal Lamps. General Electric Rev. 47:45-50, 1944.
2. Gilcreas, F., and DeLalla, L. Application of Ultraviolet Ray
Sterilization to Water Treatment. New England Water Works A.
(June), 1953, PP. 130-139.
3. Chaumeau, R. Sterilisation De LfEau par les Radiations Ultra-
violettes. LIEau 46:263-264 (Nov.), 1959.
4. Buttolph, L.; Haynes, H,; and Matelsky, I. Ultraviolet Product
Sanitation. Bull, LD-14, Cleveland, Ohio: General Electric Co.,
5. Kawabata, T., and Harada, T. The Disinfection of Water by Germicidal
Lamp. J. Illumination Soc. (Japan). 36:89-96 (Mar.), 1959.
6. Okinami, M.; Sakai, M.; Kurata, Y.; Fujita, T,. Massaubara, T.; and
Furno, H. Public Hygienic Study of Cultured Oyster on Hiroshima
Bay. The Artificial Purificatory Experiment of Oyster. Hiroshima
Prefectural Public Health Section. Rep. No. 4, 1952.
7. Kamimura, M., and Suzuki, N. The Investigated Report on the
Purification Facility of Cultured Oyster Institute, Sate Oyster
Nursery at Matoya Bay Japan Export Frozen Marine Products Inspection
8. Wood, P. C. The Utilization of Ultra-violet light in the Purification
of Oysters. International Council for the Exploration of the Sea.
C, M. 1959, Shellfish Committee, No. 41, 1959.
9. Sanitary Control of the Shellfish Industry. 1959 Manual of Recommended
Practice. Part I: Growing Areas, p. 21. Washington, D. C.: Depart-
ment of Health, Education, and Welfare, PHS, 1959.
10. Public Health Service Drinking Water Standards. Pub. Health Rep.
61:11 (Mar. 15), 1946.
11. Kelly, C. B.; Arcisz, W.; Presnell, M. W.; and Harris, E. K.
Bacterial Accumulation by the Oyster Crassostrea Virginica on
the Gulf Coast. Cincinnati, Ohio: Robert A. Taft Sanitary
Engineering Center, Tech. Rep. F60-4, 1960.
12. MacLeod, R. A.; Onofrey, Eo~ and Norris, M. E. Nutrition and
Metabolism of Marine Bacteria: I. Survey of Nutritional Require-
ments. J. Bact. 68:680-686, 1954.
13. American Public Health Association. Standard Methods for the
Examination of Water, Sewage, and Industrial Wastes(Tenth Ed.).
New York, N. Y.: The Association, 1955.
Mr. Kelly is chief, Shellfish Sanitation Laboratory, Milk and
Food Research, Robert A. Taft Sanitary Engineering Center,
Public Health Service, Cincinnati, Ohio,
This paper was presented before the Laboratory Section of the
American Public Health Association at the Eighty-eighth Annual
Meeting in San Francisco, Calif., November 2, 1960.
Reprinted from AMERICAN JOURNAL OF PUBLIC HEALTH, vol. 51, no.
11, November 1961. Copyright by the American Public Health
Association, Inc., 1790 Broadway, New York, N. Y,