Bull. Org. mond. SantJ 1973, 49, 461-469Bull. Wld Hlth Org.

Removal of viruses from sewage, effluents, and waters2. Present and future trends

GERALD BERG 1

Because large variations occur in the concentrations of viruses that enter treatmentplants from season to season and from place to place, and even during a 24-hour period,field studies on the removal of viruses by treatment processes require temporal coordinationof sampling. Quantitative methods for concentrating viruses must be developed to measureaccurately the efficiency of virus removal by treatment processes infield situations. Extendedsettling, and storage of sewage and raw waters, reduce virus levels and deserve furtherstudy. Oxidation ponds must be reevaluated with regard to temporal matching of influentand effluent samples and with special care to prevent short-circuiting. Conventional andmodified activated sludge plants must be reassessed with temporal matching of samples.Coagulation of viruses with metal ions requires field evaluation, and virus removal byfiltration through sand and other media, under constant salt and organic loadings, needsboth laboratory and field evaluation. A comparative study of water disinfectants related tospecific conditions is needed. The toxicity, carcinogenicity, and teratogenicity of productsresultingfrom disinfection must also be assessed. Other mattersfor investigation are: meth-ods for quantitatively detecting viruses adsorbed on solids, the virus-removal capabilityof soils, better virus indicators, virus concentration in shellfish, the frequency of infectionin man brought about by swallowing small numbers of viruses in water, the epidemiologyof virus infection in man by the water route, the effect of viruses of nonhuman origin onman, and the occurrence of tumour-inducing agents in water.

It is not easy to remove all viruses from sewage.But if we are going to remove them, as we should,from all effluents discharged into watercourses thatman comes into contact with, and from the sludgesspread on the land or disposed to the seas, we mustbe prepared to pay the cost.Whatever the cost, in the more affluent areas of

the world, the processes that must be used to ridsewage of viruses will also convert it into relativelyclean water. Rivers, murky too long to be remem-bered otherwise in many places, may run clearagain.Where there is yet no treatment at all, there will

need to be a beginning. In developing countries, therewill be many beginnings. And in those places whereurbanization is not yet the established way of life,the beginning may be better than the current farein the most affluent countries on this earth.

1 Chief, Biological Methods Branch, Methods Develop-ment and Quality Assurance Research Laboratory, NationalEnvironmental Research Center, Environmental ProtectionAgency, Cincinnati, Ohio, USA.

From all of this, the raw waters will benefit, andthe users of such waters and the downstreamcommunities requiring potable waters will sufferless risk.The water and sewage treatment processes dis-

cussed here have the potential to improve the chemi-cal and physical quality of effluents and waters.Whatever the value of these processes, only one(lime treatment) may be capable of removing allviruses from effluents or waters. Such treatmentprocesses for virus removal must therefore generallybe considered as adjuncts, with reliance for completeremoval being placed upon disinfection (2).

In this circumstance, one cannot justify directingtreatment processes primarily toward virus removalalthough an argument may be found for directingprocesses toward the removal from effluents andraw waters of those substances that interfere withthe disinfection process.For these reasons, the processes that require further

study, with the single exception of lime treatment,are those directed primarily at the chemical and

3141 - 461-

G. BERG

physical improvement of effluents and waters. Thevalue of knowing the virus-removal capability ofthese processes is that of better understanding, ateach step in the treatment chain, of the. extent of theproblem that has eventually to be dealt with.The single exception referred to, lime treatment,

produces high alkalinity. Unlike other treatmentprocesses, alkalinity does not remove viruses butkills them, and it can kill them quickly. Since highpH is a natural manifestation of the lime process,it may be worth while at times to increase pH suffi-ciently to achieve total disinfection.

TREATMENT PROCEDURES

The greatest shortcoming of virtually all fieldstudies on the removal of viruses by treatmentprocesses is the lack of temporal coordination insampling. Recent studies in my laboratory haveshown large variations in the numbers of virusesentering a sewage treatment plant during a 24-hourperiod. Clearly, an influent sample cannot beconsidered a match for an effluent sample taken atthe same time. A similar situation exists for rawwaters entering treatment plants.

Field studies are fraught with other problems aswell. For example, there are not sufficient viruses insewage or in raw waters to permit following thelogarithmic removal rates by treatment processeswithout concentrating the viruses first, the quantita-tive recovery of viruses being a research problem initself. Moreover, sewage and raw water compositionsvary from area to area, and even in the same plantthey may vary from moment to moment.

There is no easy way to solve these problemsexcept by greater effort. Consistent data from agiven process, studied repeatedly in several areas oneffluents or on raw waters of different qualities, willindicate the reliability of that process. Methods forbetter quantitative detection of viruses in sewageand effluents must also be developed, and it mustbe established that the effluent sampled is from thesame fraction as that from which the influent issampled.

The primary treatment system

Primary treatment, as it is normally practised,seems to remove few of the viruses from sewage.Such, at least, appeared to be the case in experi-ments where viruses were seeded into sewage (6).The same data indicated, however, that greaternumbers of viruses disappeared from sewage when

settling was allowed to continue for 12-24 h, and thatthe proportion of solids settled in this period wasconsiderably greater than that of seeded virus thatwas settled. Of course, seeding experiments mayyield misleading data because viruses in effluentsderive from faecal material, in which most of theviruses excreted are deeply embedded and adsorbed.Thus, the numbers of viruses that are settled insewage are probably largely proportional to theamounts of solids that are settled and are probablygreater than what was apparent.An increase in settling time, perhaps in static set-

tling basins to avoid the short-circuiting that oftenoccurs in continuous systems, might be worthinvestigating as an inexpensive means for reducingthe virus load upon receiving waters. The passageof time alone, especially at elevated temperatures,in the noxious environment of sewage contributesto the destruction of viruses. It might well be thatextended storage of sewage, even for 24 hours, isnot economical in some areas of the world, but inothers, especially where more advanced treatmentsystems cannot be considered, longer-term storage(extended settling) may be a most useful virus redu-cing system.

In smaller communities, batch storage or staticoxidation ponds may provide an inexpensive treat-ment method for reducing the virus content of sewageeven more effectively than primary settling systemsdo. Smafl ponds need only to be filled with sewagethat is allowed to remain for as long as preliminarytests in those ponds demonstrate is necessary.Although it may still allow the discharge of con-siderable numbers of viruses to receiving waters,extended storage may remove more viruses fromsewage than we expect today from some secondarytreatment processes.Anaerobic systems for individual households also

need to be tested. Again, emphasis should be placedon extended storage systems, even when these mustbe of a continuous nature.Determination of the efficiency of virus removal

requires better methods for recovering virusesembedded and adsorbed within faecal and othersolid materials. It is also worth investigating thevirus levels in effluents after increased settling periodsand relating this information to the costs of storage,especially in less affluent communities. Any settlingprocedure results in the removal of large numbers ofviruses, which are carried down with the solidmaterials. Some minimum settling, if only for just afew hours, should be the basis for any treatment

462

REMOVAL OF VIRUSES FROM SEWAGE

procedure, even if it constitutes the entire treatmentprocedure.

Oxidation ponds

The usual form of long-term storage is the continu-ous-flow oxidation pond, which according to avail-able reports removes viruses with erratic efficiency.New studies with an emphasis on careful temporalmatching of influent and effluent samples are there-fore needed.With the oxidation pond and other long-holding

processes, such matching will not be simple, requiringperhaps continuous sampling, and even continuousseeding with viruses, a practice that is sometimesnecessary even though it may yield misleading databecause many viruses are normally adsorbed withinfaecal material. Only preliminary testing will de-termine the conditions for further experiments thatare necessary. It would probably be wise to carry outthese experiments in the areas where the treatmentfacilities are to be installed. The aim of these studiesshould be to determine the rate of virus removalrelative to retention time, and this will be dependenton temperature, sunlight, and perhaps the conditionand microbial flora of the soil that contains thepond.

Activated sludge

From laboratory experiments and field experienceit seems clear that for removing viruses the activatedsludge system is superior to other biological systemsnow in use. Although aggregate data on virus removalby this system are less erratic than those with othertreatment systems, the field data are not consistent.The latter usually indicate moderate, and sometimesexcellent, virus removal when a plant is operatingwell. But sometimes, even when a plant appears byall other indicators to be acting well, viruses are asprevalent in the effluent as in the influent. In poorlyoperating plants, of course, this would be expectedand usually occurs.The most probable explanation for the inconsistent

removal of viruses in well-operated activated sludgeplants is the incoordination in time during sampling.Almost invariably, influent and effluent samples aretaken at the same time, and the one cannot cor-respond temporally to the other. Thus, aside fromattesting to the general fact that viruses are removedby activated sludge treatment in a plant that isoperating well, the data from past field efforts areof limited use.

It is therefore necessary to reassess the conventional

activated sludge procedure in the field by means oftemporally matched sampling, preferably of a con-tinuous kind, over a period of time, perhaps24 hours for each experiment. Owing to the variablenature of effluents, the validity of any conclusionswill depend on the consistency of data obtainedfrom different plants with sewage of differentqualities.

In recent years, modifications of the activatedsludge procedure have been studied and appeardestined in time for wider use. These modifiedsystems, e.g., extended aeration and systems thatremove ammonia, may have great virus removalcapabilities and are being studied carefully in thiscontext (Safferman et al., personal communication).

Coagulation

Coagulation, long a water treatment process, isnow an important unit process in advanced wastetreatment and renovation systems. It can be appliedto secondary effluents in a tertiary treatment mode,and may even replace biological treatment methods.It seems destined to remain the basic process ofrenovation, and is apparently the most effective virus-removing process except for disinfection.Although several good laboratory studies have

been reported on the removal of viruses from efflu-ents and waters by coagulants, these studies needto be broadened to include a greater variety ofeffluents and other waters. More information isespecially necessary on primary effluents. Eventu-ally, such studies will need to be extended to thefield, first on a pilot scale, and later with full-scaleoperations as better methods for concentratingviruses quantitatively become available.

Because of the high pH produced by lime, studieswith this coagulant will have to be expanded beyondthose with other coagulants. Since high pH levelsdestroy viruses (in contrast to virus removal byflocculation and precipitation, in which removal isalways partial), retention of coagulated effluents orother waters, either in a settling basin or afterfiltration, could bring about extensive, and possiblycomplete, virus destruction.

Studies of this kind must be undertaken at thelaboratory level first. The lime levels required toreach high virucidal rates have to be determined ina variety of effluents and other treated waters, andtime-concentration relationships must be developedto establish the economic feasibility of increasedlime concentrations and extended retention timesthat may be necessary for alkalinity to be used as

463

G. BERG

a terminal disinfectant. If laboratory-scale studiesdemonstrate feasibility, pilot- and full-scale plantstudies will be required next.

Filtration

Although sand does not adsorb viruses, organicand other substances trapped in sand often do.Thus, as organic matter accumulates in sand, thelatter may retain viruses. Salts and pH also play arole.With the recognition that organic and other sub-

stances present in a sand filter affect the extent ofvirus removal, it becomes necessary to continue theearlier experiments that produced erratic results,and to determine the relationship of filtration rateand virus removal at constant pH levels and underconsistent organic and salt loadings. The efficacy ofvirus removal by sand filtration must also be de-termined with different kinds of effluent and otherwaters over a range of filtration rates. Mixed mediafiltration must also be evaluated.

There is also value in establishing optimum con-ditions for the coagulation-filtration process. Al-though coagulation itself effects relatively goodvirus removal, an optimum carryover of floc canimprove the efficiency of virus removal in an other-wise ineffective rapid sand filter.

Carbon

As with all other treatment systems capable ofremoving viruses from water and effluents, carbonrequires continuing study. Although it is not anexciting system for this purpose, carbon does retainviruses. The extent to which it does this and theextent to which it acts as a filter by adsorbing orfiltering out other materials that adsorb viruses isworth studying.Carbon, of course, is important in the disinfection

process because it removes organic compounds thatinterfere with disinfectants. Carbon may also altereffluent quality in a way that adversely affects dis-infection. For example, iron used for coagulationmay react in carbon columns and subsequentlyreduce iodine making it unavailable for disinfection.Problems of this kind must always be monitored.

DisinfectionAlthough some treatment procedures can remove

substantial numbers of viruses, none (with the pos-sible exception of high pH processes) is able toremove all viruses. Even an extensive series of bio-logical, chemical, and physical processes (as in total

renovation) cannot be relied upon to remove allviruses. To ensure virus-free effluents, and renovatedand potable waters, terminal disinfection must bepractised.

Chlorine has been the standard disinfectant inthe USA and in other parts of the world for morethan half a century. In recent years, ozone has beenadopted in some areas of Europe and elsewhere.But the basic information necessary for the intelli-gent use of chlorine or ozone has never beenavailable.

Chlorine is, in fact, a rapid virucide in clean watersdevoid of ammonia and organic compounds and atneutral pH levels, where it exists as highly virucidalhypochlorous acid. At higher pH levels, as the hypo-chlorite ion, it is a much slower virucide; in thepresence of ammonia, as an ammonia chloramine,it is slower yet; and in the presence of organicnitrogen, as an organic chloramine, it is an evenslower virucide. Only recently have good data becomeavailable on ozone, but these are incomplete (Shuvalet al., personal communication).What is needed is a comparative study to deter-

mine which disinfectant is best suited to any givencondition. Hypochlorous acid is certainly an excel-lent disinfectant for nonturbid waters that are freeof ammonia and organic compounds, and are dest-ined for potable uses. But many questions still needto be answered. To what extent, for example, doesturbidity affect disinfection ? Since ozone is un-affected by ammonia, would ozone be more suitablethan chlorine for effluents ? What about iodine,which is also unaffected by ammonia? And bromine?Does ultraviolet light have real potential for dis-infecting shellfish depuration water, or is heatpasteurization a safer and better method ? Whatabout ionizing radiation?There is much to be done, and none of the many

isolated experiments with disinfectants can substitutefor adequate kinetic studies that will compare differ-ent disinfectant substances under essentially identicalconditions against standardized strains of viruses (4).

Soil filtration

The effectiveness of soil for virus removal becomesa most important matter with increasing interest ineffluent and sludge disposal by spray irrigation andland spreading, and with recognition of a virus hazardin solid waste leachates from landfills (Peterson,personal communication).

Initial soil filtration studies can be undertaken inthe laboratory by utilizing natural soils and deter-

464

REMOVAL OF VIRUSES FROM SEWAGE

mining the extent and rates of virus passage throughthem. Experiments in the field must also be under-taken, first by seeding and subsequently by followingthe passage of naturally occurring viruses in efflu-ents, solids, or solid waste leachates. The laboratoryand field seeding experiments may give valuableinformation about the survival of viruses adsorbedwithin the soil by following the release of virusesinto the filtrates long after seeding has ceased. How-ever, final resolution of this question must await thedevelopment of suitable methods for recoveringviruses from solids.There is no way to generalize the effectiveness of

soil filtration on the basis of tests on a few samplesoils in a few sample areas. However, parallel testson a large variety of different soils from a largenumber of different areas may be helpful in lendingsupport to generalizations. In the end, it could bethat each soil through which viruses must pass mayhave to be tested for its capacity to retain thoseviruses.

DETECTION OF VIRUSES

Detection of viruses in the water environmentIn order to eliminate viruses from the water

environment, techniques must be developed todetect small numbers of viruses in large volumes ofwater. The important matter of detecting viruses onand within solids in various waters is discussed inthe following section.The standards recently recommended for renovated

and potable waters are based on disinfectant deter-minations, supported by bacterial and viral deter-minations (3). The demonstration of the absenceof certain bacteria and viruses thus endorses theadequacy of the disinfection process, and therebythe safety of the water. Sensitive detection methodsare also necessary to demonstrate the hazard ofviruses in recreational waters and in the vicinity ofcommunity water intakes. In the USA, the demon-stration of viruses in sewage effluents, in receivingwaters, and at water intakes has been an importantfactor in reducing the level of some domesticpollution.The methods available for detecting small numbers

of viruses in large volumes of water are still ineffi-cient, however, and the development of such methodscontinues to be a prime need in water pollutionresearch. At present, a membrane filter technique isbeing developed. In clean waters containing salt,viruses are adsorbed efficiently by cellulose nitrate

membranes at pH 7, and these adsorbed viruses canbe efficiently eluted and quantitatively assayed. Theefficiency is high with volumes up to about 95 litresof water but begins to diminish after that point (5).Much better adsorption apparently occurs at pH 3when heavy organic loading occurs, as in the case ofsewage effluents (10), but the efficiency of virusrecovery by this method has still to be determined.This, however, is no easy matter, if only because ofthe problem of measuring the numbers of viruseswithin the particulate substances, especially thefaecal matter, in sewage.The efficiency of the membrane filter procedure

has been variable in tap waters seeded with virusesand in surface and sea waters. Studies with multiplefilters arranged in tandem, including some forremoving particulate matter and interfering solublesubstances, are still under evaluation (11). Ultra-filtration techniques as with hollow fibres, are alsonow under consideration. Simpler quantitativemethods for eluting viruses from membranes thanthat requiring ultrasonic treatment also need to bedeveloped (5). The high pH method currently understudy may well be one good starting base for futureefforts (11). However, there is as yet no good indica-tion that the final solution to the problem of recover-ing viruses from large volumes of water is in sight.

Detection of viruses in and within solids

During attempts to recover viruses from naturalwaters, various quantities of solids were recoveredand tested for viruses with techniques that had beendeveloped for the recovery of viruses from cellulosenitrate membranes (5). Often, the number of vilusesrecovered from the solids in river and ocean waterwas equal to or more than that recovered from thewater itself. Subsequent laboratory tests showed thatthe technique used was about 1% efficient suggestingthat the virus content of the solids tested was about100 times the amounts recovered. Most of theviruses in water, it now seems, are associated withthe solids in those waters.A major effort is therefore needed to develop

methods for efficiently recovering the viruses ad-sorbed by and embedded within solid materials. Theproblem will undoubtedly be most difficult withfaecal material, the natural source of viruses excretedby infected hosts, where the viruses may be deeplyembedded and tenaciously adsorbed. Since the trueconcentrations of viruses in faecal solids may befar greater than what we have heretofore recovered,and since faecal solids may constitute an important

465

G. BERG

part of the virus-bearing fraction of solids in naturalwaters, it is difficult to conceive the real meaningof recoveries of 50-75 plaque-forming units of virusper 378 litres of water at intakes (as we have alreadymade).The importance of developing efficient methods

for recovering viruses from solids extends beyondthe natural water environment. The hazards ofspray irrigation with sewage effluents or of fertiliza-tion with sludges cannot validly be assessed withoutmethods for detecting viruses in those materials.The turbidity standard for potable waters presents

another serious problem. How can we be certainthat a water can be disinfected when solids arepresent, if we cannot determine whether viruses orother microorganisms are present within the solids?

INDICATORS OF VIRUSES

Indicators of viruses in waste and natural waters

It has been evident for some time that faecalcoliform organisms are not adequate indicators ofviruses and that positive tests for these or otherbacteria are not inviolate indicators of faecal pollu-tion. Indeed, if they are what they seem to be, faecalcoliform organisms (by definition of faecal origin)indicate an obvious hazard. However, Aeromonasand other bacteria may appear to be faecal coliformorganisms on MFC medium on membrane filtersat 44.5°C. Moreover, faecal coliform organisms maymultiply in enriched waters where pathogens, andcertainly viruses, cannot. And faecal coliform organ-isms may disappear rapidly in surface waters con-taining certain industrial wastes that do not appearto affect faecal streptococci (Kenner, personalcommunication) and may not affect viruses, either.

It is therefore necessary to improve the test forfaecal coliform organisms so that it becomes positivefor them only, or better, to develop a broader bac-terial indicator system for differentiating bacteriaof faecal origin from free-living forms. It is unlikelythat the absence of faecal organisms from a watercan guarantee that such water contains no viruses,but it would be immeasurably helpful to know whena water is likely to contain viruses, in which caseit would also be useful to know whether theseviruses are of human or animal origin, and, if thelatter, from what animal species they originate. Forthis purpose, it will be necessary to detect faecalorganisms that are species-specific, as we can alreadydo to some degree with faecal streptococci. We shall

also need to determine the stability of the indicatorbacteria in various waters and the conditions underwhich they could multiply and make interpretationdifficult.

Indicator organisms in disinfected, renovated, andother potable waters

Faecal coliform (and even total coliform) countscontinue to be relied upon as indicators for thesafety of renovated and finished waters. It is amistaken notion that only pathogens or indicatorsof faecal contamination are of special significancein a water that has been disinfected. Clearly, thepresence of any viable vegetative bacteria in a dis-infected water suggests that the disinfection proce-dure was faulty, since such organisms are likely tobe considerably less resistant to disinfectants thanviruses. It is for this reason that the recently recom-mended tentative standard for renovated and otherpotable waters requires the absence of all vegetativebacteria, and makes no reference to coliform organ-isms of any kind (3). If there should be vegetativeforms more resistant than viruses to disinfectants,they would be of no real significance, althoughmethods would have to be developed to differentiatethem.The research need is apparent. Common vegetative

bacteria that are present in waters have to beidentified and their resistance to water disinfectantsmust be determined. After extensive investigations,if no vegetative forms are found that are more resist-ant than viruses, then any vegetative form survivingin a disinfected water may be taken as an indicatorof faulty disinfection, and a simple method shouldbe developed to differentiate vegetative bacteriafrom spore-forming genera.

If some common bacteria were found to be aboutas resistant as the more resistant viruses, and at leastas numerous, it might be possible to use them asindicators of viruses, because the detection of virusesrequires difficult and expensive procedures (suchdetection is recommended as a " back-up "-i.e. anadditional safety measure-to disinfectant stan-dards). Effort to search for such disinfectant-re-sistant bacteria in the water environment would bemost worth while, provided that the disinfection testscould be performed with the most refined techniquesavailable. These tests should be comparative kineticstudies under carefully standardized conditions thatwould allow the rate of destruction of the bacteria tobe measured against the rate of destruction of entericviruses (4).

466

REMOVAL OF VIRUSES FROM SEWAGE

CONTAMINATION OF SHELLFISH

BY SEWAGE AND EFFLUENTS

The hepatitis A virus and other enteric virusesof human origin are transmitted to shellfish byhuman faecal material, but whether shellfish harbourviruses of animal origin is not known. Manyoutbreaks of shellfish-transmitted viral hepatitistype A have occurred, but no other viruses havebeen shown to be transmitted by this vector. It isnot clear why. It may be that the disease-producingdose is much closer to the infective dose withhepatitis A virus than with other enteric viruses.It may also be that without the distinctive symptomsof hepatitis and the focus of interest on this disease,such transmission is overlooked.The shellfish industry has long been an important

one economically, and many shellfish growing areas

have been closed because of the presence of coliformorganisms, faecal or other, in numbers beyond some

arbitrary standard. Clearly, it is important to learnto differentiate animal pollution from human pollu-tion, and to identify the conditions under whichindicator organisms, whether faecal coliform organ-isms or some other group, truly indicate a virushazard.

It may well be that animal viruses will one daybe found to be hazardous to man, so that suchpollution will become overtly significant. The abate-ment of animal pollution in runoffs from the country-side and pollution of ocean waters or other watersby the myriads of individual home disposal systemsthat enter into water sources present problems oftremendous proportions.

Ultimately, it may be that thorough cooking ofshellfish that so many enjoy raw (and thoroughcooking of all animal foods for that matter) wouldbe more prudent. In any event, the shellfish problemdemands more attention than just the virus-clearingvalue of depuration that it has been getting (8).

THE MINIMAL INFECTIVE DOSE

Of crucial importance in determining the extent ofvirus removal required to assure the safety of a waterare the numbers of viruses present and the numbersthat are needed to infect. There is already availablelimited but strong evidence that the smallest amountsof viruses capable of infecting cells in culture are

capable of infecting man (7, 9). However, all of thestudies, carried out thus far have involved virusesdelivered in very small inocula. One cannot questionthe high infectivity of these viruses for man, but there

is still some question about the frequency of infectionproduced when such small amounts of viruses aredelivered in a glass of water, or in a gulp or two.Such data, though valuable, would be difficult toobtain, especially with wild strains of viruses,because of the risk inherent in such investigations.There would also be great value in determining

disease-producing doses of viruses in man, but suchdata, because of the greater risk, will be even moredifficult to obtain.

EPIDEMIOLOGY OF WATERBORNE INFECTIONS

AND DISEASES

In the final analysis, the actual production ofinfection and disease by waterborne viruses needsto be demonstrated. Past efforts were unsuccessfulexcept with hepatitis A virus and with this agentthe water supply was almost always a small one thatwas inadequately treated.

Unfortunately, all of the efforts to demonstratewater transmission of viruses have been directed atcorrelations with disease. Production of disease,however, is likely to require infection with relativelylarge doses of viruses, doses much larger than oneis likely to find in a water supply or recreationalwater. It would be more cogent to attempt to demon-strate infection of water consumers, which is a morelikely event than actual disease, and to see whetherthere is an increase in disease incidence among theclose contacts of infected consumers, who by excre-ting large numbers of viruses might well be thesource of disease-producing doses.Such studies are likely to be expensive and diffi-

cult to carry through. They would involve large,probably young, probably " captive " populations,some of whom would drink suspect water, whilecarefully matched (preferably familial) controlswould be supplied with other water sources or withsterilized water (1). Infection would be detected bystool or rectal swab sampling.A study such as this would require at least 5 years

for successful completion and a substantial, stable,intelligent population.

HAZARDS FROM THE WATER ENVIRONMENT

Effluents and sludgesSpray irrigation and land spreading are a construc-

tive and productive means for disposing of effluentsand sludges. In many places, some recycling of suchwastes has long been practised. The hazards of trans-

467

G. BERG

mitting viruses by respiratory droplets that may beairborne for long distances, by intrusion into groundwaters, and by absorption into food plants consumedby man or by animals consumed by man need tobe investigated.

In the USA, the hazards of spray irrigation ofeffluents and land spreading of sludges are underintensive discussion. QuestiQns have been raisedabout the effectiveness of sludge digestion in de-stroying viruses. Pasteurization, which is now beingconsidered as a solution to this problem, is the onlypractical method of pathogen control where largequantities of solids must be penetrated in order toreach deeply imbedded viruses and other disease-producing organisms.Some sludge pasteurization studies have been

carried out with bacteria, but not yet with viruses.Moreover, the problem of efficient recovery ofviruses from solids has to be resolved before reliabledata on pasteurization of virus-bearing sludges andeffluents can be obtained.

Toxic, carcinogenic, and teratogenic productsresulting from disinfection

For many decades, little attention was given tothe hazards of low concentrations of chlorine orof the compounds formed in water or sewage bychlorine, because there were no readily apparentill effects upon consumers of chlorinated waters oron the receiving streams into which chlorinatedeffluents were discharged. In the last several years,however, it has become apparent that very low con-centrations of chloramines are highly toxic to certainspecies of fish and to the microscopic forms onwhich fish feed. This toxicity escaped detection formore than half a century, and only after its discoveryrecently was a means ofneutralizing it developed (12).

Clearly, the short-term and long-term toxicity,carcinogenicity, and teratogenicity of chlorine andits compounds for man as well as fish and otheranimals now require careful investigation.

Other disinfectants, such as ozone, already widelyused in some parts of the world, will also have tobe studied carefully to evaluate the potential hazardsof these compounds and of their products.

Virus standards for effluents, renovated, and otherpotable waters

Owing to demand for a virus standard, a tentativeone has recently been presented (3). Since anydetectable virus is potentially a hazard to the con-sumer (7, 9), the permissible amount of virus in a

water that may be consumed should be none. Toachieve this, the tentative disinfectant standardrecommended requires essentially a combination oftime and concentration sufficient to destroy 12 logunits of a reference virus at 5°C. The extent of de-struction is based on current estimates of the maxi-mum numbers of viruses that occur in about 3.78 mil-lion litres of sewage. These are probably under-estimates and may need to be revised upward.The tentative standard also requires that viruses

and vegetative bacteria should be absent from378 litres of test waters. The research necessary toestablish optimum standards is discussed in otherparts of this report.

Effluent standards are also under consideration,but have not yet been set. Although we may consider12 log units of virus destruction to be an adequatetentative goal here too, the problems of how toachieve such destruction in effluents with all of theirsolids, and how to detect all the viruses (and perhapsbacteria as well) that are present have yet to beresolved. It may be necessary to remove fromeffluents all solids and other interfering substancesbefore disinfection can be achieved. These problemsare discussed in greater detail elsewhere in this report.

Animal (nonhuman) virusesLittle is known about viruses of nonhuman origin

in water. Viruses that infect bacteria, fish, and otherforms of water life are native to these waters, butviruses that infect animals, especially those virusesthat multiply in or about the gastrointestinal tract,must enter our waterways in large numbers from pro-cessing plants and with rural runoff.

There is a need to identify these viruses and todetermine their effects on people who consume them.In addition, the importance of this route of trans-mission to the animal species that are their sourceshas to be worked out.

Oncogenic viruses

There have been a number of unpublishedobservations indicating a correlation between che-mical, and perhaps thermal, pollution and tumoursin fish and shellfish. There is therefore clearly aneed to determine whether chemical or physicalagents are responsible for these tumours and whetheroncogenic viruses are transmitted through the waterroute. If chemical agents that induce cancers inaquatic life are present in rivers and streams, it ispossible that these same agents can also inducetumours in man.

468

REMOVAL OF VIRUSES FROM SEWAGE 469

Many aquatic species spend their entire liveswithin restricted areas, which may facilitate tumourstudies in such species. It may be necessary to ruleout certain tumour-susceptible, inbred genetic stocksthat inhabit polluted areas, but this limitation couldbe overcome by studies in several areas and by trans-ferring species from high tumour locations to clean

water captivity where the effects of defined pollutantsmay also be studied.

Studies in man, initially at least, must be epidemio-logical, correlating various types of tumours withthe quality of the water. Owing to the relativelylong life span of man and his mobility, populationgroups for these studies must be carefully selected.

RtSUMt

tLIMINATION DES VIRUS DES EAUX D'EGOUT, DES EFFLUENTS ET DES EAUX:2. TENDANCES ACTUELLES ET PERSPECTIVES

Les procedes de traitement sont en general impuissants& eliminer tous les virus des eaux d'egout et des eauxbrutes. Les recherches doivent donc porter sur les meil-leurs moyens de faire disparaitre des effluents et des eauxbrutes les substances qui contrecarrent le processus dedesinfection. Les techniques qui entrainent une e1lvationdu pH sont particulierement interessantes car une fortealcalinite a un effet destructeur sur les virus.On note de fortes variations de la concentration des

virus a 1'entree des installations de traitement d'une sai-son a l'autre, d'un endroit a l'autre et meme au coursd'une periode de 24 heures. Les 6tudes pratiques surl'elimination des virus par les proced6s de traitementexigent une coordination et une synchronisation des6chantillonnages. Des methodes quantitatives de concen-tration des virus doivent etre mises au point afin d'6valueravec exactitude l'efflcacit6 des traitements.La sedimentation et le stockage des eaux d'6gout et des

eaux brutes entrainent une diminution de la teneur envirus et meritent des etudes compl6mentaires. Ces proce-des peuvent etre specialement utiles dans les pays end6veloppement ainsi que pour les installations domesti-ques. L'utilisation des bassins d'oxydation doit etrereexamin6e en s'efforcant de synchroniser les echantillon-nages des eaux avant et apres traitement et d'eviter les

courts-circuits. I1 en est de meme des installations utilisantdes boues activees otu les nouveaux procedes d'aerationdonnent des r6sultats prometteurs. Les techniques decoagulation a l'aide d'ions m6talliques doivent etreevaluees sur le terrain, tandis que la filtration sur sableou sur d'autres milieux doit faire l'objet d'essais tant enlaboratoire que sur le terrain.

Etant donn6 qu'aucun produit ne r6pond parfaitementaux exigences de la desinfection des eaux, il importe deproceder a des 6tudes comparatives pour determinercelui qui convient le mieux dans une situation donnee. I1faut, en meme temps, evaluer la toxicite, la cancarogeni-cit6 et la teratog6nicite des produits r6sultant de ladesinfection.Des recherches sont aussi indispensables dans d'autres

domaines: methodes pour la detection quantitative desvirus adsorb6s sur des matieres solides; aptitude des solsa eliminer les virus; recherche de meilleurs indicateursde la presence de virus; concentrations virales dans lescoquillages; frequence des infections humaines causeespar l'ingestion de petites quantit6s de virus contenus dansl'eau; epidemiologie des infections virales humaines trans-mises par voie hydrique; effets des virus d'origine nonhumaine sur l'homme; pr6sence d'agents oncogenesdans l'eau.

REFERENCES

1. BERG, G. Hlth Lab. Sci., 3: 170 (1966)2. BERG, G. In: Snoeyink, V., ed. Virus and water

quality: occurrence and control. Proceedings ofthe Thirteenth Water Quality Conference, Febru-ary 1971. University of Illinois Bulletin, 69, p. 126(1971).

3. BERG, G. In: Jenkins, S. H., ed. Progress in watertechnology, Vol. 3. Proceedings of the Sixth Inter-national Conference on Water Pollution Research,June 1972. New York, Pergamon, p. 87 (1973).

4. BERG, G. ET AL. Virology, 22: 469 (1964).5. BERG, G. ET AL. Appl. Microbiol., 22: 608 (1971).6. CLARKE, N. A. ET AL. In: Proceedings of the First

International Conference on Water Pollution

Research 1962. New York, Pergamon, p. 523 (1964).7. KATZ, M. & PLOTKIN, S. A. Amer. J. pub. Hlth,

57: 1837 (1967).8. Liu, 0. C. In: Proceedings of the National Specialty

Conference on Disinfection, July 1970, ASCE, p. 397(1971).

9. PLOTKIN, S. A. & KATZ, M. In: Berg, G., ed. Trans-mission of viruses by the water route. New York,John Wiley, p. 151 (1967).

10. RAO, V. C. ET AL. Water Res., 6: 1565 (1972).11. WALLIS, C. ET AL. Appl. Microbiol., 23: 476 (1972).12. ZILLICH, J. A. J. Water Poll. Cont. Feder., 44: 212

(1972).