POSTGRAD. MED. J. (1966), 42, 703.

AIR POLLUTION,BRONCHITIS AND LUNG CANCER

P. J. LAWTHER, M.B., F.R.C.P.

Director, M.R.C. Air Pollution Research Unit, St. Bartholomew's Hospital Medical College, London, E.C.1

THE RELATIONSHIP of air pollution to disease iscomplex. The topic provokes emotional attitudesin many and is of political and economic importance.Much confusion exists regarding the nature ofpollution and its real or suspected effects on manand this essay is written in order to try to lessenthe confusion rather than to provide a formalreview of the subject.For many centuries British urban air has been

contaminated by the products of combustion ofcoal; industrial processes have added their contri-bution and in this country we have seen the additionof pollution from the burning of petroleum. Ideally,carbon-containing fuels should yield only carbondioxide and water when burned but usually manyother compounds are emitted. These may beproduced as a result of incomplete combustion,distillation of volatile matter, combustion ofimpurities in the fuel or fixation of atmosphericnitrogen. If some chemically simple fuel such asnatural gas is burned in an inadequate supply of airsmoke consisting of fairly pure carbon (lamp black)and carbon monoxide would be produced inaddition to the carbon dioxide and water. Onewould also expect to find traces of polycyclicaromatic hydrocarbons formed by pyrolysis. If thegas contained sulphur compounds, sulphur dioxidewould be emitted in company with smaller amountsof sulphur trioxide if the temperature of combustionwere high enough. If the gas, or indeed any fuel,were burned at high temperature and under pressuresignificant fixation of atmospheric nitrogen wouldoccur and oxides of nitrogen would be emitted.The production of oxides of nitrogen is greaterwhen temperature and pressure are high and thereaction products are removed rapidly from thezone of combustion. High compression ratiointernal combustion engines produce oxides ofnitrogen in this manner.The combustion of coal is more complicated;

when coal is burned in the grossly inefficientdomestic open grate a series of stages in theemission of pollutants may-.be recognised. When ashovelful of coal is added to a hot fire much volatilematter is distilled off and is manifest as greenish-yellow tarry smoke containing very little soot

since it is not a product of combustion. Eventuallysome of the volatile matter which includes coal gasignites and burns incompletely with a yellow smokyflame. Later the carbon residue - which is indeedcoke - burns smokelessly; at this stage much ofthe pyrites is decomposed with the resultantemission of oxides of sulphur. The thermalefficiency of the domestic grate is very low indeed.

Industry could not tolerate anything like theinefficiency so widely accepted in domestic heatingand the pollution produced by industrial coal-burning plants differs accordingly; a boilerman triesto achieve as near complete combustion as possibleand the concentrations of carbon dioxide, oxygenand carbon monoxide in fuel gases are carefullymonitored with this object; the emission of blacksmoke is indicative of gross failure to achieveeconomic combustion and the production of tarryvolatile matter such as that emitted by the domesticdistillation of coal is virtually impossible in any butthe smallest furnaces. In general it can be said thatoil is more easily burnt without smoke productionsince it can be dispersed in fine droplets or as avapour so that combustion may be more efficient.

There is a further distinction between domesticand industrial pollution which is highly relevant tothe study of its clinical effects; pollution fromdomestic sources is emitted at low level whilstindustrial stacks are usually high enough to avoidgross contamination of the air at ground level.Figures frequently quoted relating to gross emissionsof various pollutants may therefore be whollyirrelevant to the study of their effects for it is theconcentration at ground level which is important.Ground level concentrations are of course

influenced by the amount of pollutant emitted aswell as the height of the source but an even moreimportant factor is the weather. Normally pollu-tants are dispersed into the upper air by thebuoyancy of the plume and by the upward move-ment of warm air from the ground; they are alsogreatly diluted by wind. Episodes of extraordinarilyhigh air pollution occur during the temperatureinversions which accompany winter anticyclones.Then the normal turbulence of the air near theground ceases and stagnation results. Pollutants

copyright. on July 4, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.42.493.703 on 1 Novem

ber 1966. Dow

nloaded from

704 POSTGRADUATE MEDICAL JOURNAL November 1966

X8.............................

81.g.

..............

FIG. 1.-Electron micrograph of general pollution x 11,000.

emitted near the ground accumulate but someplumes, although originating below the inversionlevel, may, by virtue of their velocity and buoyancy,pierce the ceiling and be dispersed by the windsaloft. Fog frequently accompanies these temperatureinversions and the water droplets of which it iscomposed impair the dispersion processes and inaddition favour the oxidation of sulphur dioxide tosulphuric acid.

Air pollution is measured widely in this country,generally over periods of 24 hours, in terms of smoke(determined from the stain on a filter paper) andsulphur dioxide. These measurements are valuablefor epidemiological research but more detailedinvestigations are necessary for experimental work.

Particulate matter in the air has been the subjectof intensive study by means of optical and electronmicroscopy (Fig. 1). Average and maximumconcentrations of various common pollutants aregiven in Table 1.

Pollution may produce, or may be suspected ofproducing, several effects on man. Episodes of

especially high pollution are accompanied by sharpincreases in mortality and morbidity. The deathsand illness occur among the feeble members of thecommunity - the very young and old, and thosesuffering from cardio-vascular or respiratorydisease. These episodes are often mentioned bypatients as marking, or causing, the onset of theirchronic bronchitis; repeated exposure to highpollution or chronic exposure to "chronic" pollutionhas been blamed, with fair reason, for causingchronic bronchitis and lung cancer. These effects orsuspected effects will be considered critically.

Undoubtedly the clearest connection betweenair pollution and disease is seen during the episodesof intense pollution. Distress among patients withchronic bronchitis is common (though not all areaffected) and is usually manifest by cough, wheezingand increased dyspnoea. The effects would seemto be due to increase in airway resistance in responseto irritation. An immense amount of work has beendone in many centres in an effort to identify theirritant component. Much attention has been

copyright. on July 4, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.42.493.703 on 1 Novem

ber 1966. Dow

nloaded from

November 1966 LAWTHER: Air Pollution, Bronchitis and Lung Cancer 705

TABLE IMAXIMUM AND TYPICAL WINTER CONCENTRATIONS OF SOME

POLLUTANTS MEASURED IN CENTRAL LONDON, 1954 - 64Maximumi Typical winter

conc. conc.Smoke (mg/cu. mm.) 10 0.2Carbon monoxide (p.p.m.) 360 x loxSulphur dioxide (p.p.m.) 2.0 0.2Sulphuric acid (mg/cu. mm.) 0.7 0.01Nitric oxide (p.p.m.) 1.1 0.05Nitrogen dioxide (p.p.m.) 0.2 0.033:4-benzpyrene (pg/100 cu. mm.) 222 5x As measured close to traffic in busy streets: concentration much lower in general atmosphere.

focussed on sulphur dioxide as the most likelychemical irritant but unfortunately it has all beentoo uncritically accepted as the compound whichcauses the damage. Thousands of experimentalexposures have been made in laboratories in manyparts of the world and there is yet little convincingevidence that sulphur dioxide in concentrationswhich might be found in town air (up to 2 parts permillion) can cause the increases in airway resistanceobserved in clinical practice. There are of courseasthmatic subjects who are especially sensitive tosulphur dioxide but rarely if ever have they beenshown to react consistently to concentrations of2 ppm. SO2. Obviously some combination ofpollutants, gaseous or particulate, is more likelyto be the culprit than a single substance. Com-binations of sulphur dioxide and sulphuric acid mist(in realistic concentrations) have been tried, hithertowithout success, and following the demonstrationby Mary Amdur (1959) of a synergistic action ofsulphur dioxide and salt aerosols on the airwaysresistance of guinea pigs. Much experimental workon human subjects has been done to see how findingscould be reproduced in man. Experiments in thiscountry and by Mary Amdur and her associateshave failed to demonstrate the effect on humansubject.We have thought it possible that sulphuric acid

might be the irritant but that in laboratory experi-ments which failed to produce increase in airwayresistance the failure might have been due to therapid dilution of the acid droplets on entering thewarm humid respiratory tract. Experiments which"protected" acid, in which porous particles of char-coal have been inhaled after having been refluxed withsulphuric acid, have so far failed to give positiveresults; they are continuing. Another set of experi-ments in progress in the M.R.C. Unit is castingfurther doubt on the importance of sulphur dioxidealone. Airway resistance measurements (made inthe whole body plethysmograph) are made daily ontwo subjects immediately on arrival at the laboratoryafter a walk of nearly two miles through the City.

A correlation is demonstrable between pollution interms of smoke or sulphur dioxide in one subjectwho is unaffected by experimental inhalations of30 ppm SO2, whereas there is no relationshipbetween pollution and airway resistance in a col-league who is sensitive to sulphur dioxide in con-centrations greater than 5 ppm. These experimentsare in their early stages but a pattern is emergingfrom these results and the results of hundreds ofexperimental inhalations which indicate that theprinciple irritant is far from simple in chemical orphysical nature.

There is, of course, an urgent need to identify theirritant so that its emission may be prohibited ordiminished. The Clean Air Act (1956) prohibits theemission of smoke in specified areas but it is power-less to deal with sulphur dioxide (so far no remotelyeconomic means have been found to remove thispollutant at source) and it is of obvious importanceto determine whether or not the improvements inthe quality of air as a result of the implementationof the Act will be sufficient to diminish the likelihoodof a repetition of the 1952 London "smog."Techniques other than those used in the laboratoryare being employed and these exploit the fact thatthe concentration of smoke is diminishing whilstthe sulphur dioxide content remains virtuallyunchanged. For several years the relationshipbetween pollution (measured in terms of smoke andsulphur dioxide concentrations) and mortality fromchest and heart diseases, and morbidity (measuredin terms of requests to the Emergency Bed Servicefor admission to hospital for respiratory or cardiaccauses) have been studied. In addition, a furthergroup of specially selected patients have been takingpart in a series of winter experiments. In theseexperiments groups of volunteer chronic bronchiticpatients from Chest Clinics in London have beenrecording daily their own assessment of their con-dition as "better," "same") or "worse" than the daybefore the diary entry. These results have beenplotted against daily measurements of smoke,sulphur dioxide, visibility, temperature and

copyright. on July 4, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.42.493.703 on 1 Novem

ber 1966. Dow

nloaded from

POSTGRADUATE MEDICAL JOURNAL

humidity. The early "diary" experiments showed aremarkably close correlation between smoke andsulphur dioxide and the well-being of the group ofpatients (Lawther (1958)). The experiments havebeen continued, as have the surveys of morbidityand mortality, in order to study the possiblediminution in correlation with pollution as smokeconcentrations diminish. These experiments andsurveys need not only many subjects but manyvariations in pollution before results can be producedwhich will indicate with any certainty in which groupof pollutant this irritant lies but, though muchanalysis of current work remains to be done, thereare already indications that the morbidity andmortality would seem to be related less closely thanformerly to changes in pollution. The results to datewould favour the hypothesis that the substance orcombination of substances causing exacerbationsof existing chest disease is related more closely tosmoke than to sulphur dioxide. The applicationof computer techniques to the results obtained overthe last 10 years may be surprisingly fruitful in thatthey may enable the reactions of separate individualsor groups of individuals to different combinationsof pollution and weather to be identified.

Chronic bronchitis is found more commonly intowns than in country districts and its prevalenceis fairly closely related to air pollution (moreaccurately to some supposed index of pollutionsuch as "fog frequency," coal consumption per unitarea or to population density). The supposition is,therefore, that air pollution plays an important rolein the production of chronic bronchitis. Again,both smoke and sulphur dioxide have been blamedoften on wholly inadequate evidence. Chronicbronchitis is a disease (or even more than onedisease) with an ill-defined time of onset and a longnatural history which may or may not merge withthat of asthma, emphysema, and bronchiectasis.And few would claim that air pollution is the onlynoisome feature of urban life which need be sus-pected of causing the disease. Because of theclinical and epidemiological complexities of theproblem the development and use of the M.R.C.Questionnaire on Respiratory Symptoms is mostwelcome for by its application and the use of simplepulmonary function tests the prevalence of symptomcomplexes, rather than diagnosis, can be estimated.By means of this technique Stuart-Harris and Hanley(1957) showed that in Sheffield and in rural Wensley-dale simple bronchitis- the presence of chroniccough with the expectoration of mucoid sputumwas related to cigarette smoking rather than anyother measured factor. This important finding wasconfirmed by the College of General Practitioners(1961) who showed, in addition, that complicatedbronchitis - with expectoration of purulent phlegmaccompanied by recurrent chest infections and

dyspnoea- was related to an urban factor ratherthan cigarette smoking.These important findings might seem at first

sight to be at variance with the history of patho-logical change in chronic bronchitis as demonstratedby Lynne Reid (1958). Undoubtedly the early stagesof chronic bronchitis are characterised by thehypersecretion of mucus and the appropriate hyper-trophy of mucus-secreting tissue. Some wouldassume that this hypertrophy and hypersecretion isin response to the irritant air of our cities yet it isobvious that there is in the air no simple chemicalirritant powerful enough to produce these changes.Cigarette smoke is a powerful irritant and fits thehypothesis well. There is a greater likelihood that airpollution is the urban factor, or one of the urbanfactors, involved in the transformation of simplechronic bronchitis into the far more serious infectedcomplicated phase of the disease. Such a factorneed not be irritant at all. It could be a substancethat favours the development or persistence ofinfection either by interference with the cleansingof the respiratory tract or by actually encouragingthe growth of H. influenzae or Strep. pneumoniae.Preliminary experiments in the M.R.C. Unit haveshown that, at least, extracts of smoke do notinhibit the growth of these organisms.The hypothesis that chronic bronchitis is caused

by simple chemical irritation afforded by airpollution is barely tenable in the absence ofnotorious excesses of the disease in workers inindustries where high concentrations of chemicalirritants are inhaled.A recent study by Douglas and Waller (1966)

provides a refreshing break from the obsession withsulphur dioxide acting as an irritant sufficientlypowerful to cause bronchitis in the general popu-lation and underlines the obvious fact that thecause of the disease is likely to be much more subtle.Douglas started a unique survey by following up5,362 children born during the first week of March1946. The frequency of upper and lower respiratorytract infections was noted in addition to manyother clinical features. Waller estimated thechildrens' likely exposure to pollution and allottedthem to four pollution categories. Analysis of theresults showed that there was no apparent corre-lation between the occurrence of upper respiratorytract infections in infancy and pollution categorybut, in all social classes the occurrence of lowerrespiratory tract infections was related to pollutioncategory. The oft-demonstrated correlation withsocial class (independent of pollution category) wasalso seen.The prospect of success in displaying in full the

aetiological factors in chronic bronchitis by theuse of epidemiological methods is not very brightsince the disease is prolonged and multifactorial

706 November 1966

copyright. on July 4, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.42.493.703 on 1 Novem

ber 1966. Dow

nloaded from

LAWTHER: Air Pollution, Bronchitis and Lung Cancer

and the environmental (and demographic) factorsare altering too quickly to enable a complete pictureto be built up. Such techniques are howeverinvaluable for the comparison of prevalence indifferent countries (such as the U.S.A. and Britain)and in different occupational groups.Lung cancer is commoner in towns than in the

country. So is soot. And soot, which causedcancer of the scrotum in chimney sweeps in PercivalPotts' time contains carcinogenic polycyclic aromatichydrocarbons. Sir Ernest Kennaway's team brilli-antly isolated and identified from coal tar thechemical carcinogen 3:4-benzpyrene. Several otherhydrocarbons in coal tar, and in coal smoke, arepotent carcinogens and it would seem whollyreasonable to attribute the skin cancer of chimneysweeps, tar workers and mule spinners to the actionof polycyclic hydrocarbons. The temptation toattribute lung cancer to air pollution is great sincecoal smoke is a rich source of polycyclic hydro-carbons. There are, however, many serious objec-tions to such simple hypotheses. Lung cancer is adisease which has been becoming commoner sincethe turn of the century and the rise is still continuing.Certainly part of the early reported rise in the inci-dence of the disease was artificial and was due to themore widespread use of diagnostic radiology and togreater awareness of the existence of the diseaseas a pathological entity. There was until the 1940'smuch speculation regarding the validity of theobserved increase but for many years it has beenapparent that the increase in the prevalence of thedisease is all too real. The increase has been mainlyin tumours of the squamous and oat cell types, manyauthorities believing that the frequency of adeno-carcinomata has remained relatively unchanged.Throughout the period during which lung cancerhas increased pollution by coal smoke has decreasedsteadily.

If polycyclic aromatic hydrocarbons were respon-sible for lung cancer in man one might expect to findgreat excesses of the disease among those who byvirtue of their occupation have inhaled largequantities of these compounds. As has already beensaid the distillation and incomplete combustion ofcoal produces tar rich in polycyclic hydrocarbonsand in the past some excess of lung cancer in retort-house workers in gas works have been reported byseveral workers. Recently Doll and the MedicalOfficers of several Regional Gas Boards have studiedthe mortality from lung cancer of gas workers (1965)whilst the M.R.C. Unit (Lawther, Commins, andWaller, 1965) defined the environment in terms ofpolycyclic hydrocarbons. This work showed thatcertain workers in the old horizontal retort-housewere exposed to concentrations of 3 :4-benzepyreneabout one thousand times greater than the winteraverage found in large towns. One would have

expected to find a spectacular excess mortality fromlung cancer in such an occupation yet the excess,though demonstrable, was small and in no sensecommensurate with the vast excesses of 3:4-benze-pyrene in the environment.The urban/rural discrepancy in lung cancer

mortality in Britain has been used as powerfulevidence in favour of the hypothesis that air pol-lution causes lung cancer. Similar urban/ruralratios are to be found in countries in which airpollution is nominal. Finland has the second highestlung cancer mortality in the world yet its air, intown and country, is very poor in hydrocarboncontent. Similar urban/rural gradients in lungcancer mortality are seen in Norway and in manyother clean countries.

In recent years there has been an outcry againstpollution from motor vehicles especially thosepowered by diesel engines which has been blamedby many for the rise in lung cancer. Polycyclichydrocarbons have been found in exhaust smokefrom both petrol and diesel engines. This finding ishardly surprising since these compounds are com-monly found among products of incomplete com-bustion but this information has been applieduncritically and with singular lack of a sense ofresponsibility to the problem of lung cancer.No account has been taken of the fact that thequantities of polycyclic hydrocarbons contributedto the air men breathe are extremely small and areindeed insignificant in comparison to the back-ground pollution by coal smoke (Waller, Commins,and Lawther, 1965). Smoke is produced only bymaladjusted or over-fuelled diesel engines (a smallpuff on starting or acceleration is normal) and wehave found that even the smoke from dirty over-loaded diesel vehicles is one of the poorest sourcesof polycyclic hydrocarbons.Some justify their indictment of the diesel by

pointing out that the rise in lung cancer has coincidedwith the increase in the amount of diesel traffic.They ignore the fact that the rise in mortality fromlung cancer was already well under way when thediesel engine came into general use in the late 1930s.And again, the coincidence of the later rise in lungcancer with the increase in diesel traffic cannot beconstrued as a causal relationship since experiencewith all environmental carcinogens leads one toexpect that an exposure of many years is neededbefore cancer is produced.From this temporal point of view the petrol

vehicle merits more careful consideration since ithas been polluting city air for long enough to besuspected of causing the rise in lung cancer. But ifthis were indeed the case one would expect to see anexcess of deaths from lung cancer among peopleexposed to high concentrations of traffic "fumes."No such excess is seen among garage workers,

November 1966 707

copyright. on July 4, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.42.493.703 on 1 Novem

ber 1966. Dow

nloaded from

708 POSTGRADUATE MEDICAL JOURNAL November 1966

traffic policemen, or indeed any other workerswhose job leads them to be exposed unduly toexhaust products.

All the evidence from many parts of the worldpoints to the appearance at about the turn of thecentury of a new factor which has been spreadingto influence firstly men and latterly women in manyparts of the world. Few can doubt, having scrutin-ised the evidence, that this factor is of the smokingof cigarettes. The report of the Royal College ofPhysicians on "Smoking and Health" (1962) andthe report of the U.S. Surgeon General on the sametopic will surely come to be regarded as the mostimportant reports affecting the health of the publicto be published for many years.When due allowance has been made for smoking

habits there remains evidence that some urbanfactor is involved in the genesis of lung cancer. Ofcourse, air pollution needs to be considered as themost likely factor but the urban excess (which isdeclining) over the rural death rate may yet beshown to be due to some artifact such as betterdiagnostic facilities in towns. Much work, bothexperimental and epidemiological, remains to bedone.Although it is obvious from present evidence that

cigarette smoking rather than air pollution has beenresponsible for the observed rise in lung cancer,new pollutants must be scrutinised lest they cause afurther wave of disease. Asbestos, a widely usedmineral, has recently been shown to be moredangerous than was at once thought and the workof Newhouse and Thompson (1965), which demon-strated the prevalence of mesothelial tumours amongpeople with minimal occupational or domesticexposure should dispel complacency.The Clean Air Act of 1956, passed as a result of

the alarm felt after the 1952 London "smog", is agood act in that it aims to abolish smoke whichcan be abolished. Sulphur dioxide remains as aproblem for which there is as yet no economicsolution. But during the period when we are stilldependent on fossil fuels, with their contaminantsulphur, for heat and power it would seem reason-able to try as far as is practicable to .burn the fuelscentrally and disperse the effluents from tall stacksso that they are unlikely to contaminate the air at

the level at which it can be breathed. This trend isbeing followed by the power industries. Untilabundant nuclear energy or huge supplies of naturalgas release us from dependence on conventionalfuels it behoves us to see that combustion is ascomplete as possible and that pollution of the airat ground level is abated, for whatever the com-plexities of the case against air pollution it can betruly said that it can be lethal, it never did anyoneany good and is always disgusting.

REFERENCES

AMDUR, M. 0. (1959): The Physiological Response ofGuinea Pigs to Atmospheric Pollutants, Imit. J. AirPoll., 1, 170.

COLLEGE OF GENERAL PRACTrIONERS. (1961): ChronicBronchitis in Great Britain: a National Survey, Brit.med. J., ii, 973.

DOLL, R., FISHER, R. E. W., GAMMON, E. J., GUNN, W.,HUGHES, G. O., TYRER, F. H., and WILSON, W. (1965):Mortality of Gasworkers with Special Reference toCancers of the Lung and Bladder, Chronic Bronchitis,and Pneumoconiosis, Brit. J. industr. Med., 22, 1.

DOUGLAS, J. W. B., and WALLER, R. E. (1966): AirPollution and Respiratory Infection in Children,Brit. J. prev. soc. Med., 20, 1.

GREAT BRITAIN. Ministry of Housing and Local Govern-ment. (1956): Clean Air Act, 1956. London: H.M.S.O.

LAWTHER, P. J. (1958): Climate, Air Pollution andChronic Bronchitis, Proc. roy. Soc. Med., 51, 1473.

LAWTHER, P. J., COMMINS, B. T., and WALLER, R. E.(1965): A Study of the Concentrations of PolycyclicAromatic Hydrocarbons in Gas Works Retort Houses,Brit. J. industr. Med., 22, 13.

NEWHOUSE, M. L., and THOMPSON, H. (1965): Meso-thelioma of Pleura and Peritoneum Following Ex-posure to Asbestos in the London Area, Brit. J.industr. Med., 22, 261.

REID, L. (1958): The Pathology of Chronic Bronchitis inOswald, N. C. 'Recent Trends in Chronic Bronchitis,'p. 26. London: Lloyd Luke.

ROYAL COLLEGE OF PHYSICIANS. (1962): Smoking andHealth. London: Pitman.

STUART-HARRIS, C. H., and HANLEY, T. (1957): ChronicBronchitis, Emphysema and Cor Pulmonale. London:John Wiley.

U.S. PUBLIC HEALTH SERVICE. (1964): Surgeon General'sAdvisory Committee. Smoking and Health. Washing-ton: U.S. Govt. Printing Office.

WALLER, R. E., COMMINS, B. T., and LAWTHER, P. J.(1965): Air Pollution in a City Street, Brit. J. industr.Med., 22, 128.

copyright. on July 4, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.42.493.703 on 1 Novem

ber 1966. Dow

nloaded from