Tobacco Documents Online

The hazard of.;ambient cigarette smoke to nonsmokers has been studie& in the past mostly from the cli'nicali point of view, which leaves the question of its social impact unanswered. -The recent Hirayama study provides experimetal information inia special case on the magnitude of : the social;impact.; A theoretical study is carried out here by comparing the average dose of = exposure to ambient smoke of nonsmokers with the dose of exposure to primary smoke of smokers with due consideration of the possible nonlinear effect on low-density exposure. The maximum risk is estimated by assuming a linear response; the result is that the effect of ambient expectancy by,an.;amount between 225 days and 48 days. smoke is ;1/13-of that due to primary smoke. Based on the meager information on the nonlinear- effect,:a minimum estimate is obtained, which reduces the risk of ambient smoke to 1/60 of that of primary smoke. Thus, the ambient smoke is estimate& to cause an excess of deaths J between 50,000.and 10,000 a year inithe.U.S. population of 220,000,000, or.to reduce the 1'ife ~ ~ . INTRODUCTION definitiveTconclusions. Since then, White and Froeb (11980) have demonstrated small-airway .- dysfunction_-,,in nonsmokers chronically expose6 to smoke similar to that in smokers. For the first time,,`there .is definitely a quantitative measurement of a physi'cal change caused by the ambient smoke-, even,though its physiological andiclinical comsequences are not certain. Recently a'Japanese study of.14 years involving the health hazards to 91,540 nonsmoking! women as related to the smoking patterns of their husbands was published by Hirayama (1981). It shows statiistically significant dose-response correlation of incidence of lung cancer in non- smoking wives and smoking by their husbands. Wives of heavy smokers (more than 20 cigarettes a day) have 2.08 times as much chance of dying from lunq,cancer as wives of nonsmokers. Generally, the study concludes that the effect of passive smoking is one-half to one-third' that of direct smoking. Evaluation of the health hazar& of ambient smoke to nonsmokers hinges on two problems. First is comparison of the average dose of exposure to smoke of nonsmokers with that of smokers. If the health effect should be in direct proportion to the dose, then the ratio of exposure, R, is sufficient to determine the effect oninonsmokers--one could simply scale down the known effect on smokers according to the ratio R. There is no assurance, however, that the linear- a ~.. • "~ . . The hazard of cigarette smoking to smokers is well known (Federal Advisory Committee Report 1964; U.S.,Surgeon General 1!979). On the other hand, the effect of ambi'ent smoke to nonsmokers has not recieved adequate attention. Earlier studies on this effect, summarized'in a chapter, "Involuntary „Smokiing";-,in the 1979 Surgeon General's report on smoking! and health, do not yield U3735'106 Peter Fong, Professor of Physics, Department of Physics, Emory University, Atlanta, GA i

. response rel'ation wil'1 hold true down to very low-intensity exposures accumulated'over a long period of time. The body, in such a case, may have enough time to repair the damage so that it wilil not accumulate andlmanifest itself clinically. This nonlinear effect would introduce a reducing factor, n(n<l),, so that the health hazard of nonsmokers would be nR'times that of smokers. To determine the value of ni is the second, problem. The ratio of exposure, R', betweeninonsmokers an6smokers is a problemithat can be studied by physical and statistical analysis and is the main concerniof thiis paper. The nonlinear - reduction factor, n, on the other hand, has to be studied experimentally and clini'cally:" It is a difficult problem because of the poor statistics inherent in low-intensity experiments. The situation is similar to the issue of low-level rad'iiation hazards iin the nuclear power contro- versy (nearly the same as far as cancer hazards are concerned). However, in the nuclear_power issue, proponents andlopponents agree iniassuming,n = 1(perhaps overestimating the hazards) before any definitive information is established experimentally; this is done so that maximum safety will be assured as demanded'by the prevailiing consumerism. The same attitude may be taken here. This is all the more desirable because various environmental hazards are often comoared with one another and'the same standard of evaluation should be adopted. The study of the ratio, R, alone iin this paper (with n assumed to be unilty) will give estimates of maximum risk, which might later on be lessened'whenimore is knowniabout the reduction factor, n. value of the reduction factor, n. As far as lung cancer is concerned, the effect on non- - smokers appears to be quite considerable, indicating that the value of n for lung cancer is cl.oser to one than to zero. However, the report also indicates that the value of n for other hazards is closer to zero. The crude values of n so obtained will be combined with the value of R to arrive at a state-of-the-art estimate of the hazard of ambient smoke to nonsmokers. A nonsmoker is exposed~to ambient cigarette smoke in a variety of ways. These exposures may be classified into two kinds: (1')~low-density exposures taking place in general indoor spaces where the human density is low, such as in offices, plants, and stores, and (2) high- density exposures in commuting vehicles_(incl!uding carpools), restaurants, meeting halls, ' conference rooms, and, in general, all situations involving close proximity to a smoker. For the sake of calculation, the average daily exposure of amaverage working adult is assumed to consist of 11 hours of low-density exposure an6one hour of high density exposure. The remain- ing 12 hours are assumed to be spent at home, which ts not a public place and supposedly can be made smoke free. On the other hand, Hirayama's report (1981) may be interpreted to gain some idea of the The volume of the indoor space to be filled with smoke from one smoker is 800/37%= 2160 ft3. Therefore, the ambient smoke density is one determined by the smoke produced by one smoker and expanded to an air space of 2160 ft3 (61 x 103 L). This smoke density is the one nonsmokers are exposed to. First, calculate the effect whenithe soace i's not ventilated, whichiis not realistic. The result will be used'in the next section to help calculate the effect in more reallistic cases with venti'lati'on. 03735107 Compare the amount of exposure to the toxic part of cigarette smoke in the liung between nonsmokers from ambient smoke and smokers from primary smoke. The ratio of the two, R, in the linear assumption, is to, represent the ratio of the amounts of health hazard (of course, smokers are also affected by ambient smoke). First, consider the exposure E of a smoker to one puff of primary smoke and the exposure En of a nonsmoker to that same pu~f of smoke exhaled by the smoker and dispersed'in the indoor space of 61 x 103 L apportioned to one smoker to fi'11 with smoke. The ratiolof En to Es, defined as R1, i;a productof three factors, rl, r2, and r3 800 ft3 (about one-half of a double hotel room). This corresponds to am indoor population density that is a suitable averaqe--the general secretarial office wnrking space is perhaps more crowded than this, but the plant working space is less crowded. According to the nationali average, 37% of the workers smoke. Every working space with more than three workers will have at least one smoker, on the average, generating smoke. With central' heating and aiir condition- ing, the ambient smoke can be expected to be uniformly distributed throughout the indoor spaces. approximated by an office in whichievery worker occupies a space of 10 ft x 10 ft x 8 ft or Take the exposure of a smoker to the primary smoke inhaled from a single cigarette as the standard and try'to compare with this the exposure of a nonsmoker to 11, hours of low-density ambient smoke. This calculation is based onithe following model. The general indoor space is Low-Density Exposures in Unventiliated Spaces R, = rl r 2 r 3 (1)

( where rl, is the ratio of the densities of the secondary smoke to primary smoke in the lung, ~~ r2 is the ra~'io'of times of exposure of the two cases, and r, is the ratio of toxicities of the two kinds of,smoke, which differ because of the filtering effect of the lung as primary smoke passes through and becomes secondary. These ratios are analyzed as fol'lows. The ratio rl is the inverse ratio of the volumes of the puff of smoke in the secondary and primary state (disregard filltering, which ils taken up in r3)1. Following practices in this field of research,, consider that in each puff the smoker draws a volume of smoke of 35 mL from ;.~ the cigarette into the mouth, which ils then mixed wi,th fresh air and diluted 10 times during inhalationOne-half of it enters the lung and mixes with the 4 L of air there through turbulence. Both halves are eventually exhaled and dispersed'i iinto the space of 61 x 103 L. zi. 'The puff of smoke in the lung is thus diluted~ by 61 x 103/(2x4) times from the primary to the . secondary state. Then 2 x 4 r1 = 10 3 61 . r ~ r x , ~ .,. <, . . , , ..~ ,. The,ratio r2 is estimated as follows. The drawing of the 35 mL of smoke.into the mouth is,." usualily considered to take 2'seconds. The deep inhalationiof this volume into the lung and'thed exhal!ation that.follows take about 3 seconds. The total time the lung is exposed to primary ..~, smoke is not just the 3 seconds because not all smoke is exhaled at once. In fact, only 1 L .,; is exhaled each time and only 1/4 is removed. Each successive breath takes 6 seconds. Thus, the effective'exposure time (normalized to, the densities involved in rl,)l is 3 + 3/4 x 6+(3/4P x 6+(3/4)3'x 6+.., = 27 seconds. A nonsmoker's exposure time to secondary smoke is 11 hours i n thi s model..;`;;Thus, The ratio r has to be included because se been ondar smoke is less toxic after havin g y c filltered by the.l3ung. The lung removes 85% of the volatile material' andipa rtieulates and 50% of the CO.,:,'.``Take an overal!l removal ratio of 75%. This figure is not crucial, as will be shown later. It perhaps overestimates the filtering effect, because CO is more important in causing heart disease, which is ten times as frequent as 1!ung cancer. Thus, 25%. r3 = - = 0.25 1 The total effect,represented'by R'l is thus, -- R1 = rlr2r3 = 0.049 Now compare the exposure E5 of a smoker to the primary smoke of one cigarette and the -exposure En, of a nonsmoker to the ambient smoke produced by the same cigarette. The ratio (5) of E~ to E,, defiined as R2, is not the same as Rl because the ambient smoke consists of not only the secondary smoke of the mainstreamlbut also the sidestream smoke--that part of the smoke from a cigarette when burninq,idle. The sid'estream makes up 60% to 75`0 of the total smoke produced from a cigarette. For convenience, take 2/3 of the total to be in the sid'estream. Moreover, the siidestream smoke is not filtered by the lung, and thus, the toxicity is not attenuated. Therefore, the value of r3 for siidestream smoke iis uniity (or even greater because the mainstream is filtered by the cigarette). Thus, R2 is calculated as follows: R2 = rl r2r3r4 , (,6 ). 03735108 where r is the average value of r over secondary smoke and sidestream smoke and r,l is the ratio r amounts of the ambien~ smoke (secondary plus sidestream) a nonsmoker is exposed to and the primary smoke (mainstream) a smoker is exposed to. (The smoker is also exposed to ambient smoke;, this effect is omitted here but will be included later.) The value of r3 is the average of 01.25 and 1 weighted 1 to 2, which is 0.75. This ratio should be further refined by taking into account the fact that 94':', of cigarette smokers inhale, but only 19"' of cigar snukers and 28^S of pipe smokers inhale. These percentages should be weighted by the respective percentages of comsumption: E34% cigarette, 8''cigar and 8;!. pipe. The final result, r3 = 0.80,

is not much different from the pervious result 0.75. The ratio r, according to the amount of sidestream smoke given above, is 3. The final value of R2 is 0.46. f This means that in unventilated indoor spaces, whenever a smoker smokes one cigarette, the air is so contaminated'that to stay in this soace for 11 hours, a nonsmoker is exposed to an amount of cigarette smoke equivalent to smoking 0.46 cigarette (more exactly, inhaling the primary smoke of 0.46 cigarette). On the average, a smoker smokes a pack of 20 cigarettes a day. The effect of the ambient smoke on nonsmokers will be increased 20 times. Thiis very _ large effect, of course, is not realistic because indoor spaces are ventiTated'i to a large extent, so the actuall effect is much smaller due to the removal of smoke by air changes. The reduction by ventilation will be discussed in the next section. The large effect derived here is nevertheless appl,icablle under ailrtight conditions. Low-Density Exposure in Ventilated Spaces The actual effect of ambient smoke is largely determined by the degree of ventilation. If the windows were kept wide open all' day the year around, the effect of the ambient smoke would be very small. However, this cannot be done in winter when heating,is on. With the increased use of air conditioning, this also cannot be done in the summer. With the advent of the new architectural style (windowless buildings) this cannot be done even in the spring and autumn. The effect of ambient smoke, thus, cannot be made negligible in this way. On the other hand, in modern bui1dings with artifici'al air circulation, the degree of ventilation is well under control and one can calculate the effect of ambient smoke on a quantitative basis. °"'` Generally, in an indoor space there is a certain rate of smoke production by smokers and also a certain rate of smoke removal by fresh air circulation. The two rates become equal in a steady state, which determines the ambient smoke density, a constant in time. One can define the ambient smoke density produced by a person smoking one cigarette in his alloted space of 61 x 103'L with no ventilation by Do g/m L. The smoker's exposure to primary smoke while smoking one cigarette is defined as S g-s. (One unit of exposure is defined as the exposure of one second to one gram of smoke in the lung.) The result obtained in the last section may be rephrased by saying that in an ambient smoke of density D, the exposure of a nonsmoker in 11 hours is 0.46 S q-s. From the rates of smoke production and removal, calculate in the fol'lowinq the steady-state ambient smoke density, De, whichi may be .expressed as a multiple of Do and written as cDa. The exposure in one 11 hour day of a nonsmoker in the i'ndoor space with an ambient smoke density of cDo is thus given by 0.46 c5 g-s. S We define by a the fractional freshiair intake per unit time, the unit being one hour (analogous to the radioactive decay constant). If there were no new smoke production, the ' ambient smoke density would decline exponentially according to the law of radioactive decay, D = Dle-at (7) The rate of increase in new smoke density, corresponding to an average smoker's consumption of one pack of 20 cigarettes a day, or 1.6 cigarettes per working hour, is 1.6 Do g/mL/h. The steady-state ambient smoke denstty De(g/mL) is determine&by the following equationi, which is of the same form as that describing the steady state resulting fro m successive stages of radioactive decay. ,.. _ ADe = 1.6 D, (8) The next steplis the determination of A, from which De/D, or c is obtained. The exposure may ' thus be calculiated'as stated above (0.46 cS g-s). 03735109 In a building with closed-wiindows and arificial air ciirculation, the need for fresh aiir is not so much to counteract the depletion of oxygen and the accuir.ulation of carbon dioxide but to remove objectionable body odors. For this purpose, current engineerinq practice calls for the introduction of 5 to 30 ft3 of fresh air per minute per person (Hutchinson 1966). Because of energy costs, loss of warmed or cooled aiir through the introduction of freshi air should be reduced to a minimumiand', therefore, the rate of fresh air intake is likely to be set near the lower limit. In fact, even the rate of 5 ft3 per minute per person corresponds to a loss of heating and cooling,energy of about $1.50 per person per month, whiich is nearly 60% of the ~

c C per-person neatIng and cooiIng costs ror otr~ices ano noteis. inus,,economy does not aiiow liberal' increase of the rate of fresh air intake. Take for example a rate of 10 ft3 per minute per person This figure can be showni to agree with common sense as follows. In a typical calm day, the opening of a 3-foot window one inch wide should supply enough freshi air to remove' odor in a`double room in double occupancy. Upon the assumption of a wind velocity of 1 mile per hour, which is hardly noticable, the window opening provides a fresh air i!ntake of 11 ft3 per minute per person, comparable to the value 1'0 ft3 per minute per person adopted here. ' of fresh air intake in the model of calculation corresponds to a value of a The rate given by.the following equation: a = (10ift3/800 ft3)/min = 0.8/hi (This may be interpreted by saying that the air undergoes 0.8 turnovers per hour.) From the value of A, the mean lifetime (the reciorocal of a) of the foul air or that of a smoke particle in the indoor space is found to be 1.25 hour. Or, equivalently, the halif-life is 0.87 hours. ` The fresh 'air eirculation would achieve a 75% cleansing in 1.74 hour, which is reasonably good circulati'on. '(Incidentally, the room also may get rid of smoke particulates and aerosol drops through thei'r settling on the floor and the walls; but the mean lifetime for the particles to settle on the floor may be found to be about 20 hours an& that on the walls about 100 hours, far too long,for these processes to have any significance ini smoke removal compared wi,th fresh air circulation. Also, CO can be removed only by fresh air circulation.) , ,;.; ~-Substituting the value of a in the previous equation:. De = 1.25 x 1.6 0, = 2 D, (10) (This implies'that 8/9 of the smoke produced in a day is removed by fresh air circulation. The ' current air-conditioning standard does not seem to exceed'this, and'therefore the value of a usedihere seems'real'istiic and conservative.) Thus, vhe value of the constant c is 2, and the ' exposure 0.46 cS becomes 0.92 S. The nonsmokers in the indoor space are receiving an exposure in one day equivalent to smoking 0.92 cigarette, which is quite considerable. The current air- conditioning standard, set to remove body odor, is not sufficient to eliminate the effect of ambient smoke. (In fact, one engineering!speci~ficition calls for an air intake of 301ft3'per minute per smoker whereas it requires only 5 to 10 ft3 per minute per nonsmoker.) To reduce the ambient smoke density further by n times calls for an increase of the rate of fresh air circulation!~by nitimes, which would'in turn increase the heating and cooling costs by about n times (disregarding the loss of heating and cool'ing energy due to heat conduction through j. the watls).''-~ In the`al~ove calculation, artilcicial air circulat'ion the year round is assumed,. Actually, many older buildings`•can have windows wide open i'n spring and autumn, and even in summer if not air-conditioned, so better smoke removal can be expected. On the other hand, peoplie living in centrally heated'andicooled apartments are exposed to ambient smoke up to 24 hours a day instead of 11 hours assumed in this model. The effects of these variations tend to cancel out themselves, and the result basedion this model may represent a suitable average. The accuracy of the result obtained (the daily exposure of a nonsmoker equivalent to smoking,0.92'cigarette) depends on the accuracy of many data use6and many averages employed, most of which are reasonably certain. The crucial factor is the rate of air circulation, to which has been devoted considerable effort in reaching,a sui,table average. The final result is expected to be accurate to within 50'0. The order of magnitude of the effect is quite certain. Hiqh-Density Exposure 03735110 Next is calcullated the effect of one hour of hilgh-density exposure, using another mode. Consider a contawting bus fulil of passengers, 37"` of them smoking 1.6 cigarettes each in one hour. One can calculate the exposure of a nonsmoker to ambient smoke in the bus during this hour. In thiis case, the smoke density is much higher than in, the previous case. The air volume assigned to each passenger in the bus is about 2 ft x 2 ft x 7 ft or 28 ft3. Thus, the density Do is increased to Do by a factor of 800/28 = 29 times. On the other hand, the time of exposure is reduced from 11 hours to one hour compared with the previous case; the reduction ratio is 1/11 = 0.091. The two factors partiially compensate each other with a net increase by a factor of 28.6 x 0.091 = 2.6. Thiis means that if the air circulation rate is the same, r

one hour of high-density exposure is equivalent to 2.6 times the low-density exposure of 11 hours. However, the air circulation rate in the bus ils much higher than in the previous work- ing space because the bus door opens and closes often (the same is true for doors of : restaurants and meeting rooms)1. The smoke density decay constant is increased fromia' to a', but the value of a' is difficult to estimate. One can make the liberal assumption of a fresh air turnover rate of four times as high, i.e., a' = 4X. In spite of the uncertainty of this estimate (this is the only quantity in the whole caliculation of R that is uncertain), the error ind'uced'in the final resullt is limited, because this uncertainty'enters the final result through an additive term, rather than through a mulitiplicative factor. The resulit of the exposure is thus reduced by a factor of 4, leading to the conclusionithat one hour of high- density exposure is equivalent to 2.6/4 = 0.7 times that of the low-density exposure of 11 hours, or equivalent to smokeing 0.64 cigarette a day. (Incidentally, it is assumed here that the bus is in continuous operation the whole day, so that the smoke density has reached' equilibirum. Therefore, the previous calculation may be applied.) The total daily exposure of a nonsmoker to ambient smoke from fellow workers an6fellow bus riders is thus equivalent to smokiing 0.92 + 0.64 = 1.56 ciqarettes a day, about 1/13 ,• of that of a smoker. Since a smoker is also exposed to ambient smoke, the totali effect on a smoker is..1/1_ 3 higher than that of the primary smoke. • : . ,, , The Maximum Heath Hazard The ratio, R, of the dose of exposure of a nonsmoker to ambient smoke to that of a smoker to primary smoke is thus estimate6 to be l/13. According to the assumption of linear response, one concludes that the maximum risk due to ambient smoke to nonsmokers is 1/13'of the riisk due to primary smoke toismokers. The latter quantity will now be considered. _ f .c . The effect of smoking in the U.S. has beeni stated as an excess of deaths of 350,000 a year or a reduction of life expectation of eight years to, smokers. These figures are derived by comparing the death rates of smokers and nonsmokers, and the difference is assumed to be due only to the effect of primary smoke to smokers, i.e., the effect of ambient smoke is the same to smokers and nonsmokers and drops out in the comparison. These values thus represent the desired reference standard in the above comparison. Thus, the maximum risk of ambient smoke to nonsmokers in the U.S. is determined to be an excess of deaths of 50,000 a year (based on a smoker-nonsmoker ratio of 37:63) or a reduction of 1ife expectancy of about 225 days. In passing, a similar calculation shows that the maximum estimate of the effect of ambient smoke in just one hour of carpool commutiing daily is a reduction of life.expectancy of about 210'days because of the dense packing,in the car. Since smokers are themselves exposed to ambient smoke, the total reduction of life expectancy of.smokers, according to the assumotion of linear response, is about 8.6 years. These figures of the maximum risk estimate are resonably certaini iin their order of magni,tude and may be used for maximum safety planninas as the prevailing consumerism demands. What is uncertain is the nonlinear reduction factor n needed'to arrive at a reallistic estimate of the risk. The State-of-the-Art Realii'stic Estimate of the Risk There is little information on the nonl'inear effect of low-intensity exposures. However, the results reported by Hirayama (1981) may be used to provide a rough, estimate. According to Hirayama's figures, the rate of death from lung cancer due to primary smoke exclusively is 2.54 times that due to ambient. smoke exclusively (the contribution from! non-smoke-related lung cancer is eliminated in the calculation). The figure is much lower than the value 13 arrived at in the maximum risk estimate. Any nonliinear effect would make the figure greated than 13'. The fact that the Japanese figure is less than 13 must mean that the dose of exposure in the Japanese case, which is between husband and wife, is greater than iin the present case, which is in public places, this being quite reasonable, and that the assumption of 1!inear response is nott far from the truthias far as lung cancer is concerned. The value of n iis about unity for lung cancer. 03735111 However, Hirayama also reported that 'in his study the husbands' smoking didi not iincrease the wives' death rates of stomach cancer, cervical cancer, and ischaemic heart disease significantly. This suggests that there may indeed'be a nonlinear effect for these cases and i

( C the risk is reduced consid'erably. Fo r emphysema and asthma, Hirayama's figures show a dose- response correlation, but the result is not statistically significant. Based on these findings, one can make the crude estimate that n = 1 for lung cancer and emphysema and n = 0 for other cancers and heart disease, although this may be underestimating the effect of heart disease in the U.S. From 1940 to 1962 the death rate from coronary artery disease increased roughly in proportioni to the increase in per capita cigarette consumption, 118% (Federal Advisory Committee Report 1I964). The increase in deathirate of smokers due to heart disease cannot account for the total increase. Therefore, there could be an increase in death rate from heart disease for nonsmokers, possibly in response to the l18°o increase of the ambient smoke density. The Japanese-American difference may be due to the diet--rice and fish in Japan, and beef and milk in America--making the death rate of heart disease in the U.S. ten times as high as in Japan. When so primed by a hilgh-cholesterol diet, ambient smoke might have an effect. :. Among the excess of deaths due to smoking, 16% are from lung cancer and 40116 from emphysema; heart disease, which represents the majority of cases, and'i other cancers account for 80% (Federal Advisory Committee Report 1964). The average value of n is 0.2, meaning that the non- linear effect cuts downi the hazard of ambient smoke by a factor of 5 in comparison withi that of primary smoke. Using this average value of n, one can obtaiin a state-of-the-art estimate of the hazard of ambient smoke in the U.S. as 1/60 of that due to primary smoke. This corresponds to an excess of deaths of 10,000 a year or a reduction of life expectancy of 48'days. As already stated, this may be a minimum estimate. These figures are less certain because of the uncertainty of n but may be more realistic in their interpretation. The maximum risk due to the effect of ambient smoke in the U.S. based on the assumption of linear response is 1/13 of that due to primary smoke, implyinq an excess of deaths of 50,000 a year, or a reduction of life.expectancy of about 225 days. A state-of-the-art realistic estimate of the risk, which is 1!ikely to be a miniwum estimate, is 1/60, of the same, implying an excess of deaths of 10,000 a year or a reduction of liife expectancy of some 48 days. Whichever set of figures iis used, the effect of ambient smoke to nonsmokers is evidently appreciable and makes it a serious environmental hazard. The figures may be compared'~withi those of other hazards of life (Cohen 1974). The reduction of life expectancy due to air polllution is variously estimated to be from 8 to 160 days; automobile accidents, which account for one-half of all accid'entali deaths, 140 days; ozone depletion in the stratosphere due to the use of aerosole spray cans, 1 .1 days; routine emission of radiloactive materials from nuclear power plants in an all!-nuclear ellectric industry, one hour; average effect of all nuclear power plant accidents iln such an economy, one hour. REFEREMCES. Cohen, B.L. 1974. Nuclear science and society. New York: Doubleday. Federal Advisory Committee Report. 1964. Smoking and health, Public Health Service. Hirayama, T. 1981. Brit. Med. J1., Vol. 282, p. 183. Hutchinson, F.W. 1966. "Air conditioning," In McGraw-Hill Encycl'opediia of Science and Technology, Vol. 1, p. 1451. New York: McGraw-Hill. 03735112 Smokin and health: A re ort of the Surgeon General. 1979. Washington, D.C.: Government Printinq Office DHEW'publi,cation no. 79-50066). White, J.R. and Froeb, H.F. 1980. New En~cland J. Med., Vol. 32, p. 720.