Tobacco Documents Online

2241 T. Eieher, F. M1111er, and D. Steinhoff very broad basis to the solution of the problems that still have to be overcome. WYNDER: I understand that, and I compliment Drs. Eicher and Muller on the excellent science. My question is, do we believe that these tobacco substi- tutes will have or already evidence acceptance anywhere in the worfd'? KEITH: tT-H: I think that undoubtedly tobacco substitutes have capabilities of reducing tar, nicotine, and undesirable components, and by a variety of biological tests they can be shown to have utility. When the science and the biological effects can be exposed to the public, such products have received some degree of acceptance. In areas where you can't say anything about the material-onfy that "tobacco substitute" is in the pack-the public just doesn't buy the cigarettes in any appreciable quantities. You have a viable product when you tell the public you are improving the cigarette. BOCK: Was the bioassay (Table 2) expressed in an equal number of cigarettes basis? MULLER: Yes. BoCK: Did you do sufficient dose-response information so that you could predict how it would be on an equivalent amount of condensate basis? MOLLER: For the 30% mixture, that was equal forcondensate, and for the 5017v mixture there was a specific reduction too. BOCK: So the specific activity was lower, as well as the absolute level. ooRl: Artificial substitutes may have some significance in the future, particu- larly, as Dr. Muller pointed out, if cigarettes with high nicotine and low tar become desirable. But first the safety of these products must be proved beyond reasonable doubt, which in my opinion has not yet been done. Second, there is an economic question because in the U.S. we can achieve the same results by filtration and aeration, as some of the speakers have indicated before. Also whether the country where RCN or ATS may be used is tobacco-producing or not influences the acceptance that one gets there. oL19zsEZOz I Physical Methods for the Modification of Tobacco Smoke CHARLES H. KEITH Celanese Fibers Company Charlotte, North Carolina 38232 In the past 30 years, there has been a dramatic change in the cigarettes commercially available in the U.S. and throughout the world. Back in 1950, the usual cigarette was a short. 70-mm, unfiltered cigarette which delivered 30 mg or more of tar and correspondingly large amounts of other smoke compo- nents. Conversely at the present time the usual commercial cigarette is an 85-mm, filtered cigarette with a tar delivery of 17 mg or less. A rapidly growing segment of the market consists of low-tar cigarettes with deliveries ranging from 10 mg down to I mg and greatly reduced deliveries ot' other smoke components such as nicotine and carbon monoxide (CO). Although many factors have contributed to this profound reduction in delivery of smoke and its components, by far the most important changes have been the addition of a filter to the cigarette and, more recently, the introduction of controlled air dilution systems in the filter and the tobacco column. These physical modifications of the cigarette have provided most of the 45-95% reduction in smoke and smoke component delivery. The purpose of this discussion is to describe brietly these systems, illus- trate their individual and combined effects on the smoke stream, and show that a combination of' filtration and dilution can be used to achieve a controlled delivery of a number of smoke components. The advantages and limitations of these filtration and dilution techniques will also be discussed, as will future trends in their usage. In gathering the material for this review, a number of previous discussions of filtration have been utifized. In particular the works of Kiefer and Touey (1967). Stiiber (1970), Reynolds (1970). Kiefer (1972), Keith (1972, 1975, 1978) and Williamson (1974) have been useful. Reviews and articfes. by Schur and Rikards (1960), Norman (1974), Moric (1976), Baker and Crellin (1977), Owens wens (1978), Selke (1978), Selke and Matthews (1978), and Durcx:her et al. (1978) have provided much information about the dilution process both in cigarette paper and filter tips. Recent work by Browne et al. (1979) has provided much of the basis for a discussion of the interaction between filtration, dilution, and tobacco combustion. 225

22tl / C.H. Keith FILTRATION Although cigarettes with an attached filter have been commercially available since the mid-1930s, they were not widely accepted until the mid-1950s. Since that time filtered cigarettes have grown from a market share of about 1% to almost 90% in the U.S. and are continuing to displace the classical unfiltered cigarette both here and abroad. This dramatic change in a space of 25 years has occurred primarily because of two important factors. One, of course, has been a growing public awareness of the health aspects of smoking, which has engendered a desire for lighter, milder cigarettes. The second was the intnxluc- tion of the cellulose acetate filter, which can be manufactured and attached to cigarettes at high speed with a high degree of uniformity at reasonable cost. Unlike many other filter systems, the acetate filter does not seriously distort the taste spectrum of tobacco smoke, and it imparts a smoothness by selectively removing some of the irritants in the smoke stream. Both of these factors apparently contributed to its widespread public acceptance. As a result almost all of the commercial cigarette filters in this country are made from cellulose acetate fiber. . The usual cellulose acetate filter is made from a bundle, commonly called a low, of 10,000-15,000 continuous filaments of cellulose diacetale polymer. The fiber dimensions and the number of fibers are generally specified in terms of denier per filament (DPF) and total denier (TD). Denier is a textile term that gives the weight in grams of 9000 m of f7ber, and is directly related to tlTe cross section area of the fiber or group of fibers. For cigarette tow, the range of DPFs is from 1.25 to 12 which represents a span of cross-sectional area of' 100-1000 µm"/fiber. The range of total deniers is 20,000-100,000 but most ordinary filters use tows in the range of 35,000-45,000 TD. The fibers are not cylindrical but have various shapes dictated by the geometry of the orifice from which they are extruded. Most commonly this is trilobal in the shape of the letter Y, but 1-.beam and multilobed fibers are available. Usually a tow will be specified in terms of DPF, cross-sectional shape, and total denier, as for example 3.3Y/40000 or 4.01/42000. In the tow manufacturing process, a crimp is imparted to the fiber and during filter making this is partially removed. Considerable crimp remains in the finished filter in the form of a zig zag or sinusoidal wave with a frequency of 10-20 crimps per inch. The net result is that average fiber orientation is 25-45° away from the longitudinal axis of the filter rod. This crimp provides a degree of firmness to the filter so that a reasonably rigid cylinder is created even though the fibers occupy only about 10% of the filter volume. This firmness is increased by the addition of a partial solvent or plasticizers to the tow during the rod-making process. This serves to weld the fibers together at their points of intersection, thereby increasing the stiffness of the fiber mass. In addition to the fiber properties, the other important filter parameters are the length and diameter, the latter usually being expressed in terms of circum- ferencc of the filter. Circumferences of most normal brands are around 24.8 I Modifying Tobacco Smoke / 227 mm (7.9-mm diameter) but slim brands have lower values of 21.5-22.5 mm (6.8-7.2<mm diameter). Filter lengths cover quite a range, going from 10-15 mm on short 70-mm cigarettes to 20-22 mm on 80-_ and 85-mm cigarettes; 100-mm brands frequently have 25-mm filters. The primary purpose of a cigarette filter is to remove part but not all of the smoke particles. These particles have diameters ranging from 0. I- I µm and are captured by the filter fibers by several mechanisms which are shown schematically in Figure 1. The first and most important mechanism, particu- larly for the smaller particles, is Brownian diffusion. In this filtration process, a particle traveling on a streamline in the vicinity of a fiber is actually following an erratic path because of frequent and irregular collisions with surrounding gas molecules. This path may be such that the particle strikes the surface of the fiber, and is permanently lost to the smoke stream. Since smaller particles have greater displacements because of molecular bombardment, they are more likely to be deposited on the fiber. Thus the filtration efficiency of the filter increases with decreasing particle size.. From theoretical studies of smoke filtration (Keith 1y75), it is found that this mechanism accounts for about two-thirds of the removal efficiency. A second important filtration mechanism is_ direct interception of particles by the fiber. This occurs when a particle is so situated on a streamline that it touches the fiber during its passage around it. It then is permanently removed or filtered from the stream. This mechanism is more effective for large smoke particles, and is found to account for r almost one-third of smoke particle removal efficiency. A final mechanism, illustrated in Figure I, is that of inertial impaction of particles onto the fiber. Even though the particle is rather small, it does have mass and inertia and hence it tends to continue on its original path as the gas stream turns to flow around the fiber. This leads to a deposition of the particle on the fiber surface. This mechanism is also more effective for large particles, but does not appear to be very important in cigarette smoke filtration, amount- ing to about 1% of the overall particle removal efficiency. Other filtration mechanisms such as thermal diffusion, electrostatic attrac- tion, and gravity deposition, are thought to be negligible in tobacco smoke DIRECT INTERC€PTIDN €FFECTIYE FIBER RADIUS i FIBER Figure 1 Schematic diagram of methods of particle capture by filter fibers. z4z9jsEZOz

I 228/ C.H. Keith fiftration because of the small thermal gradients in the cigarette filter, dle uncharged or very lightly charged nature of the smuke, and the small particle size and short residence time of the smoke particles in the filter. A direct sieving action by the tifter also appears to be unlikely as the average interfihcr separation is of the order of 40-50 µm, which far exceeds the less than 1-µm size of'the particles. In any consideration of fiftration, the draw resistance or pressure drop of' the filter has to be considered along with the removal efficiency of the filter. The two are closely related because high-efficiency filters are almost invariably accompanied by high pressure drops. This occurs because those factors such as smaller fiber size, closer fiber packing, and increased filter length that increase filtration also increase the pressure drop. It is very important to control the pressure drop of the filter as this has a direct bearing on the consumer acceptance of the cigarette. Generally cigarettes with pressure drops in excess of 150 mm of water are rejected by smokers because of hard draw. Using basic physical principles, it has been possible to develop theoretical expressions for pressure drop and particle removal for acetate tow filters. As shown in Figure 2, values calculated by these equations agree well with experimentally determined pressure drops and nicotine removal efficiencies. These data cover a series of fifters ranging in length from 17-25 mm and DPFs ranging from 1.8-8.0. A similar correlation exists between particle removal efficiency and tar removal efficiency, although the latter is generally 7-8 units greater than the nicotine removal efficiency. This increase occurs - because tar contains some volatile and semlvolatlfe components such as phenol that arc selectively adsorbed by the filter material, thereby inflating the removal efficiency. To summarize the effects of filtration, Table I briefly lists representative effects of various filter parameters on the pressure drop and particle removal efficiency of' acetate tow filters. In this table it is apparent that increasing filtcr length, decreasing fiber size, and increasing fiber weight, circumference, and degree of dilution all increase the removal efficiency for smoke particles. Pressure drop goes in the same direction with length, fiber denier, and weight, but decreases with increasing circumference and difution, the latter being important in low-tar cigarettes. The 20-60% range of efficiencies shown is essentially the practical range for acetate fifters. Thus it is difficult to reduce tar deliveries much below 13- 15 mg per cigarette by means of conventional filters alone. Other filter systems such as corrugated 6lters, or molded filters with a special configuration can achieve somewhat lower deliveries; however these are less desirable from the point of view of uniformity and economy. DILUTION To extend the range of tar deliveries to lower values and to achieve a reduction in the amounts of gaseous components in cigarette smoke, the physical process I Z4i9zsCzoz ModifVinp Tobacco Smoke / 229 ® 01 0 1:1 LINE 0 IU .11 30 40 50 60 70 80 90 -IUU-IfU l1U M€ASUNED PRESSURE DROP (MM NLO) PANTICLE REMOVAL EFF (j;1ENCY Figure 2 Computed versus measured pressure drops and removal efficiencies of air dilution is usually employed. Air dilution or ventilation can be achieved by a combination of techniques. One approach is to increase the porosity or permeability of the cigarette paper enclosing the tobacco column by using an inherently more porous paper or by mechanically or electrostatically perforating a lower porosity paper. A second method is to introduce a controlled degree of dilution in the filter section of the cigarette by using perforated or porous plug wraps and tipping papers. Since these two methods give somewhat different results in terms of delivery of smoke components, it is desirable to discuss them separately even though they are interrelated. The reduction of delivery by means of porous or perforated cigarette papers was first described by Schur and Rikards (1960) and has been the subject of a number of articles since then. Cigarette paper, which usually ?

„ ;30/ C.H. Keith Table 1 Effect of Filter Variables on Pressure Drop and Filtration Efficiency Variable Amount Pressure drop (mm H,O) Particle removal efficiency (%) Length (mm) 15 42 26 20 57 33 25 71 40 30 85 46 Fiber size (DPF) 1.1 159 55 2.0 87 43 3.1 56 33 5.6 31 20 Fiber weight (g) .101 46 29 .111 56 33 .121 67 37 Circumference (mm) 22.6 63 31 24.6 56 3 3 26.6 50 36 Dilution (%) 0 56 33 25 42 38 50 28 46 75 14 61 consists of cellulosic fibers from flax with added clay and carbonate fillers and small amounts of burning modifiers, is naturally porous to a degree. By changing the physical form of the fibers and fillers and by modification of the paper-making process, this level of porosity can be changed over quite wide limits. For a normal low-porosity paper, air flows of 12 cm'/(cm" • min • cb . pressure drop) are found by the CORESTA (1975) recommended method, which is generally referred to as a permeability of 12 CORESTA units. By suitable modification, inherently porous papers of up to 120 CORESTA units can be fabricated. However increasing the porosity of paper also increases its burning rate, which in turn affects the burning rate of the cigarette, resulting in fewer puffs to reach a given butt length. Alternatively, a cigarette paper may be perforated to increase its porosity. This is usually accomplished by means of an electrical spark gap that burns small holes through the paper. Using this technique, CORESTA permeabilitics of up to 200 units are achievable. Since the perforation process does not change the burning characteristics of the base paper, it is possible by this technique to obtain either slow- or fast-burning cigarettes. As shown by Owens (1978) an appropriate choice of paper and amount and type of perforations can control both the cigarette burning rate and degree of dilution. C4i9isEZOz Modifying Tobacco Smoke /?31 This ability to control both the rate of bum and the degree of'dihuion is imponanl in that it allows greater than expected reductions in the delivery of some smoke components. This is illustrated in data reported by Owens (1978) who showed that increasing dilution by means of inherently porous paper gave reduced puff counts_ and sizeable reductions in tar, nicotine, CO, and NO. Conversely, when perforated paper was used to achieve similar dilutions, ns, increased puff counts were observed along with large reductions in CO and NO. Tar was also reduced, but to a lesser degree than previously, and nicotine delivery was essentially unchanged until high dilutions wece achieved. These data suggest an interaction between dilution and combustion as will be dis- cussed later. The large reductions in the permanent gases, CO and NO, partially result from the diffusion of these materials out through the cigarene paper during the smoking process, as has been found by Owen and Reynolds (1967), Morie (1976), Baker and Crellin (1977), and Durocher et al. (1978). Although air dilution through cigarette paper and the outward d diffusion of gases is universally present in all cigarettes and is useful in controlling the relative amounts of smoke components, it has two drawbacks limiting its usefulness in designing low-delivcry cigarettes. The first drawback is that the dilution and its attendant effects decrease as the cigarette is consumed. Thus the normally higher deliveries of smoke components in the latter pufls are fiirther heightened because of the progressive loss of the initial dilution effect. This provides a feeling of imbalance during the smoking process. The second drawback is that the practical levels of dilution are fairly low. For inherently porous papers of the order of 80- 100 CORESTA porosity, the top level is about 25%. Electrostatically perforated papers generally have dilutions of 35% or less. These limitations are imposed by the mechanical, burning, and varia- bility characteristics of the paper. Also if much higher dilutions were incorpo- rated in the cigarette paper, it would be difficult to light the cigarette because of a very high dilution in the initial puff. Because of these drawbacks to dilution through cigarette paper, much of the activity in recent years has been in the development of filter dilution systems. These have the advantage that a relatively constant dilution can be achieved during the whole smoking cycle, and, by appropriate choice of filter plug wrapping and tipping paper, almost any dilution level can be achieved. This is illustrated by the fact that ventilated brands currently on the market have a range of filter dilutions from a low of 15% to a high of.80%. The structure of ventilated filter lips varies from brand to brand, but, in general, the dilution system involves the passage of air through two layers of paper. The inner layer is the paper wrapper for the filter tow and is called a plug wrap. In unvented cigarettes this is a low-porosity material (less than 10 CORESTA units), but for vented filters a range of porosities from 600 to 26,000 CORESTA units is available. The latter material resembles the paper used on tea bags, and essentially presents no barrier to air flow. A recent development is a process for preparing filter plugs with no paper wrapper, ~.~. ~,.

i 2321 C.H. Keith which are called nonwrapped acetate (NWA) rods. The outer barrier consists of the tipping paper, which attaches the tobacco column to the filter. This is usually vented by a single row or multiple rows of perforations, which are formed in the paper by mechanical punches, laser beams, or electrical sparks. Considerable ingenuity has been shown in designing equipment to fonn these holes with very close tolerances at high speeds. Another type of tipping paper has an inherent porosity built into the paper. This provides a dilution over a considerable area of the filter rather than at discreet, localized holes. In both types of tipping, control of glue application in the plug-making and cigarette- assembly processes is important to avoid closing the ventilation holes. A pioneering investigation of the effects of filter dilution on smoke component delivery was conducted by Norman (1974). Using a vented acetate filter and a companion vented, empty-tube mouthpiece, he compared the re- duction in delivery of 13 components or groups of components with the degree of dilution. If the only effect occurring were a simple replacement of smoke with air, it would be expected that a plot of yield reduction against dilution would be a straight I : I line. In most cases this was not true, indicating an interaction between dilution and other processes. As shown in Table 2, Nor- man found that permanent gases and some volatiles such as CO, CO, NO, hydrogen cyanide (HCN), and a group of aldehydes were all delivered in lesser amounts than would be expected from the dilution. This indicates, as men- tioned earlier, that the outward diffusion of these gases from the tobacco column through the cigarette paper reduces their delivery. As the filter is more highly vented, the rate of passage of the smoke stream through the tobacco column is decreased, thereby allowing a greater diffusional loss. As has been discussed by Durocher et al. (1978), there does not appear to be any appreci- able outward diffusion through the filter vents. In another investigation (Norman 1974), filterable components of low to medium volatility were removed about as expected from the dilution levels or, in some instances, were reduced much less than expected. Wet smoke solids and tar were removed at about the expected amounts in the open tube mouth- piece, but in greater than expected amounts in the vented filters. As was discussed earlier, this is an indication of an increase in the filtration efficiencies for these materials because of a slower flow rate in the tobacco column and a portion of the filter. This effect is heightened for particulate water, which is efficiently filtered both by tobacco and acetate filters. Nicotine, menthol, waxes, and volatile phenols gave less than expected yield reductions. Most of these materials are considered to be distillable smoke components that are directly transferred from the tobacco to the smoke stream. Evidently the slower combustion occurring in the more highly ventilated cigarettes allows a more effective transfer of these materials backwards through the tobacco column with less destructive loss. In the case of nicotine, this enrichment is offset by a more efficient filtration, again resulting from the lower stream velocity. But for phenol, which is largely adsorbed from the vapor phase, the extra dilution Modifying Tobacco Smoke / 233 Table 2 Yield Reduction by Filter Dilution Material Acetate filter Vented mouthpiece 1'crnment gases CO + + CO, + t NO + + Volatiles HCN + + Isoprene + + Total aldehydes + + Distillables Nicotine 0 '' Menthol Phenols _ I 0 WaJ(es Other Wet stnoke solids .} (I Particulate water + Tar + 0 Data from Normaa (1974). •Greater yield reduction than dilution level. "Yield reduction equals dilution level. ' Yield reduction less than dilution level. reduces its vapor phase concentration and thereby its adsorption and selective filtration on an acetate filter. Although filter ventilation can remove more or less of various components and is largely constant during the smoking process, it does not seem to be highly dependent on the way that the dilution is achieved. McKee and Parker (1978) investigated a wide variety of filter constructions, using various tow, plug wrap, and tipping paper combinations. From their data the delivery of all smoke components appears to mostly depend on the total dilution level, i.e., the sum of the filter and tobacco column dilutions. Additionally, there does seem to be a dependence for filterable materials on the cigarette tow incorpo- rated in the filter. As found both in Norman (1974) and McKee and Parker (1978), filtration efficiency for these materials increases in a nearly quadratic fashion as dilution is increased. This effect has also been described by Keith (1978) and Kiefer (1978). Thus an interaction between dilution and filtration is well established. tLz9zsEZOz

4341 C.H. Keilh One of' the fortunate aspects of dilution is that it lends to decrease the gas concentrations more than expected, whereas filtration is more effcaive in reducing particulate concentrations. Thus these two physical techniques are complementary, and by a proper choice of levels of each, considerable flexibil- ity in cigarette design is available. A further fortunate occurrence is that nicotine, which is associated with the taste and impact and hence acceptability of smoke is less affected by dilution than tar, which is a measure of the total smoke exposure. Because of this it is possible to create low-tar cigarettes that maintain an adequate level of consumer acceptance by using a combination of dilution and tiltration. INTERACTION WITH COMBUSTION In the preceding discussion, the effect of the interaction between dilution and filtration on smoke delivery has been outlined. Dilution unlike filtration has a further effect in that it alters the combustion process, and thereby affects not only mainstream but sidestream smoke. As shown by Browne and colleagues (1979), increasing the level of filter dilution decreases the puff volume applied at the burning end of the cigarette. The effects are qualitatively similar to those obtained by changing the puff applied to the cigarette. Slow gradual puffs of one-half (he standard volume on an unvented cigarette were found to give about the same delivery of tar, nicotine, CO, and CO. as those of standard volume from a 50%-dilution cigarette. This point is illustrated in Table 3 for these components in mainstream particulate and vapor phase smoke. Table 3 Mainstream Smoke Data Including Material Captured on Filters escription Number puffs Puff volume at coal (ml) Tar (mglcigt) Nicotine (mg/cigt) CO (mg/cigt) CO, (mg/cigt) 35-ml f'uff volume No dilution 8.7 35.0 29.4 1.7 18.6 52.2 33% Dilution 8.8 23.5 27.4 1.4 12.9 40.3 48^k Dilution 9.8 18.2 15.7 1.1 6.6 27.4 83% Dilution 10.6 6.0 9.0 0.6 2.4 14.7 17.5-m1 1'uff velume No dilution . 9.6 17.5 21.5 1.2 9.3 32.4 3396 Dilutiort 10.3 11.7 13.0 0.9 5.0 22.2 50,mi Puff yolbme No dilution 7.4 50.0 34.3 2.1 20.4 56.9 33% Dilution 8.3 33.5 29.8 1.8 17.0 51.3 sLi9zsCzoz .,....~_. _ Modifying Tobacco Smoke / 235 The interaction between dilution and combustion is clearly apparent when the deliveries of components in the mainstream and sidestream are combined to provide a total delivery. As shown in Figure 3, the total delivery as a function of dilution differs for various materials. As dilution is increased distillable components such as nicotine and water have relatively constant total deliveries. Both are reduced in the mainstream at higher dilutions, but these reductions are offset by increases in sidestrcam delivery (considerably more of both compo- nents is found in the sidestream). The total delivery of tar that can be consid- ered as an incomplete combustion product is_ strongly reduced as dilution is increased. On the other hand, CO., which is a final product of combustion, ustion, is strongly increased. An intermediate combustion product, CO, also decreases slightly as dilution increases, which indicates that air-diluted cigarettes do not, as is fiequently supposed, provide more of an environmental burden than their undiluted counterparts. The pattern shown here indicates that as dilution is ror tu4. N1[OTINE 13 0 l I 1 I a I I 1 I I j lo 70 30- 40 so 60 )o to 90 100 0 DILU11014 (S) Figure 3 Tqrl dalivery as a funcuon of dilution l

$384 C.H. Keith increased there is a more complete combustion of the tobacco approaching that occurring in a free-buming cigarette where all materials are going into the sidestream. CONCLUSION To summarize, it is found that filtration and dilution are powerful tools for the reduction of smoke component delivery. Filters that operate principally on the particulate phase of tobacco smoke can remove one-ha)f or more of the smoke particles presented to them. Dilution, particularly filter dilution operates on both the gaseous and particulate phases. Because of interactions between dilution, fiitration, and combustion, vented cigarettes deliver more or less of many smoke components than might be expected from the dilution levels. By an appropriate combination of these physical techniques, it is possible to design cigarettes with almost any level of delivery of tar, nicotine, and gaseous components. The fact that there are commercial brands available now with extremely low deliveries of these materials indicates that the tobacco industry is responsive to the need for such products. The only real barrier to a wide usage of these products is their acceptance by the smoking public. As for future trends, it does not currently seem as though there will he any major change in the make-up of cigarettes, such as a revolutionary new filter or dilution system. In years to come, more and more cigarettes will be highly filtered and will increasingly incorporate dilution as a means of modification of the smoke stream. Low and very low-tar cigarettes will continue to supplant current brands, probably not at the very rapid pace of the past few years, but before too many years have passed they should represent the major portion of the U.S. cigarette market. REFERENCES Baker, R.R. and R.A. Crellin. 1977. The diffusion of carbon monoxide out of cigarettes. Beirr. TahukJorsch. 9:131. CORESTA. 1975. Recommended Method No. 3. Determination of the air permeability of cigarette paper. CORF.STA Bull. 1975-3-4:40. Browne, C.L., C.H. Keith, and R.E. Allen. 1979. The effect of filter ventilation on the yield and composition of mainstream and sidestream smokes. Beirr. Trehuk- Jnrsch. (in press). Durocher, D.F., C.F. Mauina, and W.A. Selke. 1978. Diffusion of gaseous compo-s nents through the wrapper of a cigarette. Beirr. Tahakf'e)rsch. 9:201. Keith, C.H. 1972. Modification of tobacco smoke. In The chemisrry oJ tobacco and tobacco smoke, p. 149. Plenum Press, New York. . 1975. Experimental and theoretical aspects of smoke filtration. ACS Symp. Ser. 17:79. Modifying Tobacco Smoke /237 _ 1978. I'hysical mcchani.ms of smoke fiftrabon. In Recent udrunrc•s in mhucev, .science, vol. 4, p. 25. 32nd Tobacco Chemists Research Conl'erence, MontrcaL Kicfer, J.E. 1972. Fihration of cigarette smoke. In "!'he rhemi.,nv of te,hu,eu and mhue e u srnpke•, p. 167. Plenum Press, New York._ . 1979. Ventilated filters and their effect.+ on smoke comfNosition. In Rev rnr uel„enres in rnhua u science, vol. 4, p. 69. 32nd Tobacco Chemists Research Conference, Montreal. Kiefer, J.E. and G.P. Touey. 1967. Filtration of tohacco smoke particles. In Tahucro und u.lxeccu smoke, p. 545. Academic Press, New York. - McKee, J.L. and J.A. Parker. 1978. Filter structure effects on tip ventilation and cegarctte smoke yields, pan I. Paper presented at 32nd 2nd Tobacco Chemists Re- search Conference, Montreal. Moric, G.I'. 1976. Some I:ectUrs that affect Ihe diffusion of carbon monoxide out of cigarctle.. 7i,h. Sri. 20:167. Nonnan, V. 1974. The effect of perlorated tipping paper on the yield of various smoke components. Beirr. Tubukfirrsch. 7:2ti2. Owens, W.F. Jr. 197N. Eftect of cigarette paper on sneyike yield and composition. In Re•ceru udrunces in tohuccu science, vol. 4, p. 3. 32nd Tobacco Chemists Research Conlerence, Montreal. Owen, W.C. and M.L. Reynolds. 1967. The diffusion of gases through cigarette paper dunng smoking. 'l'uh. Sri. 11:14. - - Reynolds, M.L. 1970. Theory of smoke filtration. Paper presented at 21st Tobacco Chemists Research Conference, Montreal. Schur, M.O. and 1.C. Rikards. 1960. The design of low yield cigarettes. Toh. Sri. 4:69. Selke, W.A. 1978. Dilution of smoke through ventilation of filters. Beirr. TubukJursch. 9:190. Selke, elke, W.A. and 1.H. Matthews. 1978. The permeability of' cigarette papers and cigarette ventilation. Beirr. Tabukfiu.sch. 9:193. Stiiber, W. I970. Smoke aerosols, formation, change, filtration. In Procerdheg.c 5th hnernuriunul Tuhuccu Scientific Congres.s, p. 79. Verband der Cigarenenindus- try, Hamburg. Williamson, 1.T. 1974. The elTect of filters on cigarette smoke compounds. CORE;STA Bed l. 1975-special:79. 9a3TsEZOz