Document by0JEg9j0EN0Y6nN6wzY97BKD

I i OUT OF NOTHING Legal limits on asbestos exposure are becoming By Jeffrey A. Boggs increasingly less. I I H t is critical that decisions to re- But how are valid air samplings open removal sites to the un- protected public are supported by statistically valid test data. made when Today a method for estimating lowlevel airborne asbestos fiber concen the values trations with statistical confidence, both prior to and following abatement measured are near zero? projects, has been developed. This "Poisson" distribution method has been tested and found the best statis tical model for determining both unexposed membrane filters (back ground levels) and sample membrane r filters (exposed levels) of asbestos. The "limits of reliable detection" for the USPHS/NIOSH methods P&CAM 239 or 7400 are the limit of qualitative detection (LOD) and the limit ofquantitative detection (LOQ). Standard calculating procedures utilizing known variables--such as microscope field area, filter effective collecting area and sample air volume--provide guidelines for deter mining detection limits. The protocol has been simplified for laboratory analysis of low-level airborne fiber concentrations by the presently preferred NIOSH methods. Additional formulae, which determine the re quired air volume (Q) to "reliably detect" the filter surfece fiber concen tration standards (S) help laboratories and hygienists specify appropriate sampling strategies, a priori. Thus, the proposed standard methodology offers two significant contributions for determining low- level airborne asbestos fiber concen- PLAINTIFF'S EXHIBIT AL-986 MAY/JUNE 1988 ASBESTOS ABATEMENT 23 trations: 1) provides point estimates based on the calculated, known "limit of detection" for any volume of air sampled, and 2) permits statistically valid "re-entry" decisions. OSHA first regulated occupational exposures to asbestos (See Figure 2) in 1971 when it adopted a 12 f/cc limit. In June 1972, OSHA pro mulgated the present standard which is found at 29 CFR 1910.1001. This standard established an 8-hour TWA permissible exposure limit (PEL) of 5 f/cc (with a ceiling limit of 10 f/cc), and mandated further reduction in the PEL to 2 /cc effective July, 1976. In 1975, OSHA again proposed to lower the PEL from 2.0 f/cc to 0.5 free and the ceiling limit from 10 f/cc to 5 free. This notice of proposed rulemaking was based on then-available evidence which OSHA believed had "accumulated to warrant the des ignation of asbestos as a human carcinogen." In 1982, the State of Maryland Figure 1 REAL-TIME ASBESTOS FAM-1 Laser Fiber Counter No special training required High correlation to PCM Immediate notification of methods potentially hazardous Sensitive to .0001 fibers/cc exposures Completely portable Digital readout in fibers/cc Selectable sample/result Differentiates between fibrous and nonfibrous particles times as short as one minute Optional battery pack The FAM-1 can be used in schools, factories, and other public buildings to determine airborne asbestos concentra tions prior tg, .during, and after removal operations. For complete information, call or write to: cMIE MIE, Inc.* 213 Burlington Road Bedford. Massachusetts 01730 (617) 27' 5444 Telex: 92-3339 (Formerly a division of GCA) 24 ASBESTOS ABATEMENT MAY/JUNE 1988 promulgated asbestos regulations that required final air testing on any large project. Areas could not be opened to the public until airborne concentra tions were less than 0.1 f/cc. In November 1983, OSHA published an Emergency Temporary Standard (ETS) which mandated an immediate reduction in the PEL to 0.5 free. In June 1985, the^SPA published "Guidance For ControHIhg AsbestosContaining Materials in Buildings". This text is generally regarded as the latest information on asbestos hazard evaluation. Appendix M of the docu ment recommends that if any phasecontrast microscope samples exceed 0.01 free, the abatement site should be recleaned. The Veterans Administration lias specified that authorization for removal of barriers of small area asbestos removals will be authorized if the air sample is 0.005 free. These current re-occupancy stan dards would suggest very low and zero fiber counts in post-abatement environments. The problem, there fore, is how to handle "zero" or "near zero" data values. In industrial hygiene work, zero value levels are generally undetectable. These levels do occur, however, and (here is no the 1500 liter specification where 10 fibers must be counted in 100 fields. Our result, however, permits the "detection limit'' of 0.01 f/cc to be used when 1112 liters have been sampled and fewer than eight fibers are counted in 100 fields. Conclusions We have shown how to compute the required volume Q to detect a given airborne fiber concentration. However, analytical laboratories do not usually control the volume sampled for individual analysis. As such, we have algebraically solved for the detection limit based on the volume of air sampled given as a con stant K. This computation uses the same features as the relationship for Q, i.e., the effective filter area times the LOQ, surface fiber concentration as known. However, this method solves for the concentration of air borne fibers/cc as a function of the volume sampled. Here, we show for our own laboratory, that K equals 11.12 MAGNITUDE OF? ERRORS ENCOUNTEjffci WITH LOW FIBER CONCENTRATIONS:3 - *- CONDITIONS: 37M& DIA. FILTER T .006"SQ MM RELD-AREA^rs 960 LITERS 2 FIBERS IN 100 FIELDS r BY CAiicuuvnoiip . vXSs. FIBERS/CG = %J|Q0 RBERS) (855 V (.006'SOi MM) (960- UTERS)KiqQ^g^^|| f = .0h0h3a ':iLr* r -Ctr v--.... AX 3 to 25 LPM With adjustable flowmeter removable telescoping arm compact and rugged AISO 25mm and 57mm filter cassettes Model vm-s FREE call or write UT-E^TM Sfap/ex Air Sampler Division 779 Fifth Avenue Brooklyn, n.y. 11252-1695 TOllfree 800-221-0822 (718)768-5553 fibers/cc divided by the dimension less value of the volume sampled. Somewhat obviously, if we divided K by the required volume Q to detect the previously shown re-occupancy limit, we solve for 0.01 fibers/cc. Once analytical laboratories have derived a constant LOQ, they can derive the analytical detection limit based on the volume of air sampled when the fiber count LOQ is not ex ceeded. The results shown here (See Thble 6) for volumes varying from 100 to 5000 liters give corresponding detection limits of 0.12 to 0.002 fibers/cc. We show here (See Thble 7) in the two computations the type of error that results from calculating the air borne asbestos fiber concentration when the LOQ fiber count is not known. This is not a worst-case situa tion but rather, unfortunately, a con dition that arises quite often for samples submitted for re-occupancy. In the upper calculation there were two fibers in 100 fields measured, a value even lower than the LOD, for a sample volume of 960 liters. This is essentially an 8-hour TWA at 2 liters/min. Were the laboratory to release the result of 0.003 fibers/cc, the area would be improperly cleared for re-entry of the unprotected public when the limit of detection is actually 0.012 fibers/cc. This, again, is a moderate problem when compared to a fiber count of zero whose detection limit is still 0.012 fibers/cc--but without this LOQ relationship the analytical laboratory has no quantita tive estimate for the airborne fiber concentration. The value is not truly zero but rather the concentration was undetectable. (About the Author: Jeffrey Boggs is on analytical consultant with Aerosol Monitoring & Analysis, Inc., Hunt Valley, Md. Also contributing to this article were M.J. Cirri and B.E. Lippy of the same laboratory, and N.A. Leidel of the federal Centerfor Disease Control, Atlanta, Ga. 38 ASBESTOS ABATEMENT MAY/JUNE 1988 Asbestos Abatement The Solution EUROCLEAN has developed a unique, patented system for the safe, efficient removal of asbestos and the safe, efficient clean-up of asbestos removal equipment. The main component of the system is a combination cyclone and cleaning filter. The cyclone separates out the heavier dust particles into a replaceable, removable plastic bag. The FlipStrip Filter The FlipStrip filter, a five stage cleaning filter, is an easily removable cartridge that traps the finer dust particles. Air enters through the core of the tubular FlipStrip filter and sets hanging strips in motion that agglomerate the microscopic dust particles into larger particles that work down to the bottom of the cartridge. After four more stages of filtration, the air leaving the filter is virtually clean -- the outside of the filter remains clean at all times and the filter can be handled safely. The five stage filtration process of the FlipStrip filter also enhances filter capacity and efficiency. Plastic Bag Intended and designed for heavy industrial use, the EUROCLEAN system easily outperforms statutory requirements and meets maximum permissible hygienic levels. Since the FlipStrip filter and the cyclone are internal components, they remain completely clean and dust-free on the outside. This means that the FlipStrip filter and the plastic bag can be changed at any time without risk of contamination and without the need to wear protective garments. In addition, the FlipStrip filter outlasts conventional micro filters by a factor of ten times. 907 West Irving Park Road, Itasca, Illinois 60143 800-323-3553 (In IL 312-773-2111) for a .006 mm2 field area. The value of S for Method 7400, where the microscope field area equals 0.00785 mm8, was computed to be 12 fibers/mm2. Having computed the values for LOD and LOQ concentrations, look again at the cumulative observed and expected frequency distribution for re-occupancy sample fiber surface concentrations (See Figure 5). The vertical arrows indicate the LOD and LOQ fiber concentrations of 6 and 13 fibers/mm2, respectively. As can be seen, we have observed sample con centrations less than or equal to LOD at a frequency of 45 percent and less than or equal to LOQ at at frequency of 72 percent. It appears, therefore, we have chosen the appropriate dis tributional model and confidence limits under the condition of low-level airborne concentrations. The relationship for the required air volume Q, in liters, to detect a given concentration level, such as a re rjcv* [TIFIABLE -i)ETECTI0^JtL;03ai) '/Si- ' KwT:^;%V: 5&V - .. sirs) '||f" ''' - ` *"*- . - s-Vl-v. ,.0.11 .T(S$, Taife 6 occupancy standard, is computed by the effective collection area of the filter used, times the fiber surface concentration LOQ equals S, divided by a given concentration we wish to detect (See Table 5). In this example, we are using P&CAM 239, i.e., a 37mm-diameter filter and S equals 13 fibers/mm2, where we must be able to detect the typical re-occupancy limit of 0.01 fibers/cc. The computed required voume of air, Q, is 1112 liters. Those familiar with the EPA guid ance for the limit of reliable quantifi cation of 0.01 fibers/cc using P&CAM 239 method will recognize that this Q of 1112 liters is somewhat lower than GET SPECIALIZED & PERSONALIZED ASBESTOS ABATEMENT We'll help you with... The Availability of Marvels The Availability of Proper Limits Naming Additional Insureds Finding Premiums and Deductibles that Make Sense Covering Second and Third Parly Employees We'll help you understand... fteirocoverages Tailcoverages Reporting Periods The Difference Between an Incident and a Claim Those Stipulations Which Can Void Coverage Al Rose-Tillmann we are experienced specialists in Asbestos Abatement General Liability Coverage. Your insurance program will make sense and provide Ihe proper protection you deserve. We also become involved in your business and deal with your particular requirements. Our product is nol only insurance but also profes sional counseling and service lo you General Liability Coverage If either saving money or true transfer of risks is important to you. ptease contact S. Anthony Brown at (312)480-3570 Rose-Tillmann, Inc. 630 Dundee Road, Suite 200 Northbrook, IL 60062 36 ASBESTOS ABATEMENT MAY/JUNE 1988  to LOD (N) fibers in 100 fields, where the true average count (n) is then the LOQ fibers in 100 fields. The 95 per cent probability is given by one minus the sum of all event probabilities from 0 to (N-l) events. Unfortunately, the procedure to solve for the LOQ (n) is a "trial and error" series of computations. A less burdensome evaluation can be made from the statistical tables of relative expected frequencies of the Poisson distribution. By successively sum ming the expected frequencies given in these tables for several LOQ (n) values by increasing LOQ (n) plus 2, plus 3, etc., one can quickly home in on the N-l value approximately equal to .05 the 95. percent prob ability. In our example, the LOD-N-4 fibers in 100 fields, the LOQ (n) was found to be 7.8 fibers in 100 fibers. Thus there is a 95 percent probabil ity of counting less than or equal to LOD-4 fibers in a single 100-field count, where the true average count of LOQ is 7.8 fibers in 100 fields. Now that we have solved for LOQ fibers in 100 fields at the 95 percent detection power, we can use the rela tionship shown to compute the average surface fiber concentration, S, that would yield a count of 7.8 fibers in 100 microscope fields (See Thble 4). Here, S is in fibers/mm2 and the count field area is in mm2. In this example, using the LOQ equals 7.8 fibers in 100 fields and the microscope field area of .006 mm2, we solve for S equals 13 fibers/mm2. This is now our value for the limit of quan titative detection, LOQ. It is necessary to solve for the LOQ fibers in 100 fields from the LOD fiber count which is a function of the microscope field area. The importance of using the LOD fiber count for an individual micro scope field area is inherent to the logarithmic nature of the cumulative expected frequencies of the Poisson distribution. For example, we have computed the LOD fiber count based on 6 fibers/mm2, the 95 percent detection power, LOQ: and S for com monly used microscope field areas. In the case for P&CAM 239 method us ing a Porton Reticle with a 0.003 mm* field area, S was computed to be 16 fibers/mm2 not twice the value of S - T V'." y t/. ^RELATIONSHIP FOR REQUIRED|^~ 0 AIR VOLUME; Qi ^ -'Wfc. . :.& ' ^.: < ......... ............ -'jSli (FILTER AREA) (S) .-W# DETECTABLEjAI RBORN E ASBESTOS CONCENTRATIO N EXAMPLE: SQ or- - (1V3- F,i-BE-->R. - S^ .-3*<: '*r- . - -s* ,,v ;010 F----I--B----E----R-----S----/--C- C- ^^* ., v !:VU ^ > ,'-r' Tables "THIS 'N^^EOUmEMENTS FIRE-PROOFING ASBESTOS ABATEMENT ACOUSTIC MATERIALS STUCCO POOL PLASTER INSULATING PLASTERS CQ? CEILING TEXTURES SELF-LEVELING FLOOR MATERIALS The economical HY-FLEX 30 SLURRY PUMP uses a quiet electric motor, and rolls through 32" openings on pneumatic tires into your work site for easy application. Remote ON-Ofl control at 24" wand, forward/reverse controls, variable speed, a 30-gallon hopper designed for tree material flow, fast clean-up and no oil seals or bearings in the mix are among other features. A 6 cu. ft. HY-B0Y MIXER is also available to empty directly into the Hy-Flex 30 Pump. Small in size but big on performance: visit our test area, or call for more information about our full line of pumps, blenders, and mixers performance proven to meet your special needs. Cambridge City, IN 47327 R.R.1, Wilbur Wright Road, 317/332-2924 MAY/JUNE 1988 ASBESTOS ABATEMENT 35 BEST TECHNOLOGY AVAILABLE 1 FOR ASBESTOS ABATEMENTS a ENVIROMAJOR A lightweight, self contained 3cubicle airlock entry/exit unit with decontamination tacilities. A ENVIROMASTER A compact unit offering similar facilities to the Enviromajor. Easily dismantled for transportation then assembled on site in 6 minutes. VENTMASTER AIRMOVER Compact 2-part fan/fiiter system which creates negative pressure and captures airborne contamination. Suitable for airflow requirements up to 1000 CFM. Roving prefilter available for maximum versatility. A CUBEMASTER A latch-on unit for either end of Enviromaster to expand changing area. Locker facilities can be included. A SMOKE GENERATOR Indicates airflow conditions within the work area and provides a simple visual check to prove effective tent sealing. REGENCY ENVIROMASTER INC. P.O. Box 3541, West Chester, Pennsylvania 19382. Telephone: (215) 344 0637. Toll Free 1-800 USA ENVI PRESSURE MONITORING Monitor-- displays negative pressure with safe/unsafe Indication and alarm. Printer-- provides continuouslrace of monitored pressure. Power Switching Unit... automatically switches on standby fan/filter unit If pressure becomes unsafe. qualitative detection, LOD. From our data set (See Thble 1), in fact, the 95-99 percentile occured at the sur face concentration of 6 fibers/mm*. Such clumping is a result of the limited classes available from the counting of either whole or half fibers only. The comparison of our data set with the results obtained by the NIOSH Division of Physical Science and Engineering shows good agree ment. There is no significant dif ference between the means of the two data sets. However, the proposed LOD of 6 fibers/mm2 is based on the tested Poisson model. The normally distributed mean plus three standard deviations (or 99th percentile) has been given by NIOSH as the esti mated LOD for Method 7400. The fiber count LOD in fiber/100 fields (See Thble 2) is a function of the microscope field area for counting (mm2) and the 95th fractile value for the distribution of background fiber surface concentrations for unexposed filters, or the LOD in fibers/mm*. For example, for our laboratory us ing a microscope with a count field area of 0.006 mm2, the fiber count LOD is four fibers in 100 fields. There are two conclusions which can result from the comparison of the fiber count LOD to a measured fiber count that exceeds the fiber count LOD: With regard to quality control, any filter lot from which representative filters have a measured fiber count greater than four fibers in 100 fields (LOD) should be discarded. Also, four fibers in a 100-field count is the minimum total count that must be ob tained from any exposed filter sam ple before concluding the exposed Table 4 ProCounsel ASBESTOS ABATEMENT PERSONNEL PLACEMENT NATIONWIDE GENERAL SUPERINTENDENTS PROJECT MANAGERS WORKING FOREMEN GENERAL MANAGERS SAMPLING TECHNICIANS COORDINATORS SALES MANAGERS ESTIMATORS SALES HYGIENIST Placements are made with CONTRACTORS. CONSULTANTS, MANUFACTURERS, and DISTRIBUTORS. Fees paid by employar. All Information kept In strictest confidence. We respect your needs for the right geographical location, the right compensation and the right challenge. Contact SHANNON SMITH Toll Free 1-800-545-5900 or 214-9390635 (24 hour ans. srvc.) 32 ASBESTOS ABATEMENT MAY/JUNE 1988 filter has a significantly greater sur face concentration of fibers than an unexposed filter. (This may not always be the case as shown earlier.) We now have the tools necessary for estimating a sufficient surface fiber concentration to yield a specified power of counting at least the fiber count LOD of fibers in 100 fields, for a given set of measurement condi tions (See Table 3). First, we assume a distribution of many 100-field fiber counts is adequately described by the Poisson distribution. Next, we can solve for the 95th cumulative proba bility on counting less than or equal ... HEPA-AIRE 2000 ASBESTOS FILTRATION MACHINE D.O.P. LEAKAGE (penetration at 0.5 JU D.O.P.) CFM OUTPUT Bp hepa-aire 2000- tested over 30 times better than the competition* S' HEPA-AIRE 2000- tested over 20% better than the competition* * Results of H2000A compared to average results of 5 competitive brands. HEPA-AIRE 2000... blows the doors off the competition We knew our engineers had developed an out standing product, and now the experts agree. In recent tests conducted by professional Indepen dent testing services the new HEPA-AIRE 2000 Model H2000A was the hands down winner when evaluated against five leading competitive brands. In fact, it wasn't even close. HEPA-AIRE 2000 blew the doors off the competition! For example, HEPA-AIRE 2000 tested up to 60 times better against D.O.P. smoke penetration. So, why buy the rest when you can have the best? Call or write today for more Information about the HEPA-AIRE 2000 or for the name of your local Abatement Technologies distributor. HEPA-AIRE 2000-the state-of-the-art in asbestos air filtration products. Now available in 3 models... H2000A-2 speed full feature electric H2000C-2 speed contractors electric H2000P-variable speed air driven model ADVANCED PRODUCTS FOR THE ASBESTOS REMOVAL INDUSTRY When you go info bidding with o 60% reduction in labor and material costs, you've got one big competitive edge! And an opportunity for greater profits. The HEPAVAC automated vacuum system reduces bag handling from 5 or more workers to 2. Cuts bag use by 75%. Why? Because wet or dry asbestos is moved, not by muscle, but by a vacuum hose up to 1000 ft. long at the rate of 9-18 tons/hr. Certified HEPA air filtration assures system will meet or exceed OSHA and EPA safety regs. From pick up to bagging, fibers are captured in a fully enclosed, negative pressure system. The HEPAVAC never leaves the site, working continu ously until the job is done. And John Deere diesel engine, quality construction and proven design means the job gets done faster and more safely than the oldfashioned manual way. A faster job. A safer job. A better job. A less expensive job. Bring on the competition! Engineering \SXSSSkCorpora tion 3374West Hopkins Street Milwaukee, Wisconsin 53216 414/444-4010 portion of the curve--comparable to the previously shown background fiber concentration distribution--sep arately from the higher concentration portion from 9 to 23 fibers/mm* For the purpose of Chi-Square testing for fit of the two curves we therefore chose to have just one degree of freedom. Under these testing conditions the Chi-Square measured value was less than the critical value at 50 percent. Thus we now have two separate tests of significant agreement with the Poisson distributional model using both unexposed and exposed filter fiber concentrations. These results provide the proper statistical frame work from which we can evaluate the limits of qualitative detection (LOD) as well as the limits of quantitative detection (LOQ). Limits of Detection We have chosen the 95th percen tile of the observed, Poisson distributed, unexposed filter fiber concentration for the limit of ^-LIMITS fo^FlQUAlllT^IVEDETECTION(l^orD: ""-r ',w,*^^E^psEblifirER GONcbrn^TipN^'' .MM) ........... 3S.D. JOISSONi 7.0* . i'u'Sx Table 2 kwiONsii^|Fci^:-&$ UNTL.O.D rVirisxv AREA)(FlBERS/SQ^MM)i^i ^ 6 FBERS \ SQ.MU-' \ [l0bMELDS#^lJr abase-**. ASBESTOS LIABILITY INSURANCE $1,000,000 Limit Occurrence Non-Cancelable Oomestic Carrier "A" Rated A Not-For-Profit Association X^ 1. Gross Asbestos Removal 2. % Liab. Ins. 3. $lns. Cost Average Actual 1987 A.A.A. Member $1,460,000 2.9% $42,000 Typical Comparable Competitor $1,460,000 13% $190,000 SAVINGS =$148,000 How Much Could You Save 1,460,000 ?% ?$ =? ? ? Asbestos Abatement Association 101 S. Stratford Rd., Winston-Salem, N.C. 27104 919-722-9895 30 ASBESTOS ABATEMENT MAY/JUNE 1988 A striking feature of the data is the large frequency of zero concentra tions, 33 percent of the total. Addi tionally, the clumping of 0-2 fibers/nun2 results in the observed relative frequency of nearly 60 percent. These results indicate that back ground fibers concentrations might be adequately described by the Poisson distribution. The Poisson distribu tional model also seems appropriate based on the low mean value of 2.3 fibers/mm2 with a high coefficient of variation of 87 percent. The purpose of fitting a Poisson distribution to numbers of rare events is to test whether the events occur independ ently with respect to' each other. If they do, they will follow the Poisson distribution. We have tested this observed dis tribution as well as the observed fre quency distribution of re-occupancy sample fiber concentrations against the expected Poisson frequencies and evaluated the comparison by the ChiSquare test for goodness of fit. The plot of cumulative, relative observed and expected frequencies of fibers concentrations (f/mm2) from unex posed filters shows good agreement throughout the range of measured values (See Figure 3). The shape of the Poisson expected frequency curve is completely defined by the mean of the observed values. The Chi-Square test value was less than the Chi-Square critical value at 10 percent with four degrees of freedom and by interpolation was ac tually less than the critical value at 30 percent. This significant agreement, therefore, supports our hypothesis that the observed fiber concentrations of unexposed filters are Poisson distributed. Having adequately described the Poisson distribution of background fiber concentrations, we have ran domly selected 100 re-occupancy samples and plotted the observed fre quency distribution for fiber concen trations in fiber/mm2 (See Figure 4). Not too surprisingly, .as predicted from these presumed clean air en vironments, there is a relatively high frequency of low, 0-3 fibers/mm2, concentrations accounting for more than 20 percent of the distribution with more than 50 percent being less than 8 fibers/mm2. The observed mean for this data set was 8.8 fibers/mm* with a coefficient of variation of 76 percent. These sam ple concentrations are measured from filtered volumes of 1000-1500 liters and have not been corrected for background fiber concentrations. With such a high coefficient of varia tion and a relatively low total fiber concentration, we felt this observed frequency distribution might also be Poisson distributed. Again, we have plotted (See Figure 5) the observed and expected cumu lative frequencies of the fiber concen trations in fibers/mm2. We wished to evaluate the low-levl concentration MAY/JUNE 1988 ASBESTOS ABATEMENT 29 Figure 2 single accepted way to handle them. One method is to obtain from the analytical laboratory the least amount of the contaminant which the analy tical method can detect. This value is then used to determine the least detectable concentration in the amount of air sampled, and is then substituted for all zero values. By resolving the r roper distributional model, one can hen apply the ap propriate probability function, and thereby the detection power, to the analytical methods. Methods and Results We have accumulated data over the past several years (See Figure 1) for fiber counts on unexposed filters as a routine quality control measure for background fiber surface concentra tions. The observed frequency distri bution of these fiber concentrations, shown as fibers per square millimeter, was obtained from phase-contrast microscope counts using the NIOSH analytical methods P&CAM 239. Some 100 separate factory lot num bers of 37mm-diameTer, celluloseester membrane filters have been analyzed with a microscope optical system using a Patterson-Globe graticule that has a calibrated field area of .0062 square millimeters. Figure 3 28 ASBESTOS ABATEMENT MAY/JUNE 1988 BATEMENT *T* ECHNOtOGIES 3305 Breckinridge Blvd. Suite 118 Duluth, GA 30136 404-925-2761 FAX 1-404-381-7454 AIR FILTRATION WATER FILTRATION GLOVE BAGS ADHESIVES SURFACTANTS SIGNS LABELS BREATHING With the new Duo-FloTM and Duo-TwinTM Pressure Demand Respirators from MSA, you can enter, exit and move station to station in asbestos work areas without an air-line* When you remove asbestos for a living, you know what a hassle it can be to have an air-line hose trail you around. It's there wherever you go, in whatever you do. Well not anymore. MSA's new Duo-Flo and Duo-Twin Pressure Demand Respirators? featuring ' Ultravue* or Duo-Twin Facepieces with unique D/PD "switchable" exhalation valve, allow you to easily disconnect from your air-line and still breathe freely. It's like having more breathing room to work with, without more air-line to worry about. The Duo-Flo and Duo-Twin Pressure Demand Respirators meet EPA guidelines, are fully approved by NIOSH, and are com pletely retrofittable from MSA's current straight Pressure Demand and Duo-Flo Pressure Demand Respirators. For more information on the new Duo-Flo and Duo-Twin Pressure Demand Respirators, or any of MSA's full line of asbestos abatement products, call 1-800MSA-2222. Or write MSA, EO. Box 426, Pittsburgh, PA 15230. J Ultravue Facepiece with D/PD valve. MSA - "This device is NIOSH approved for entry, egress and moving station to station in the air-purifying "HEPA Filter" mode of operation.