,varl,:atio-n of Cn:--brid~-e fillers as a rep-qs for the of s7o;-.e durin the e-eaution of anal",%rses on ci.:Tarette s-oke. RN.T Hazourg- Sz/*diek/Xaussmaxm D 15021, 7-o.1-19. Intro,t--~:tion -ti,,hin the scope of revision work- on the standard f.-:! smoking, the q----e3tvion arose as to the (Yhoice of that method of pre--ipitation which --;.ould. provide scientifically accurate yields, rf..,ic-- -.-Iould not be poli-.ically disputed by Institutes for Consurier Tests avid Government Departments, and ~.?hich would constitute an optinum nethod as regards its anolytical potential. Sinqe the results presented by the scientists on the Tobacco Resear::h Com:;Attee raised a doubt as to the exclusive use of Ganbr-J--'.-,-e filters, a joint investiGation was planned and eaarried out. The reoults -.7ore-made availcnble to the Pedoral -U-stry of Health (Prof.Krbller) for purposes of comparison v-rith iI-0s o,,-n invent.- 7ations by the re.1ference, method a ccording to Para 35 of the Pede.-a-- Pure Food Act, and to the BUF Gos. as evidence In wapport of its arguments. I. B--i se--ies In-f-luence of thia Canbridr-e filter on the. condenr.-ate :r--*eld ,ks a fixn2ti on of its st-ate o.," c1large yzith -sno'zo-concte:w"ate. Rzring t;,Is series of exprriments, cigarettes vithc-L-.4w filters from t---e "strong" sector of the maelcet were consumed by =echanical smokci=- repeatedly in each experinent accordinG to DIIII IC240: -oking -siachine; ILI 20, const.volume, bell-shaped pxcfile, closed. S201ce top: Tvio Ccw~brid!-e filters (central dfz=,eter q3.-m,:I in series behind each other. 41tib N.) I BATCo document for Legal Services : Health Canada 22 October 1999 2. random samipler3: n = 6 x 20 individual ciga~ettes Extent o L per experiment Extent of determination: Ghange in weiGht of the CaTabridGe filters. Kll the exprriments were, carried out by a single laboratory assistant-.on a single snoking at practically constv=t ., ma c h enviror-riental conditions (room temperature, relotive huniditY) and in a regular time sequence. I.1 Both of V!i, t-,-.,o uricharg;ed Cambrid,-e filters arr,-n~Ted in series bellind each other. Tabl3- 1: Incr,?Eise in weight of the Cambridge fil-er (mp.) 1 2 0 first (uncharged) second (uncharged) 1 649*3 10.9 2 688.4 8.3 3 712.1 13.0 4 674.2 8.2 5 676.5 9.2 6 677.5 4.0 7 1 680 + 21 8.9 3.2 a s 20 5 3.0 9 v 3:01,., 3 3. 71&-' The limiting values specified in line 7 cha-racterise, 4.,-.-- limits as at a 95'.1 prtbability of signi, --"i cone e. of reliability of the men: This also applies to the lines in Tables 2 and 3 for t, Js s ries h.: e of ey-p .erineats. The second Caibridge filter in somes has still increased by about L-1 of th.*?t proportion by weight which has been folznd to be present as "condensate content" on the first Cx-bridge filter in series. The high r2lative differonce3 in devia-11,ion in.line 9 bet,.,-iec-n columns I and 2 can be interpreted in terns of -the non- uni- 'ilter., This evidence ic of f0l-M clu-l-*t,,,, of the C-et'Aridtio -L CD --li CD 4 BATCo document for Legal Services : Health Canada 22 October 1999 3. particular intcrest because the Car-bridf-, ,se filters originated from a very narrow basic ran!~e of homogeneity, so that the phenomenon icnornn as charge differences, which must certainly be expected, need not be allowed for in respect of this unfavouz-able result of deviation. 1.2 First C--T-,br,--*d!-e filter in sez-ics wicharged, and second Ca -brid~7e filtor charged. During this e~mcrimetnt, the Cai-ibridl,7e filters of the previcus experiment I (Table 1, column 1) charged with srlolke condensate were used as the second Caubridge filters beyond uncharged Ca:-ib;idGe filters. Table 2: Increase in %-,reir-ht of the Cm-ibridg .,efiltr?rs (mg) ' 1 2 0 fi'rst (wncharged) second (charged, see Column 2 of Tnble 1) 1 676.9 7.1 2 664.8 4.9 3 656.0 4.4 4 655.0 2.5 5 677.0 8.7 6 667.7 4.2 7 666 10 5.3 ñ 2.3 8 S 9. 2.2 9 v 1.4',j 41. 5,'~; There are only random differences between the means sv,eci--7-'--;ed in lines 7 of Tables 1 and 2. The conclusions to be possibly inferred from this, that the previous condensate ch--:Lr.-:Ln- ----ould in this experinent exert no influence on the increase in weiGht 4cc.%%'k ~X'-04t '~ of the second Cambridge filter ~ZrijpqLuid Twith that in DXperimentl are not however valid, as wtll be shoim now during 13.~.erinent 3 in this series. The relatively high differances in deviation betrieen colummis I and 2t as ",,ecifiod in line 9, correspond to those in the r1o 4 BATCo document for Legal Services : Health Canada 22 October 1999 preceding -7- eri-nient 1 and muctt be interpreted in the =rle wa-,r. The evidence there encountered is further confizned by the results of this experiment. 1.3 First Canbridre filter in the wa-ries ch..-.rged and second Canbri--r- filter unchf~rred. Purging rdth 8 "-puffs" of air -,.t 35 nl/2 se,2. Darina this experiment, the Canbridge filters ch-irged with condensate from the previous Experiment 2 (Table 2, Coliwn-n 1) were purr.,,ed with 8 "stand-urd puffs" of air each at 35 ml/2 see. 'a:-.bridge Cilter was arranged beyond these filters An imcharged C as the second one in the series. 0 Table 3:. Increase in weight of the Cxibridr7e filt2r (mg) 1 2 0 first (cr~arged, see. Column 1, Table 2) recond (uncharged) 1 38.7 1.7 2 3 -.3607 0.3 4 - 33.2 0.7 5 - 32.8 - 6 - 34.1 7 - 36 ñ 2.8 0.9 8 S 2.U - 9 v 7.2c,j The Cambridge filters charged rrith more tl= 600* of smoke condensate thus experience a decrease in weight oñ about 35 ng! corresponding to about 5,'~ or about 1.8 mg/cigarette, -ahen puffs" of air are pacsed through them as simul:-, tied zero- smoke puffs. The results in 3xp ,eriment 2 are subject to a ma:-.inum error corresponding to this "lose". Accordingly, the increase in weight noticed in this experiment is only apparently similar t.- that in 7xperinent 1. In fact, it should be bt~sed on an incref-s in mass of about 40 mg. a BATCo document for Ledal Services : Health Canada 22 October 1999 5. It follows that: At a given constant supply of smoke, the increase in wei-ht of the charged Cxmbridge filter is about 4 times as large as that of an uncharged filter under these experimental conditions It also follovis that: U'hen Cambridge filters are. used, it is apparent that the sr'..1oke condensate values for the cigarettes in the "lightll sector of the maricet are favoured (in terms of relatively lower condensate results) compared ;rith cigarettes of the "stronger" sector of. the market. 0 11. 13xperimental series 2. Comp-arison.of the condensate -orecivitation by Cambridge filters yzi th thnt due. to an el rctrostati c smoke trap. Glass fibre filters (diameter 44 mn) were arranged in series beyond either. a central electrostatic smoke trap or a central Cambridge filter, and their change in weirght vras, established. Filter-tipped cigarettes f ro-m the "mediu n. -strong" sector of the market were employed as the experimental cimarettas. Consixnption by smoking was in accordance with D12T 10240. Smoking machine; M-1120, const ant volume, bell-shaped profile, closed a) electrostatic trap followed by a Cambridge filter (diameter 44 mm) b) central Cmabridge filter followed by a Can-bridge filter (diamaeter 44 mm) -es 'Extent. of rc,,indori sam."'Ple: nj = 2 x 100 individual ci~,arett 4 BATCo document for Legal Services : Health Canada 22 October 1999 6. per experinent n = 5 x 20 individual cigarettes 2 per experiment Extent of detcr=ination: Change in -ieight of the second CambridGe filter in series (divnnoter 44 mm ) in the. moist and dr,' state. M I Quantity of th,~, =ro'ke conclersate xrociritate by the different types of S1110IL3 t_-I S. Lr,- I s 'P Daring this experiment, 100 individual cirarettes ---iere consuired by smoking in each test. A Ca:;-tbridge filter was arranged beyond-each smoke trap (electrostatic smoke trap or CavibridGe filter), and its weight was established -it the beginninr,; and Rr end of the ez-periment. In addition, this filter ,,ias aftcnwards dried (24 hours over silica gel) and weighed naain. Table 4 reprCduces the results. Table 4., Ourtntity of precipitate. Bleotrontptic p-..iokctr,-P/ Canbridge filtor. Increase in weight of the Gantbridre filter arranred be.,.?ond it .1 2 3 4 5 moist dried 0 Total(mg) (mg/cia) Total(mg) (mg/cig) I :Electrostatic 6.4 0.064 2 trap 2.6 0.026 2.2 0.022 3 CambridZe 2.0 0.020 4 filter 1.3 0.013 018 0.008 Columns 3 r-nd 5 cont-iin the c*,arges in weiCht in relation to a single cigarette consumed by smokin~g. These data can therefore directly relate to the other drata on snoke condensate to be specified in accordeance with DrNT 10240. 1, 13ATCo document for Legal Services : Health Canada 22 October 1999 These data ropresent the MaXiMUM pos-sible levels for the ar-ounts of condensate not cov(~red by the smoke traps. The experime.-atal results indicate the following conclusions: The p:-oportion of condensate retained in the snoke traps is clearly greater than 991,t, for the electrostatic trap as well as for the Ca:-,-.bridge f~lter. An increase in weight can be established in the Cambridre filter arrar-Ced beyond either of the two smoke traps, wad this cannot be interpreted in te--..,is of the possible further dcposit- ion of water vapour. The increase in weight o.1" these Ca---ib.ridge filters arranged beyond the central Cambridge filter is not significantly different than that in those arranged beyond th e electrostatic anoke tr.,--.p, and is 'less than about one po,.-.er of ten th--Yi the last signifimnt figure'to be quoted for the s:,iolce condensate content in accordance with DIN 10240, md about t-~-.,o powers of ten belov, the last significant figure to be specified for the condo-asate data on cig-arette paol:ets since 1976. The lovier results in lines 3 and 4 compared vith the nuLqorical data in lines and of Table 4 r.,Lay be interpreted as evidence for the. fact that pe-..-haps a Cvibridge filt~,r is possibly after all better capable of retaining more of the particle phase than the electrostatic smoke trap, but we shall show bolow th~-..t such an inference is rrrong. 2.2. Influcance on chqrginr;- of the Cambridge filter vith condcns-te on its ch.,.npe in weij-,ht due to the non-particle phase no -condensate). Durin.- this ex-nexament, the e.-,m,riment of 2.1 nodif4ed in NJI 9 BATCo document for Legal Services : Health Canada 22 October 1999 8. 0 that t,,Ie Cambridge filter arranGed as the second in series Tras ri-e-r-orsly charged vrith the condensate fron 5 cigarettec, 3,11d in that only 20 cigarettes were consumed by s-n"iAinFg. .,3--red nble 5: __R Increase in y~qLrjit of.-:t~hecc tho second in s;,rics. 2 3 'th condensvte uncharged charged nn- 0 ng/cig M G I Mectrostatic smoke trap 0.18 0.24 2 Cambridge filter 0.12 0.38 I., a charr This experimental result clearly shows tl ;ed Cambridge' filter absorbs nore of-. the fractions of smoke not belonging to the condensate from the non-Dcrti cle --phase of the tobacco, si-~.iohe t Uh-n an. uncharGed Cambridge filter. It folloas that: A CaT:foridg~,e filtor rei-ioves Gas phase constituents frcm the tobacco smokc from being avalable for wnr'.lytical ectim.-ation, de,)ondin,-,- on the e A.'ent of the charge on it, T-1L. also follow-- that: Possihlo diff orences in the data for --7ioke conde--is:.-!te content as obtf--~inod af"ter electrcstatic =oke preciyitntion or on Cambridoe filters a-ro not neco;%sarily, due to dif--:-,e:fe-'It efficiencies in capturinj the actual smoke condensate, but due to errors introduced into the results for srmlpce conde!Lqate results caused by the Cambridee filter. .;.hese t-,~,o c:ncliisions are further axpported by the evidence of 'h- exporiment,-.1 series 3 and 4. E:-zoftrimentzal s=ios 3. During t1lis experiment, filter-tipped oijarettes fre-m the 9 BATCo document for Legal Services : Health Canada 22 October 1999 9. "light" sector of the markot were consumed by sroking, whilst a Canbridge filter of diameter 44 mr, wac arranged beyond the electrostatic mnohze trap. HCIT was estimated by a potentio- graphic method on the gas/vapour rhase of the cic,,~arette rmoke. S,mokiag nachine: -R-,! 20, constant volume, bell-sh.--oed profile. smohe trap: electrostatic trap, with a Ca-,7zbridge filter (dianeter 44 mn) arranged beyond it. ent of rrndcm sa:,iplinrg 10 x 20 individiml nI to 3 cigarettes n 5 x 20 individual ci,;arettes 4 per excperirent Extent of deterninetion: dry syr:olce condensate content (mg&;g' nicotine in smoke (mg1cig) IICN in gas/vapcur phase (mcg/ciC;). Darin- 2xp-ririeirlb I in this serien, no Canbridge filter was arranged beyond the electrostatic snoke trap, during 2 the Cambridge filter arrwiced beZrond it r4as. uncharged, du-~ing B,xperinent 3 it had D.',:,eviously been charged with the s-,,:ol:e condensate from 5 ciraretten, and dl=inr,, 7,moriment 4 it had Previously been charged with the anwhe condc-noate fron 8 ciGarettes. Table 6 repr-,duces the results. The statenents in this Table are mcans f-ron 10 and 5, respectively, ropeated experi:.ients 0 (see extent of randon s,,:i:plinrr),, The limiting values indicate the linits of reliabilit-y for tho Ljea, .1c. at a 0,5,? of p robability significance. N.) I BATCo document for Legal Services : Health Canada 22 October 1999 10. Table 6: 1 2 3 4 5 6 Dry ITicotine in HO in gas Incr,.-~ase in -na.~ss or second 0 conderrIete vpmour ph,,se Cambrid-rie filter:Lri seriC7 (ml-1/cig) (M*E/cig)-"! mg/20 cig mg/cig 1 9.9 ñ 0.20 0.53 + 0.02 139.0 + 4.3 2 + CL-Pi 10.0 + 0.38 0.54 + 0.01 128 + 6.5 2.6 + 0.8 0.13 uncharged 3 D + 10.0 0.19 0.54 + 0.02 121 + 4.9 6.1 + 0.8 0.31 charge 5 4 E + 6'- L 9.9 :t 0. 30 0.53 + 0.03 117 + 6.0 8.1 + 0.8 0.41 ch,,,xge 8 (K should be unleos intended to kill. Tr,-.--,slntor) E = electrostatic s.-cke tra-P + CA-Fi = vAth Gambriid3e filter charge 5(3) = charged rath the condensate from 5 (8) cigf?'.rettes In this Table, the nwaber a.,ainst the line is the -lame as the PiLaber of the erneriment within t~ais series. The results in line 2, Column 6 corre spond to those in Column 2 of Table 5, and those in line 3, Colu= 6 correspond to those in Colvmn 3 of Table 5. rn Mus, tht?se ex-p-rimer-to confinn the results previously indicated. 37at in addition. the increase in weight of the second Cambri-rag,a- filtIc:..- in sories v-s it is increacin.71y charred viith c-.~ohe -ain T)roviden aupnortinG evidence for the -e,:ult:i condonzate. This 4 I in 1.1, 1.2 vnd 2.2. 2~',t Cjolw-m 4 of Cable 6 ,.hows thrit as the conde-2ioate ch-arge o-n the sc-co-nd C.-?L.1bridGe filte::- rzioe3, there is. a decreace in the HCE content o--F' the Gq,.7/vapou_- phase. It ollows t1hat: A Cr1ribridge filter 13 cap.:.,.ble cf Pbsorbin~3 diffe~.-ant arounts of K-C:T .a Irorl tile gao/vccrour ph_-ne of cig:-'r-re"u-to r-:-o!:o, dere:idlin- on tho e-*.tent of its charge with smoke condenzate. it also follo'::3 that: .,hen CaiibrJd,-e filters are errploved, ci-crettten in the I'strong" cxD 4 BATCo document for Legal Services : Health Canada 22 October 1999 il. sector of the maricet are more greatly faveured Ue_Zs 17M.1) in -2snoct of their IICIT content in the gas/va-nour pha.-e, cor.-Irared 4* Id 0 o i ,,h cicarettes in the middle strength or Ii lit ea o- IV. 4. In this series, an inventipation was carried out cn the gptsl anour p.7-as e f zrcn. a non-frilt er cilgarette in tI e st n c-'-Io r cf the rnan%et (Cir. ,aretute I) P-nd on th-.1-t frozi -- filter-tipped cigarette izi the lierht sector of the m=het (Ci,rra-ette 1T). U ._-aparison vas raade of results viAch Inad Ibean obtzE~iined after z parattion of the conden3ate a) vrith a C&,-.bridge -filter ch.-.rged rith s-,..ol-.e condensate and b) vrith rai une arg ~7 1; f h, ed Cc, _',,rid e filter. the gas phase 1.7as conducted by gsz cnromvtorf c#- :he amilysis of -.-ia the p-rint-out pattern -LLnder the folloain;,- C=' _-Zitions: I .ras chromatograph: Carlo Tiodel 2101 '0 lumn: PFG 5100 (100 n glass capilla-_y tubinr of internal diarmeter 0.5 mi) 2eriperature: OOC, isothe-rnal 6 min., then ed at 30C rise -,er inin. to 1100C. gas: ,a=.ier 3 =1 per min. ,21, e sn ohirn, conditions correspond to those in t'---. -p-_-ecez~ing -3:cp erin ent al, s ezri es e "he results in Table 7 below aro peak. a:--.as in It BATCo document for Legal Services : Health Canada 22 October 1999 T 1 e. 7 : 4 5 6 2 of ch,-_rjre with cc*-m-.,'ie-*- perl,, Tzain constituent Ch-=Ss in Si-71r-l for Cigarette in peak 2 (F Cigar t-1-e I I MM e 2 Mri + 25 + 6.6 68 - 20.7 1soprene + 23 + 5.7 57 - 16.7 3 + 10 + 9.2 - 7 8.0 4 + 7 + 10.4. - 10 10~.2 8 3?ro,)! - onaldehyde + 5 + 1.4 9 Acetone + 52 + 10.2 - 54 - 29.2 11. ixrol Pin - a - 5.4 - 10 - 5.4- 21 Acetonitrile + 48 + 17.1 + 16 + 11.3 25 Diace"'Y1 + 11 + 4.4 20 Penzene + I + 0.7 30 -fur,-me + 40 + 17.4 + 20 + 6.3 37- To?uene + 89 + 21.2 the The patte=' provided by re 1,1ts is co-:ipletely n(' honofreneous: .uibridCe filter al Jherear, the ccnde.-nsate chnrginr; of the C, caUses MoTe of the ricasured gns phase constitne-ats to reac' the gas/va7)our phase from Cigarette I, this trend is mainly in the case of GiCarette !I. It follo-s that: The (Jarbridge --filter aff,clts the w,.ount and camposition of the ga.s/vapc,ar phase, as a fanction. of the extent to w'tich it fis charged. it also follows that: 5e filters are e=loyed for the sep.- a n Uhen Ca7bridg, __,~ tio SMOAte condensate, the cha:igos cwazed in resract of ruriouilt -:nd composilion of the gas/vapo-a= phase eigrorel;,te, with ra --!7h condensate ccnta~it are different to those e:--porie-Aced in t'--e case of oigarnttes with a low condmisate content. Vo It Three b-,a,--103 of cigaretteo vic~re simkod durir-_- t",rie. M-) I BATCo document for Legal Services : Health Canada 22 October 1999 13. expe-ri-.ental Series: filte-,'?-tipred cir-arette from the "ncdiun strclnf.-th" sect? .or of th? cif-?.-~ette 3 m-ilter-tinned cimarette fron the I'light"' sector Q - L. of the ncL-ket "i r:aret' Vhe llat=ong" sector of v 6e C = no'n Mter cisarette f=o2:i -I- the market. Durin,r:* this serie,-, the smoic-e condensate rns mnturcd in a central electrostatic smohO trap and also deposited on P. central filtcr. A ca-1 -7rison was nedc be-tween the two zIPt"-0(13 of p On precipitation with r--,-I-rd to the mutual devi:-~tionz o ' their r -- ilts. k poosible cha-Ze in deviation due. to the slo-.- c,.;nbu!7:tion st-reayi vzs alr-.o invacti,atod dui-in,.- this e:-pcri-riontal series. CoftstLmption by snok-in:,- war~ in accordance with DTIN 10240. ~hokin.7 nachine: M-1 20, constant volump., bell-shaped pr,.;file a) open b) closed Taokn trap: a) elcetrnstotic evrolco trap b) oentral Ccr.brid..-e filter (dial- cter 93 =i-.W dual ciC;r-rette3 pe'r . xtent of randr)m 10 x 20 indiva exp-riment -on: contents of s- xtent of deter.'--inati noist mid dr, riolke condensate (mc-/cig) 13;.I.ent of ri -czlt: 3 ci-arette brands in each case a = OT)en ,qc = open, electrr,static trap r Go filter ad = !.-pcn, C=.:ibrid~j b = closed bo = closed, ell*?ctrostritic trap bd = closed, Cczibridge filter 120 s--oke pac-sa-es from 20 ci!rarettes. The e-_--_p,--rimo_nts in this series were condacted by a iinglo ~ laboractoi-i ncniztm-.t on a sinrftc s---.okin-- n-cchdr-r, r.'U rcgal:-~r N.) NJ 9 BATCo document for Legal Services : Health Canada 22 October 1999 e interra-is under pract4 (.I- ly cons ant enviroi-liaental I;ure, relF%tive atriosrpheri c hum-iditY) con&itions (te,:,.-ocra I was intended to eliminalte the possibilit-.- that individual repetitions of t',he e.--,poririents night be subject to w-rateMatic influences of deviation coi-mared with others. Control of riethods hf~d been confirned for the t,.*;,,) typf?s of procipit*,~.+_f -,-n,. and esp. e ci al a y fo r moi sture de te-ri-an-at ion in t he rnr oke conden-oante duriiig preliTrinaa7j e:?perijtents %.aich are not to 'he z;p-ncified in detail here. Table F3 re-oroduces the results. Table 8: 1 2 3 4 5 6 0 Conparison of deviations between -,:', (electrostatic and Canbxridi,~e CA filter n=!- 01 Trap open 2 F. closed2 02 a S I 'Ck Cig ,. condensate Moist 0.64 0.4096 2.11 0.37 - 0.1'7-i:o 1-12 1 2 dry 0.74 0 - 5 476 -645t. 68 0. 4 !-'-f' 32 1.1i 3 CIg.B condermate noist 0.64 0.40962-410-32 0 . 10 -- 41- 3 '0 1'4L 4 dry 0.63 0. 7744 S-0~O - 37 0-13*-~ 5 Cig.C condencate noict 0.44 0 19363-3~-):63 9 0 6 aj 0 , 0:38 06 0. y 2 35~ -V 44 Ir A.1 -( 7 E Cig ". condensate noist 0.44 0. 1936 0.35 0.1225 dry 0.31 0.0?,61 0.36 0.1-,C' Ci,-, ,3 conde.-noate Moint 0.41 .0.1(81 0.31 0.cq-~! 0 - d ry 0.39 0.1521 0.31 0. C'-;~-' Ci-,.C conde-nsate moist 0*24 0.0576 0.35 12 drv 0.49 O.N.C! Columns 3 =-d 5 contain the standard devitation3 of t-he data, and columns 4 Pmd 6 j)-r3vide the varirtncen. lines. I to 6 specify the deviations for de-( position on the Ca:'1101-idGa filtor, and lincs,7-12 the deviations for precin- V J-i ot. a in the el-~,ctrostatic smoize trap., t n comparison of the variances in t.,-.e first -,up c-L - cz lin :_Inli I BATCo document for Legal Services : Health Canada 22 October 1999 15. the corresponding variances in the second group of lines, it is, evident that the deviations for the first grOUIP 0' lines are U sometimes significantly greater. This is even more clearly marked in Col=n 4 than in Co lu;-n 6. In addition, it is worth noting the mean increase in deviation in the-t"i rst group of lines from the odd to the next even number of the line. This phenomenon is missing in the second 'group of lines. This has aqain been more markedly expressed in ColiLmn 4 I th-con in Column 6. It follovis frora. t-ese rpsults that: k determination of the smoke condensate content by means tion on a Ca,.ibridge filter is subject to the of deposit. influence of greater deviation than that in accordance with the electrostatic method. -This randers a Cambridge filter less reliable, and it can more readily provide incorrect results. Im estimaticn cf t','~ie viator content in the pxecipitated smoke condensate can lead to ifurther increase in deviation when Cambridge filters are used, but not, .-vhen electro- static precipitation is er.,iployed, and hence produce even more -unreliability in the canalytical results. The slojq combustion stream can reinforce this unfavournble effect of the CambridGe fil-~Ier. It also follows that: Ciacrettea frcia the "strong" sectox of the mr-unket can adver sely affect the reliability of analytical evidence based on tne use of Cambridge filters than those fron the BATCo document for Legal Services : Health Canada 22 October 1999 16. tllight,, sector of the market. Si zw " a rV During five initial ex-or--inental series, sone typical ci,:--rette b.---ands from the Ge='an manket have been investi,;cted, in order to assess the pr~cipitation of ciSarette snoke condensate on Cuabridge filters, in particular on comparison Ath the use of an electrostz p accordinn to ~i atic s.-ioke tra DIP 10240. As a result, vie find that: 1. A Cambridge filter can lead to objectively incorrect results of a smoke analysis, de ;.> pending on the type of cigarette. 2. A Ga4i-bridge filter can affect the gas/vapour phase from cigarette snoke, dependinT on the type of ct7aretto -.And on thq condensate. It may remove from it some subotances, vehich can then no longer be detected in the gas/vapour phase. 3. The effect of Cambridge filters of causing changes in the results of exoeriments 1--cks uniformity st~ th;at it is impossible to apply correction factors. 4. Quite from the phenomena in 1-3, the use of Cambridge filters leads to fLirther unreliability in the re---ults of analyses on conparison with electrostatic precipitation. Hcwever, sinc!i vie have neither exhaustively covered the whole of the cigarette market nor taken a re resentative sari-ple o-,-' it, the results can only have an orientatinS charaelter, and me~rely indicate the direction to bQ followed in further invest:~4r tions. They are however sufficient to indicate the advisability of not conter,.plating the us,- of a Cambridge filter rr, the only method for the precipitation of smoke condaina-te in accordance with a Standard. NJ -tt- I BATCo document for Legal Services : Health Canada 22 October 1999