Tobacco Documents Online

, A SOt~,' STATISTICAL OBSERVATIONS ON A COOPEItATIVE STJDY OF HUMAN PUIX0FLARY FAT80IQCY by Edwin B. Wilson and Mary H. Burke Office of Naval Research, Boston, Massachusetts, and Tobacco Industry Research Cosmnittee, New York, N.Y. Because of the reported persistent increase in the incidence of cancer of the lung, Dr. Stanley P. Reimann of the Iankenau Hospital in Philadelphia arranged for the Scientific Advisory Board of the Tobacco Industry Research Coa®ittee a cooperative study by a number of pathologists in different centers across the country to examine systematically the conditions at autopsy in the bronchial mucosae of run-of-the-mill (not specially selected) cases. HS has given a general account of the design of the atudy and of its major findings.(1) It is the object of the present paper to give some of the statistical details with comments upon them so far as the total picture can be put together at the time of writing. The conditions found on the pathologists' slides were classified in ascending order of pathogenicity from normal to carcinona in six ranks: 0. Noraal, 1. Eyperplasia, 2. Sa,uR*M+uA Metaplasia, 3. Atypical Metaplasia, 4. Carcinoma-in-situ, 5. Ca.^cinoma - based upon the worst condition found anywhere on the slide. It is obvious that in such a classification a considerable degree of -judgaent is involved as the line between adjacent classes is not so sharply definable as to be interpreted in identical menne:- by different judges; indeed it is not always poseib e for a single person to repeat precisely a classification he has previoualy made.~2) Thus the pooling of data is almost certain to involve vhat statisticians know as mixed classification, but it is one way to get a sort of standard with which s:aller aanples aay be compared instead of co=paring each and every sat ple with every other. The pooled data for males by age is as follows: Table 1 Males Percentage Distribution Over Classes Aae SlSdes 0 1 2 3 4 5 Under 25 666 86.6 6.0 6.3 1.1 - - 25 - 44 1,240 70.2 11.9 11.5 5.7 - 0.7 45 - 64 3,982 70.0 12.4 14.0 2.3 0.1 1.3 65 and u 2,702 68.8 12.8 13.4 3.2 0.2 1.6 Total 8,590 70.9 12.0 12. 3.0 0.1 1.2 25 and up 7,924 69.6 12.5 13.4 3.2 0.1 1.3 As Dr. Reimann pointed out(l) there is no substantial increase in the percentages in each class after age 25, except in class 5 (carcinoma) where It would be expected. CTR.I COh`FIDENT4AL: MttiN8s0'lA TOBACCO LITIGATION 142G ~,r tm t'w.. i tH 00,3'°°t"48

- 2 - If it be desired to compare males ar.d feaales it vill be necessary to restrict the coverage to those places vhrch sent in data for both aen and vomen because it so happens that those pathologists reporting on only male autopsies had a slightly different distribution of percentages by classification than those reporting both. Table 2F Feaales Percentage Distribution Over Classes Age Slides 0 1 2 3 4 5 __ t nd-r 25 1+0o 91.8 6.0 2.0 0.2 - - 25 -4-+ 803 88.2 5.2 6.2 0.4 - - 45 - 64 1,157 81.8 8.3 7.4 1.4 0.2 0.9 65 and up 1,301 86.2 7.3 5.2 0.9 - 0.4 Total 3,661 5.9 7.0 5.6 0.9 0.1 0. 25 and up 3,261 85.1 7.1 6.3 1.0 0.1 0.5 A.-.d for the ccrresponding males: Corresponding Males Are Slides 4hble 2M Percentage Distribution Over Classes 0 1 2 3 4 5 Under 25 615 86.5 6.2 6.2 1.1 - - 25 k5 65 - 44 - 64 and up 926 2,871 2 04L 73.3 74.1 72.6 11.4 10.1 11.6 11.3 13.1 13.0 3.2 1.6 2.2 - 0.1 0.1 0.8 1.0 0.6 Total 6,45'- 74.7 10.4 12.1 2.0 0. 0. 25 and up 5,6L'3 73•4 10.9 12.8 2.1 0.l 0.8 a is seen that at ail ages women have higher percentages of normal slides an3 lor+er percentages in each other class except for an insignificant excess in carcinoma-in-situ at ages 45-64 based on two cases. Further, it Is seen that for women as for men there are no really substantial changes after age 25. As to the slides for cases vith prisary lung cancer in males, the n=cers and percentages are: Table Percentage Distribution Over Classes A,Ee _ Slides 0 1 2 3 4 5 25 and up 596 50.3 8.4 18.6 5.5 1.2 16.0 Comparison with the last line of 7able I shows that the percentages of normal and hyperplasie are down whereas the percentages of inetaplasia, squamous and atypical, and of carcinoma-in-situ and carcinoma are up, as might be expected. (,1kI i 1427 CONPIDENTIAL: MINNESOTA - TOBACCO LITIGATION C-1-r•gh NN 0103449

-3- Even if adjustment be made for the 16d of carcinoma by dividing by 0.84, the statement remains true that normal and hyperplasia are down and metaplasia, seuamous and atypical, and carcinoma-in-situ are up; for the percentages become 59.9, 10.0, 22.1, 6.5, 1.5 compared with 69.6, ~ 5, 13.4, 3.2, 0.1. unfortunately these results cannot be compared with Auerbnch'a'•1) because hls classification azii his method of collecting data are both different from those employed in the study. Corresponding to the last lines of Table 1(al1 males 25 and up) and Table 2F (all females 25 and up) the diffexent cities for which we have records of individual percentages give these results for the means and for the standard deviations of the distributions about them: Table 4 Percentage Distribution Over Classes 0 1 2 3 k 5 Males 65.2± 13.5 15.6 'f13.1 15•0'~'5.3 2.9 ±2.0 .15 t.16 1.1 ~_. Fe=ales . E1.2 ± 9.2 12.0 ± 10.4 5.8 ± 2.9 0.7 ± 1.0 .03 ±.02 0. 2 ±C . A1:boug~-, the means of the percentages do not differ materially from the mean percentages, the relatively large estimated standard errors of the distributions about ther are noteworthy and signify that either the pathological conditions in the different places a_re different or that the different pathologists were cl.assifyir.g 'ifferently or both, aM there is no way to distribute the variation over these two a-.,urces. A^n experinent vas made by asking each of the twelve participating patholog'_sts tc send in a few slides, some typical and others problematical. Frcc= these Dr. Rei=anrn selected 40 which were classified indepe^.dently by a±i twelve. ZT:ere were therefore 480 items spread over the 6 classes 0 to 5 respectively as 46, 120, 223, 50, 7, 32. Division by 12 gives the average values 4, 10, 16.6, L.2,.0.6, 2.7. For simplicity one may cor.:oine 0 and 1 (normal and hyperplas:a) to get 11+, keep 2(scuaaous metaplasia) at 18.6 a~ combine the three most serious conditions 3, 1: a^.d 5 at 7.4. A.'•hi-aouare on the table of 3 x 12 observed velues gives about 75 for 22 Q~~ees of freedom which is very significant of differer.ces in the classification l. Tbere is no way to estimate what would have been the differences or their significance had the twelve pnrticipaiing pathologists each read the 12,000 or so slides in the main study or any ra.•sdom sample of them sufficiently large to establish a reasonably stable set of data for the purpose of codparison between them. If one were satisfied to make three groups, as above, namely classes 0 and 1, class 2, and cla-ses 3, 4 and 5, it would be well to have 400 or 500 randoa.ly selected slides reac by several pathologists. Another way to deal with the matter is to make a correlation table of the pairs of classifications by the 66 pairs of classifiers. The table may be constructed in a syffinetrical way resulting in 132 x 40 = 5,280 entries, each reader appearing as Reader I and as Reader I:. 1428 CONPtOENT(AL; MINNESOTA ~""~, TOBACCO LITIGATION CT N N 0 ~'~ 3- 4 E-5

_ Table 5 4 - Syrsnetric Correlation Table Tor Readers 0 1 2 3 4 5 Zbtal c lgo 189 115 34 - - 52& 1 189 584 409 129 1 8 1,320 2 115 409 1,696 208 19 6 2,453 3 34 129 208 136 13 28 550 4 - 1 19 13 12 32 77 5 - 8 6 28 .2 278 352 Total 528 1,320 2,46 550 77 352 5,280 - In this table 54.9% of the entries are in the main diagonal shoving that degree and only that degree of ar- eement between the pairs of pathologists. If the two adjacent diagonals te included the result is 87.1% of the entries, indicating that 12.c,)% of the entries were for cases in which two pathologists differed by tvo 9- more classes in their estimates of the most serious condition on a slide.,4) One car.not convert the correlation table into a standard correlation coefficient without implying that the six clas~~i'icatior.s C to 5 are proper r.;~erical measures of the relative pathogenicities. ) Eut i_' c::e uses tbis assurptio: rI -,I = 0.692 with 5% and 95% points at r: 0.576 a nd r= 0.787 as figured by Fisher's z-transfor-•.ation. And on this index the mee- is 1.883 (* .190 = t1.183/F391. If o7e is villing to rank the different readers by using this r.=erical measure of relative pathogenicity as an index one finds for the twelve foliov:ng: 1.575 ± .196 1.600 ± .211 1.650 { .204 1.675 ± .201 1.700 * .173 1.875 ± .169 1.9CO ± .179 1.925 ± .177 1.950 ? .179 2.125 ± .161 2.225 ± .198 2.400 t .167 It was interesting to us to note that the order of these means for the different vathologists correlated well with the order of leniency appea.ring in their reading of all their ovn slides. °'aSs suggests that the distributions of their findings in their own hospij.,als are due in part to their habits of interpreting their slides and that inter-institution comparisons of their fir.3ings should not be made without some allowances for these differences, I.e., for what astronomers are wont to call the "personal equation" of the observer. Classifying an individual slide is a "clinical" problem as is any exercise of jud gaent about any individual--case whether pathologic, meterolog;c, social, economic, medical or psychological. Statiatics deals with a population and there mey be such wide scatter about means or trends that the statistical CTF.1 1429 CDNFJDENTIAL: MINNESOTA TOBACCO LITIGATION C 1 '° R ~ ~( ~ ; D ("~

5 findings may be of little use to the clir.ician. Moreover, no statistical stud,y can include all the variables which may be acting. An important function of the clinician is to be alert to the special i.Sportance for the particular case of some variable vhi:h is not in the statisticians' findings for populations but may be controlling in the case before him; and moreover there are aspects of a.ny clinical proble.m which may be sensed by the keen and experienced cl.in~s~a.n but are not yet generally incorporated into standard statistical nethods (6. Even had it been practicable, the cooperating pathologists should not have been forced to agreement vith respect to standards of classification for this study at the danger of modifying their individual standards; for they were all experienced scientists primarily responsible for long series of pathological determinations in their respective hospitals and should, therefore, be concerned vith the continuity of standardb in those series. Science is built upon the deductions one can draw from data freely assembled by different scientists vit.h proper allowance for the inevitable differences between scientists. One can come to forced agreements giving an artificial sense of security. The cooperatir.g pathologists after getting vell into the study were unaninously of the opinion that through it they were learning much that vas new to them about pu.L:ana,.ry patholog~,, and profiting greatly from their mutual contacts relative to it. - The original hope of Dr. ReimAnn's plan was to correlate the pathological findings with such environmental va_riables as amount and type of smoking, rural or urba-n residence, geographic location, occupation, etc. We hope that developvents not ready for this meeting may later make it possible for us or others to extend this statistical discussion to some such matters; but in the short time available for the presentat.on of this paper v- have been unable to exhaust even those materials es.d calculat=on.s which we cotaider germane to the topics here discussed. Q T,A 1430 CONAIDENTIAL: MINNESOTA ~~ t-=~ ` TOBACCO LITIGATION C T Rl~"~C)

- i - (1) Reimann, S. P., Pathologic-Anatomic Study of Eur$r. Lungs. Tobacco Industry Research Committee Coc..unication, June 1957 (to be revised for publication). The agreement was to use approximately 15 sections in certain definite parts of the tracheo-bronchial tree. At least ha1S of these sections were to have as =ch as 50% of the mucosae intact. The average number of such sections proved to be about 12. This vas not ao detailed a study as that of Oscar Auerbach (Nev England Jour. Med., 2-r,,6-, 97-104, January 17, 1957) who took a great mnny sections, but his study vas at a single place and has reported only slightly over 100 cases whereas this study involves over 1,000. (2) Studies of "errors of observation" hcve long been made in the exact sciences on the assumption that there vas something under observation vhicb vas far less variable than were the observer's possibilities of measuring it. In the medical, and other inexact, sciences the vord "error" has sometimes bee^n used where differences of information or of judgaent in complicated situatios are involved to such a degree that to attribute error to those who differ is undesirable. A few of the earlier studies of reliability are: R. C. Cabot, "Dia$nostic Pitfalls Identified During a Study of Three Thousand Autopsies", Jo•,L-. Amer. Med. Aasoc:, 59, 2295-98 (1912). E. E. Southard and A. W. Stearns, "The Margin of Error in Psychiatric Eospital Diagnoses", Boston Med. Surg. Jour., 171, 895-900 (1915). Edwin B. Wilson and Julia Dening, "St.atistical Comparison of Psychiatric Diagnoses in Some Massachusetts State Eospitals During 1925 and 1926", Quart. Bull. Mass. De:~t. Mental Diseases, Vol. XI, 15 pages. Reprinted in Schizophrenia: Statist:ce: Studies froW the Boston Psychopathic Eospital, (1925-34), collected for private d:stri'outiorn by Dr. C. Macfie Campbell. C. R. Doering ar.d A. F. Raymond, "Reliability of Observations in Psychiatric and Related Characteristics", Amer. Jour. Orthopsychiatry, k, 249-5? (1934), reprinted in Schizophrenia. r. E. Slegler, "Microdiegnosis of Carcir,o-_a-in-Situ of the Uterine Cervix: A Co_-pa-retive Study of Pathologists' Diagnoses'; Cancer, Q, 463-69 (1956). C. C. Birkelo, W. F. Chamerlain, P. S. Fhelps, P. E. Schools, D. Zacks and J. 7erushalr:y, "Tuberculosis Case Finding: Comparison of Effectiveness of Var:o•.a Roentgenographic and Photofluorgraphic Methods", Jour. Amer. Med. Assoc., 1J3, 3 59 -66 (1947). (3) If one does not wish to have so small a theoretical value as 7.1+ !n the Chi-Square test, one may aske tvo divisions, viz., 0 and 1 and 2, 3, 4 and 5 with the result that C5i-Sauare is about 30 for 11 degrees of freedom whereas, the one per cent point is at 25. (4) This shows that actually in practice the spread in the judgments stated is not just a borderline affair. Another vay to show the differences is by co ,mxring the least severe and the most severe readings of the individual slides: CONFIDENTIAL: MiNNESOT& ~ TOBACCO LITIGATION C T L-~ ~y

Most Severe ~ Least Severe 0 1 2 3 4 5 Totals 0 - 1 8 10 - 19 1 - - 4 6 1 1 12 2 - - 2 2 2 6 3 - - - - - 2 4 - - - - - 1 5 - - - - - 0 Total 0 1 14 18 3 4 40 Fote the peculiar distribution of the frequencies of the differences in class between the least and m.ost severe readings, namely, 2, 8, 18, 11, 1 respectively for differences of 0, 1, 2, 3, 4 classes. Still another way is to tabulate the 40 alides according to the rnL:.ber of different classes used in classifying each. The frequencies are 2, 11, 17, 10 respectively for the use of 1, 2, 3, 4 classes. There is no slide for which all readers agreed that it vas normal or that it vas carcinoma. (5) It would be a separate as.d probably involved study to deter-m:ne how an index of pathogenicity should be made up. Paturally "normal" would be taken as 0 and hyperplasia tcight be assigned the figure 1, but apart 1'rom some intricate considerations of pathology and statistics one could only assign classes 2, 3, 4, 5 general values m, a, s, c to stand for vhatever m;g.ht be the pethogenicit:es of those classes relative to hyperplasia. That 1<m<a <s <c would be assu.med because of the basic assumption that the classes are 1n increas:r.: dee.ee of pathogenicity. Consider a si=,-le correlation t~&ble bettireen two readers of the 40 slides such as: Reader B Reader A 0 1 2 3 4 5 0 4 1 2 - - - 1 6 2 4' - - - 2 1 1 13 - - - 3 - - 2 1 - - 4 - - - - - - 5 - - ~ - - - 3 It is clear that these readers excess of Reader B over Reader agree in 23 of 40 cases and A on the index wouid be disag*ee in 17. The E - lt2n+4(m-1) - 6 - m - (m-1) - 2 (a-m) = - 8+6m - 2a and any values assigned to m>1 and a> m which make E positive vill make Reader B severer than A. If we set m- 1= m'>0 and a- m a'>0 ve have E=-4+t4m' - 2a' as the excess. If we plot a' as ordinate and m' &s abscissa we can say that for 1432 CONF(DENTIAL• ~ ~~~ ~•~ N N 0 ~ ti ~ ~ ~ '~ ~ ' TOBACCO LITIO rjp~ rA

any point (r.', a') in the first ouadrant which lies above the line -4 t 4m' - 2a' = C Reader B will be less severe than A, whereas for any point below it B vill be more severe. If we take m= 2, a= 3, m' = 1, a' = 1 as we did, then E-2 and B is the less severe. (6) These matters have been discussed by P. E. Heehl in his book Clinical Versus Statistical Prediction, University of Hinnesota Press, 1954, and before the 195 convention of the American Psych.ological Association under the title "tithen Shall We Use Our Heads Isistead of a Formul.a7" and in his article "Configurational Scoring", Jour. Consulting Psychology, 14, 165-171 (1950). , CTR,/ 1433 C01\F(OF.NTIAL: MINNESOTA ~'~ `i~R, NN t~.~0,1`4`D:``_ ~ TOBACCO LITICA7YON

. wA-7./~`g.Cr SC, J)',' CS i'nc-c7i)t(- f((/.Js -~~-~Y) % for a thort time beyond the Latent pcriod. The period of bgarithmic increuc ap- pean to represent primarily the .ucmbl) phasc looa TAatK Reclrferler /wrtiture Provocaion of Maaked Hog Cbolera Vuw in Lunrworm•lnfated Swiae by Mcaris Lar.ae Conaideration of the epidemiolo6y of bog cholera suggests that its eauaative .irur must be perpetuated in nature in a oonporc.inc teservotr intermediate ban. Ezperimean conducted with the twiae lunw+orm indicate tAat this oematode un Krve as a rtyervoir hoat for the bo( cbolcra virut but ako that it darbon th v'uut in a muked or occult notunfective form. Swine fed lun"orm larvac eontaio- iag the maaked bog cholera virus ordi- oarDy do oot come down wich bos eholera. However, the appcarance of good bealth abo.+n by tuch animals -u mialeading be- eaux all that ia roquired to bring them down with a fatal attac] of bog cholera is the appheatioo of .ome relat'rvely iaaocu- out prvvocsitivt atreu In the earperimenu rtported, ml(Tating AlCaril IarYa< fup- plicd the stimulus that prvvoked maaked bog cholera virus to infcctivity. Rtcsa~w E. Saot•e Rockrfeun lartit~te Amino Acid Ttunovcr in Brain Compared with Turno.cr in Other Tcssucs A litur of mice was produced whose tissua were labeled by feeding C"-glyrinc and aerine to thc'v dam for 19 days (6 days before and 12 da.s aftcr the young were born). One young mouse wu ki7led the da,v C" feedins waa stopped, the re- maioder at intervals up to 136 days there- after. Five aeu of tiatues from tbcu anima2s were analyzed for ndioactiviry: (i) braiq (ii) heart and lungt. (ii.i) lives, (iv) kid- ney and tpleen, and (v) akeletal musde. The liver, kidncy, and apleen in the 136-day animal abowed tso radioattivity, wherw the brain and the tinua eonuin- ing muscle were tti.11 at high levela. Tbe implicationa of these ti.ndingp were brieAy discttaaed. JoatrR W. SnLL Geerfe {S'uhialtoe Univerrity ScAoot of Afedicine Eahanced Folic and Fotinic Aad Cootrnts of ErTthrocTteS Iraf ected with Milaria Yarasi ta There is good evidence that folic acid playt some knporunt role in the meub- olum of malaria parasites . Thus urtun antimalarial drugt appeu to act by in- terfering with the utilization of folic or folinic acids The eatraceUular aun•ival ie vino of a species of bird tnalaria (Plar. rnodium fophurer) has been found to be prolonged b) the presence of folinie acid in the culture medium. Assays of the 1236 folic and folinic acid contcnu of the eryth- rocytet of normal ducks and of ducka in- fected with P. fojhurae dave oow shown thas infected red cells contain much more of both of these growth factors than un- inlectcd oncs. Only about one-fifth to one-third of this increase u conuined in the paruitet themselves, the other four- 6ftha to two-thirds being contained within the infected boat ceU material. TLc ob.u- vatioeu art coruiutent with the hypothesis tbat invuion of an crythrocyte by a pa»• tiu alters the folic aud metabol;.m of the boet eeU in a manner of besx5t eo the pars.site. Roclefel4r Irutitrte is uaed. Scrotonin is )0,000 timet u acti, e aa C.CI, on rat uterut. Finally, it is ..-c11 known that indolcacttic acid un be rc- placed by compounds such u 2,4-dichlor- opbenoryacetic acid u agrowth regulator for planu, and that both aurtinliie and anttau-idn effecu of 2,4.1) can be thown The corraponding atnine-that is. 2,4- dichlorophenoaryetLylamine-.howt uro- toninlike and antiurotonin actions on rarioue an~"al tiaauet. A close correlation aoan between the types of rwctural altmtion whicb chaage the au3tin activity of 2,4-D aad those which affect the Kfo- loninlile actions ef 2,4-diehloropheno- Wat.twat Ta.woea stbytamine. Some of these aubt6tuted I phem:yethylamina are very active u Some Satiatieal Obxr.atiom on Cooperati.c Study of Htanan PalmotaarF Pa&oloty Statilcy P. Reimaan of the Lanketuu Hoepiul in Philadelphia arranged, for the Seientibc Advisory Board of the Tobacco Industry Raearcb Committee, a study by a number of pathologists in diAertot eets- ten to ezamiae the eonditiona at autopry in the broncLial mucotae of unaelected eaau; be haa reported thereon. The ob- ject bere is to give aotae of the autiatical detai.L with special ydexnce to differ- enca berw"n differsnt eroten. Thc pa- tbologiru daai6ed their slides atxording uo the worst condition found thereon, from normal through bypesplaua and meu- plaaia to easeinoma. The prrtentaaes of alida in tbe different elauer differed widely from plau to place. To aee whether the patbolo6iau were differing, a.elccted aample of 40 tlida wu presented to all of tb=, and marked differ-ences were found FtTther studies ue in prosreas to enable the pat'ologuta to allow for their rystem- atic di,°cntoces in makina eompari.oos ber..cen their findings E.owtrr B. Wu.aor+, Mwav H. Btract Otrr of Naval Retiarch and Tobacco lrtdsutry Retrarch Committrr . . _~ • robablc Evolutiottar7 Relitionship o( Serototaia and ladeleaoerie Aiid TLia paper presents evidence for the idei tkat aeroton;n (S-hydroxytrypu- taitu) is the bornwne in aniatala eorre- rponding to indoleacetic acid in planu. Then is 6nt the reLtionahip in chemical rtruerure, aince aerotonin is the bue (modi6ed by bydrorylation) correspond- ing to indoleacttic .dd. Secondly, there ir mounting'evidcnce of a rexmblance in basic mada of action of the two tub- ttanca lodoleacetic acid changes the per- taeibility of plant eclL, causing tLent to uke up water and thus to elongate. Sero- tonin u known to cause muscles to eon- tract. It is now ahown that this action is aasociated with a change in the permeabil- ity of the muscle eeU to c.alcium ions. When ulcium ions are xquestered with Vencne, aerotonin no longer is able to cause isolated rat uterine muiclc to eon- vut. In the absence of Veraene, calcium ions can rrplace terotorun in causing eon- tnction, provided that tufficient calcium pro and antisetotonint. D. W. WOOLLeI' RockefeRe+ Irutitrte A=nic Serial Calttve in Ce11-Frec Medium of Enumoeba in.adcns, a Pathogenic Amoeba of Snakcs A rtrain of Eetnnoeba iavaderu, ordi- narOy frown in eultura of mixed bacterial species, was enabliahed bacteria-frec by 1Siller (1951, 1953) with the aid of fres!. liver tissue. With earaca of raw liver atenaiud by filcration, Miller reponcd aome but no continued fr•o.+th. T'he pneaent study tuned with a Miller culture received (from Dr. E. McervvitcE, Institute of Paratimlogy, 1(.cdonald Co:- lep, Quebec) in aerum=saline eontainioc a pieu of hamrta liver. Axenic 6rowtL ic a txll-frec mediun was rubacQuently ini- tiated, and aerial cultures vrere mainta:nec for 16 months thereafter. Tbe buic medium .nt autodavcd lis cr b:nth, used with or without pcptone, and supplemented with a crude utract fror7 raw liver prepared u a ehi:icc, acid ir.- ftuion and tterili.ted by Seitz EJaatier (StoU, 1953). When, in addiuon, muc.r is tupplied to this medium o! eoc.binec beat-uable and heat-labile liver infusion3, ere is ttracing enhancement in yicld ar.c atrviva) o! the entamocbac. Nc ar.t:bicc:c! a.rc employed. The cultures arc .-ax:.r• .ealed. One interest that atuches to the s_:- eaaful erootinuous eultivation of E. s- s.drw a.zrn'seally in the absence o! cci~~ is that it tnay be a ttep tow'a:--; a ric.::a: recult with tucL a species of the genus a: E. Airtofytica, the pathogenic amoeba c! man. Another ia that the avidit,for muc:r. diaplayed by E. itoadnu in axer.ic cultcrc, u found by Rateliffe (1934 ) in r..:=ec ba: - Wxial cultures with the paruite. ma, be a elue to the saatural eute of aftain in thr boat-panaite relation o( this gcn's Noaw- R STO.: Rockefeper lutitrre Hypotbemu for Mecbanism of Action of CbTmotr7Pain Investigations of the active site of ch, - motryptin have shown that a serine mo:r eule is acylated during the reaction, ar.c they have elucidated a partial pept:c- aequencc in the neighborhood of this sr. rine rnidue. Kinetic in.esu5at~ons ha- 1434 SCIENC£. VOL 126 C(-- <- I'-, CONFIDENT[AL: MINNESOTA. TOBACCO LITIGATION C T N 0003 ~t ~ L