Collection Czechoslov. Chem. Communs.
(68'2) Witkop, B., Goodwin, S., Beiler, T. IT., J . Am. Chem. SOC. 76. 5813
24, 645 (1959). (668) Wagner, H., Berg, H., Electroanalytical Chemistry 1, 62 (1959-60). (669) Waksmundeki, A., Oscik, J., Frelek,
(1954). (683) Wituicka, J 63 (1955).
Z., Ann. L'niv. X a r i a e Curie-Sklodowska, Lublin-Polonia, Sect. A A 10, ?io. 2, 17 (1957). ( E O ) Wang, E., 2. physik. Chem. (Leipzig) Sonderheft, July 1958, 134. (671) Wawzonek, S., Fredrickson, J. D., J . Electrochem. SOC.!06, 325 (1959). (G72) Wawzonek, S., Wearring, D., J . Am. Chem. SOC.81, 2067 (1959). (673) Weidner, G., Hoppe Seylers 2. physiol. Chem. 312, 202 (1958). (674) Wenger, P. E., Rlonnier, D., Vogel, J., Mikrochim. Acta 1957, 405 (675) Whisman, RI. L., Eceleston, B. H., AXAL.CHEM.30, 1G38 (1958). (676) U'hitnack, G. C., Gantz, E. St. C., J . Electrochem. SOC. 106, 422
, Posfepy Biochem.
1,
(684) Zbid., p. 85. (685) Zbid., p. 101.
(686) Witwicki, J , Zjlo, J., Ibid., 1, 113 (1955). (G87) Woggon, H., Ernahrungsforschung 4, 35 (1959). ((388) Wronski, hI , Roczniki Chem. 31, 1065 (1957). (G89) Zahradnik, K.,Chem. listy 51, 937, ( 1957): Collection Czechoslov. Chem. c o m m h n s . 23, 1529 (1958). (G90) Zahradnik, R., Zbid., 23, 1435 (1958). (691) Zahradnik, R,., Svatek, E., Chvapil, A I . , Chem. listy 51, 2232 (1957); Collection Czechoslov. Chem. Communs. 24, 347 (1959). (G92) Zahradnik, 11.) Zuman, P., Chem. l i s f y 52,231 (1958); Collection Czechoslov. Chem. Communs. 24, 1132 (1959). (693) Zhdanov, S. I., 2. physik. Chem. (Leipzig) Sonderheft, July 1958, 235. (691) Zimmer, A. J., Huyck, C. L., J . Am. Pharm. Assoc. 44, 344 (1955). (695) Zobov, E. Lyalikov, Yu. S.,
(1959). (677) Wienhaus, H., Dassler, H. G., Chem. Ber. 91, 260 (1958). (GT8) Wild, A. M., Chem. & Znd. ( L o n d o n ) 1957, 1543. (679) Williams, A. F., Brooks, J., J . Polarog. Soc. 1958, 5. (680) Williams, A. F., Kenyon, D., Talanta 2, 79 (1959). (681) Willits, C. O., Ricciuti, C., Ogg, C.
Izvest. A k a d . A'mk Turkmen. S.S.R. 1958, ?io. 1, 93. (696) Zrilimi, G., Lanza, P., Ricerca sci.
L., Morris, S G., Riemenschneider, R. W., J . -4m. Oil Chemists' SOC.30,
420 (1953).
26, 2512 (1956).
(697) Zumun, P., Ceskoslov. f a r m . 7, 8 1 (1958). (698) Zuman. P.. Chem. listv 52. 1310 ' (1958); Colleciion Czechosov. ' C h e m . Communs. 24, 2027 (1959). (699) Zuman, l'., Chem. lzs!i/ 53, 154 (1959); Acta. Chim. Acad. S c i . j l u n g . 18, 141 (1959). (700) Zuman, P., 2. physik. Chem. (Leipzig) Sonderheft, July 1958, 218. (701) Zuman, l'., Cerny, V., Cheni. Iza/y 52, 1468 (1958); Collection Czcchosloz~. Chem. C m m u n s . 24, 1925 (1959). (702) Zuman, P., Chodkowveki, J., Potesilova, H., Snntxvy, F., Xalure 182, 1535 (1958). (703) Zuman, P., Fedoronko, >I., Z. physik. Chem. ( L e i p z i y ) 209, 3T6 (1958). (704) Zuman, P.,Kabasalkian, l'.j J . Electrochem. SOC.105, 758 (1058). (705) Zuman, P., Kru icha, J., Collection Czechoslot'. Chem. 8ommuns. 23, 508 11958'1.
(7OG) Zbman, P., Kuik, &I.,'Vaturwz'ssenschoffen 45, 541 (1958). (707) Zuman, P., Sicher, J., Krupicka, J., Svohoda, RI., Chem. listy 51, 1068 (1957): Collection Czechoslou. Chem. Comm'z;ns. 23, 1237 (1958). (708) Zuman, P., Zahradnik, R., Z. physik. Chem. (Leipziy) 208, 135 (1958). (709) Zuman: P., Zumanova, R., Tetrahedron 1, 298 (1957). R., Teisinger, J., (710) Zumanova, Zuman, P., Chem. zcestz 11, 517 (1957).
Review of Fundamental Developments in Analysis
Statistical Methods in Chemistry Benjamin
N . Nelson
Westinghouse Electric Corp., Pittsburgh, Pa.
A
of statistics applied to analysis appeared in 1949 in ANALYTICAL CHEMISTRY (210). Since that time the series has continued in 1952 (SQ),and in 1956 (131) and 1958 (132) with the present title. The purpose of this review is to continue the coverage of statistical applications in chemistry and to review some of the ne\v developments in statistics, hooks, periodicals, courses, and meetings of pertinence. This review covers the literature of the period from the last review to September of 19.59. Coverage is principally of U. S. origin with limited review of the foreign literature. Papers covered are those that have been published in the open literature-not meeting transactions or documents of limited distribution. h-o attempt is made to review the field of statistics or quality control in its entirety; only certain pertinent developments of present and potential interest to chemistry are disci:ssed. Regarding chemistry, coverage is limited to papers in which there are other than routine precision, bias, and lea$t-square fit applications. There is REVIEW
some coverage of industries allied to chemistry. BIBLIOGRAPHIES, OTHER REVIEWS, A N D JOURNALS
BOOKS,
As sources of information on n h a t articles and books have appeared in the fields of statistics, quality control, and their applications, the reader is referred to Internntional Journal of Abstracts on Statistical Methods in I n d u s t r y and to Quality Control a n d A p p l i e d Statistics Abstracts. These lournals are published by the International Statistical Institute, The Hague, Xetherlands, and by Interscicnce Publishers, Inc., New Tork, respectivelv. For mathematical statistics literature, the reader is referred to the proper sections in Mathematical Rei i e u s . A very useful series of reviews has appeared in Tndztstriat and Engineering Chemistrg yearly since 1953 (107, 172-7). The last mentioned series emphacizes chemical engineering applications and should be examined to bridge the gap b e t n e m that field and chcmistrj as cowred in this ANALYTIC A L CHEMISTRY series. -1s a review on n h a t books are avail~
able for the basic library of a chemist and scientist interested in statistical applicationu, there are two articles (182, 158). Since the last review, a number of new books have appeared in the general field of statistics. A government agency statistics manual, written in 1955, has become more generally available (57). A newer edition of the popular Moroney book on the elements of statistics (148) has appeared. Statistics for engineers has bren the main theme of books by Volk (206) and by Bowker and Lieberman (19). A hook by Laccy (117) is designed as a first cqurse for t,lie man in industry who occasionally uses statistics. Saunders and Fleming (184) h a r e n-ritten an elementary book for use by pharmacists, biologists, and chemists. Books by Johnson (106) and by Tuker (240) deal with statistical exercises in analysis of variance and other calculations. For the analytical chemist, a book by Blaedel and Meloche ( I S ) has an escellent chapter on uses of ,statistics in that field. Li (123) h a written a book on statistical inference ivhich deals with many fundamental concepts in statistics. Kendall and Buckland (111) VOL. 32, NO. 5, APRIL 1960
161 R
have compiled a dictionary of statistical terms and Sreedharan Pillai (196) has a book of concise tables for statisticians. I n the field of quality control, new editions of books by .Juran (108), Schrock (1871, and Duncan (66) have appeared. The complete proceedings of a symposium on design of experiments (45)has appeared in part as a published hook (44). Cox (.54) has written a n excellent book on planning of experiments, the philosophy, and statistical aspects. In the field of nonparametric statistics, there have appeared books by Fraser (76) and Tate and Clelland (199). I n the field of multitivariate analysis, there are books by Anderson (6),Kendall (110), and Roy (161). Gumbel (66) has m i t t e n a book on the statistics of extremes and Acton ( I ) on the analysis of straight-line data. There have been a number of books on operations rcsearch, linear programming, game and decision theory. Two comprehensive bibliographies on operations research have appeared (161, 243). For coverage of current, statistical literature, the reader is referred to a number of jouinals. .irticles of interest in the applications of statistics and quality control in the physical sciences appear in lndustrial Quality’ Control, J o u r n a l of the .-1merican Statistical Association, .-I pplied Statistics, J o u r n a l of the Royal Statistical Society, Series B, Biometrics, Eiornetrika, the -4nnals of JIatheinatical S t a t i s t k s , Operations Research, Operations Research Quarterly, and the n e w r journals. Technometrics (200) and International J o u r n a l of -1bstracts-Statistical Theory and -Ifetiiod (101). Technometricc is particularly designed for the physical, rheniical. and engineering sciences. The international journal gives a complete coverage of papers in the field of statistical theory and new contributions to statistical methods published after October 1, 1958. MEETINGS, COURSES, A N D SEMINARS
The chemist and the scientist interested in iitilizing statistics in their tvork can take any one of a number of short courses. section in the periodical Industrial Quality Control appears regularly called “Clinics. Conferences, and Courses,” which lists short, summer, and formal courses as well as meetings. Summaries of short Courses in quality control and statistical methods have appeared in Industrial Quality Control (Sovember 1958, page 22) and in a revieu by Rose and others (177j. The training of engineers and of statisticians for industry has been treated in two articles (185, 166). Meeting3 of particular interest to chemists are those held by The Bmerican Society for Quality Coctrol, Chemical Division, and the annual Gordon Research Conference on Statistics in Chem162 R
e
ANALYTICAL CHEMISTRY
istry and Chemical Engineering. The Gordon confcrence, held a t a small New Hampshire boys’ school, provides the condiicive atmosphere for the presentation and frank discussion of the newest tools in statistics and their applications. The American Chemical Society occasionally has statistics symposia at its national meetings. At the meeting in September 1959, there was a symposium of the Industrial and Engineering Chemistry Division entitled “The Use of Statistics in Improving Plant Operations and Chemical Processes.” The Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, February 1960. had a symposium entitled “Statistical Control of Laboratory Operations.” The .Inierican Society for Testing Materials has an E-11 Committee on Quality Control of Materials with its subcommittee functioning to help members on statistical aspects of standards programs. .it the ASTM meeting in June of 1938. a y n p o s i u m on bulk sampling & a s hcld, nhiclh has since been publihhrd (3). Altthe ,June 1959 meeting there was a ymposium on practical and statistical significance of tests and properties of bituminous binders and a cement committee (C-I) symposium on statistical treatment of interlaboratory test rewlts. GENERAL C O M M E N T S ON STATISTICAL APPLICATIONS
The role of statistics in scientific research has been corcrcd by 1 I a j ne (248). His article relates the contribution of statistical techniques to the problcm of induction and other phases .In article of the ccientific mcthod b\ 1Iandelsohn (247) emphasizes the need of the engineer anti stntidcian to n ork together. Lurie (246) discusses the questions that the statistician asks of the scientibt. He points out that the questions, although thry nlaj seem impertinent, are, in reality, pertinent ones that the statisticinn has the responsibility of asking but not ansn ering. This article has been replied to in a letter by Crain (55). Crain does not interpret the statistician’s questions xs nyessarilj- impertinent and comments further on the relationship betmeen the statistical method antl the scientific study of problems. STATISTICAL A N D QUALITY C O N T R O L APPLICATIONS T O CHEMISTRY, ACTUAL A N D POTENTIAL
Experimental Design. Some of the major developments in statistics during the past few years h a i c been in t h e field of design of eupcriments and related subjects, in u hich applications in chemistry have become more prevalent. This revien hxs alreadv mentioned the symposium edited by Chew (44, 45) arid the nenect book by
Cox.(s/t) which deal with thc planning or design of experiments. The reader’s attention is called to the column that Youden has in Industrial a n d Engineering Chemistry on statistical design (223-2%). I n this column, he presents many of the new developments in stat’istics and design to the chemist and chemical engineer. I n one of his columns (237),he discusses systematic and nonsystematic designs; in another (224), fractional factorial schemes. Bingham, in a two-part series (16, l 7 ) , discusses various designs and the type of information yielded by each. The articles are designed to help the enginecr select the design most suitable to the problem a t hand. Zelen and Connor (-341) cover the “why‘s and n-herefore’s”of themultifactor experiments. Finkner (74) discusses various designs that can be utilized in collaborative testing betm-een laboratories. Tingey (603) explains factorial designs for chemical processing work and gives a discussion on the analysis of t’hecomponent sources of variation. Baker (9) gives a scheme for the analysis and presentation of results of a factorial espcrinient with an example In the tensile strength of a polymer. Fractional factorials are portrayed by diagram in an article on economy in planning experiments (155). K i t h regard to applications of factorial experiments in chemistry, there are a number of pertinent articles in the yarious fields. I n analytical chemistry, there are applications pertaining to wet chemical methods (32, $8,59, 70, 73, 76,.140, i s , 178, 1 ~ 6 , i 9 i , i 9 5197) , and to spectrographic methods (76, SO, 139, l S , 167! 203, 204). I n organic chemistry, there are papers dealing with melting point determinations (8, 227). cataiytic hydrogenation (119), synthesis (142), and polymerization (l68). There are application papers in physical testing (9, 12, 35> 130, 137, 2%) and a theoretical paper by Scheffb on experiments n i t h mixtures (165). There are pertinent papers in food chemistry (33, 43, 5,9,67, 92, 96, 118, 121, 127), in rubbcr chemistry (6, 29, 30: 52, 53. 201, 2’08, 21c?: 2?fc5). in paper (.%. 36, 1 6 ~ in) ~oil (igS),in ceramics ( i O 4 ) , and in metals (113, 136’). I n leather, there are a number of interesting factorial applications (5. 41, 93, 137, IS^, 14s-146, 164, 170, 180, 206) and fractional fact,orial applications (39, 147,220). With regard to other types of designs, applications of incomplete blocks appear in an interlaboratory analytical x-ray fluorescence study ( 7 ) ,an analytiea1 flame photometry study (151), a leather paper (169), two of Youden’s series pertaining to metals (225, 2%), and a radiochemical measurement paper (133). Delate and Walz (62) wrote on the theory and use of the Youden “linked-block” design in evaluating
alternative ingredients in the manufacof transferring from a laboratory to a ture of a product. Applications of latin full-scale plant production and how square designs appear in two leather EVOP offers one approach to the probpapers (91, 188) and one of Youden’s blem of determining the best conditions series on sampling of metals (228). for production yield. Koehler (114) In another of his series on a plastics has written on practical experience with experiment ( 2 2 4 , he discusses a graecothe method in the organic processing latin design. field. Whitwell (217) has related the The class of response surface designs EVOP method to chemical processing by Box and coworkers has been the problems. Box and Hunter (245) have subject of further investigations and a paper on condensed calculation techapplications in chemistry. The preniques for EVOP. vious reviews in this series (131, 132) Another statistical development gave some basic references on this which is creating considerable interest subject. The symposium edited by is the random balance experimental Chew (44,45) has papers on response design technique developed by Sattersurface theory and applications. Dunthwaite. The design is a technique for can’s book (66) has a section on rescreening a large number of variables sponse designs. An excellent series has in a limited number of samples to find appeared by Bradley (20) and Hunter the few which are the largest contribu(97-99) giving the basic ideas behind this tors to the effects under consideration. technique of the determination of opAccording to Satterthwaite, random timum operating conditions by exbalance design has had industrial apperimental methods. An article by Li plications since 1956. He presented (122) proposes a convenient worksheet the idea a t the Middle Atlantic Conferfor tabulating optimum condition data. ence of the American Society for QualCaplin (34) deals with the use of digital ity Control in 1957. Budne has an incomputers and gives some optimization teresting series of three articles (24-26) techniques for the Box-Mrilson approach. in which he gives the philosophy, techDealing particularly with the chemical niques, and some case histories of processing field, there have been good random balance. He discusses case articles on basic ideas of the response histories in the fields of plastics manudesign technique by Andersen (4, facture, product development, and anaNeuwirth and Naphtali (160), and lytical chemistry problems. I n TechSmith (191). Cull0 and Colman (58) nometrics, Satterthwaite presents the point out the value of the approach in ideas of random balance experimentadeveloping a chemical process. Bingtion (1,SL) and Budne gives some apham (16) points out, in one particular plications of designs (27). The applicasituation, the limitation of a conventions are in the field of foundry casting, tional factorial design over that of a analytical chemistry, and product imresponse design. provement. Following these papers are There have been application papers printed discussions by several wellin analytical chemistry ( I @ ) , in inorknown statisticians and responses by ganic chemical processing (50), and in the authors on the pros and cons of the organic chemistry (77, 120, 194). Two papers and the technique of random paper industry applications have apbalance design. The topic appears to peared (23, 68) and one article attempts be a controversial one a t this point, to convey to the rubber industry how as has been true for most new ideas. the two can be useful (53). It is being used in certain companies Response designs, as developed by and is being taught in short training Box and coworkers, have as their aim courses. Satterthwaite and Shainin the finding of the optimum conditions presented a paper on this type of defor a process. Box has also introduced a sign at a meeting of the Northeastern new concept called “evolutionary operaSection of the American Chemical tion” which has since become known as Society, December 1958, a t which time ET’OP. EVOP is a technique aiming they distributed a pamphlet called for optimum conditions in production, “Polyvariable Experimentation, Backand is a method in which some statisground Information.” This pamphlet tical concepts are incorporated in norhas since been published (183). Ansmal production routine to lead to evencombe (242) has written a paper on tual increased yield. The idea appears quick analysis methods for random to have been first presented by Box in balance screening experiments. Paris in July 1955 a t the Internationa! Homogeneity Studies. Closely Conference on Statistical Quality Conallied t o experimental design is t h e trol organized by the European Prostudy of the homogeneity of materials. ductivity Agency of O.E.E.C. Box Frank, Dellemand, and F r y (76), gave a paper on the subject a t the North Ottolini (167), and Farrell, Harter, Carolina Symposium (45) in 1956 and and Jacobs (69) utilize analysis of subsequently published an article (244). variance in studying the transverse One of Youden’s columns (236) brings and linear homogeneity of metal samthis technique to the attention of his ples to be used as spectrographic standreaders. He discusses the difficulties ards. Youden, in two of his columns
(226, 235), investigates the homogeneity of titanium sheet and alloy by novel incomplete block experiments. Interlaboratory Comparisons. An application in which experiment design is sometimes used is the comparison of the results of various laboratories. Finkner (74) discusses the use of various designs in collaborative testing of analytical methods. The ASTM has published interlaboratory studies on the melting point of chemicals ( 8 ) , on the x-ray fluorescence method ( 7 ) , and on the effect of ozone on rubber (2), which utilize analysis of variance (abbreviated ANOVA, for convenience). ANOVA is used in collaborative testing in articles on leather (206), food (118, l a y ) , and gas analysis
(197). Youden has developed a very interesting graphic scheme for the comparison of various laboratories on two samples. I n one of his columns (229), he points out a sign test to detect if there are constant errors among a series of collaborative laboratories. From this idea he evolved a graphic picture of some cement compressive-strength comparisons by a series of laboratories (231). The Youden scheme yields a certain amount of information that is not detectable or as readily apparent by any other method. Utilizing two materials only, information is readily conveyed on the constant biases of the various laboratories being compared and whether certain laboratories are erratic or out-of-line in their results. He continues discussing the scheme 111 a later column (232). An article in the NBS Technical News Bulletin (156) and another by Youden (238) discuss this scheme more fully. hiandel and Lashof (130) have a n article on another novel interlaboratory graphical evaluation technique, for which the background is given in another article (129). The scheme is for studying linear systematic differences between laboratories and the relation to it and to ANOVA is shown. Mandel and Lashof present a designed interlaboratory round-robin program in the physical testing of paper and give an analysis and interpretation of the results. Graphical Interpretations. Some grdphical methods for evaluating interlaboratory comparisons have just been discussed above. Also in his columns, Youden discusses another graphical technique for comparing the output of two processes (233). He shows how the graph gives much of the mformation t h a t the analysis of variance would indicate-and in a very vivid manner. I n another of his columns (234), he discusses a similar graphic technique for pointing out systematic errors in analytical methods and interactions between methods and materials involved. Brooke ( b d ) , in VOL. 32, NO. 5 , APRIL 1960
163 R
the field of analysis of petroleum catalysts, utilizes the same graphic technique in comparing operators. Quality Control. Aside from t h e books mentioned previously (66, 108, 187), there are a number of pertinent papers in the field of quality control and its related graphical technique, the control chart. Bingham (13-15) discusses quality control and the control chart in chemical processing work. Davidson (60)also discusses quality control and the control chart in chemical processing work. Kroll (115) gives the details of a quality control program in a n analytical laboratory; Nelson (153), in regard to spectrographic methods. Basic concepts of quality control are related to rubber industry workers by Thornley (201, 202) and Wescott (211, 214-21 6). James (104)has written a pertinent article in the field of ceramics and Martin (156) in metals production. Bingham (14) deals with control charts specifically in batch processes. Behnken (11) has an article on moving trend control charts in a physical testing laboratory. Specifications. Closely allied n ith quality control are t h e specifications of materials. Youden (232) writes on the need of good test procedures to determine if materials meet specification. Davies (61) discusses this in more detail in relating analytical cost to release of materials and the risks involved. Sheppayd and Ginsburg (190) have an article on the use of statistics to prepare specifications from the consumer’s and producer’s viewpoints. Smith and Waters (193) point out how =INOVA is used in product specification. Sampling. Some interesting articles have appeared on sampling in relation to quantity control and t h e ascertaining of acceptability of materials. T h e ASTM (3) has published a symposium on bulk sampling. Wescott (212) and Coutie (52) have papers on the use of sampling schemes in the rubber industry. navies (61), Smith and Waters (193), and Sheppard and Ginsburg (190) deal with optimum sampling plans. Youden (223) gives a pairedsampling plan useful in detecting vendorconsignee disagreements. Fiskell (75) compares the effects of several sampling devices on the variability of analysis. Crane (56) writes on paper sampling, and Hodus and Stubbings (93) on various designed sampling plans. Correlation, Regression, a n d Covariance. Acton’s book (1) is designed t o give t h e scientist t h e relation between regression and t h e analysis of variance. T h e symposium edited by Chew (44,45) has a paper on simple and multiple regression. B u r r o w (51) has a paper on straight-line relations, with an example from heat transfer data. The National Bureau of Standards deals 164 R
ANALYTICAL CHEMISTRY
with the fitting of straight lines to cumulative data (167). Greyson and Cheasley (87) address the petroleum man on the subject of getting better curves from data. Opfell and Sage (162) and Chou (47) write to the chemical process engineer on the background and applications of least-square methods to their work. hlarquardt (134) has a paper on least-square fitting of nonlinear models. Orths (165) conveys to the steel industry and Williams (218, 219) to the rubber industry some uses of correlation. I n the field of organic chemistry, Watterman (209) utilizes the graphic technique in conveying correlations of physical constants and organic chemical structure; Fcigen, Trapani, and Hurd (71) utilize regression, and Chinai (46) correlation. I n the paper industry, correlation is utilized in several articles (63, 64, 7 2 ) . I n two of these articles (63, 64), analysis of covariance is used to assess the significance of certain relations. I n two papers previously mentioned (1I , l 4 ) , regression is used in trend and correlation studies in the chcmicnl processing field. I n ana1ytic:d chemistry, there have been a numhcr of papers utilizing the correlation technique (32, 48, 85,85, 105,125, 189) with several in the spectrographic field of analysis (86, 100, 198). Least squares regression technique is iiscd in a physical testing paper (124) and in one involving radiochemical measurements (133). In the leather industry, there have been a number of pertinent papers on correla-, tion (37, 90, 137, 179, 180) with one (90) utilizing analysis of covariance. Miscellaneous Statistical Topics. Multivariate analysis, related t o repression, is a statistical tool of potential interest t o the chemist. I t is used when differences are being compared on two or more related variables. Pertinent books have already been mentioned (6, 210, 181). There has been a n article by Jackson (103) in which such techniques have been used in color-matching. Another statistical tool which has been utilized in a number of articles is that of multicomparison and multidecision testing. The book “Experimental Design: Theory and Application,” by W. T. Federer, SIacmillan, New York, 1955, mentioned in the previous review in this series, has a large section on this subject. The symposium edited by Chew (45) has a paper on selecting the best of several normal populations. Youden has discussed Beckhofer’s multiple decision procedure for the best of k means in one of his series (230). There have been applications of Duncan’s multiple range tests in the grouping of means in analytical papers in the organic (178), spectrographic (76), food (79), and agricultural (48) Qj
fieMs and one in the paper processing field (68). Extreme value theory is another tool of potential interest to the chemist, particularly with regard to physical testing. The book by Gumbel (88) has been previously mentioned. Mandel (1%) has brought the subject to the attention of ASTM in their bulletin. Related to regression are some interesting graphical techniques. Joint confidence ellipses are utilized in the correlation between slope and intercept with respect to the uncertainty of linear calibration curves (125). Contour ellipse areas are applied in density scatter diagrams in spectrographic analysis (198). Instrumental drift techniques of Youden are used in flame photometry analysis (151). There has been a utilization of Bartlett’s test for homogeneity of variances with respect to physical measurements (171). Youden has nritten on the uses of random numbers and random normal variates in one of his columns (226). Aids in Statistical Calculations. A fesv interestinq nomographs have app e n r d nhich are of use t o thosc involved in makine statistical conclusions. Nelson (159) has designed one giving the t-value a t a glance, for any degree of freedom. British Chemical Engineering (22) has designed one for comparing two variances and another for comparing two means. A few papers have appeared on the use of computers in statistical calculations. Muller (150) writes of the role of computers in such calculations and gives references to some codes available. Orchard-Hays (163) writes on applications in design and operation research and Caplin (34)on optimization and linear programming techniques. Opfell and Sage (162) show how digital computing equipment is used in chemical processing and engineering problems. I n analytical chemistry, Kienitz (122) points out applications in mass and ultraviolet absorption spectrometry and Hopp and Wertzler (95) in mass spectrometry analysis. REVIEW OF PAPERS IN BIBLIOGRAPHY BY FIELDS OF CHEMICAL INTEREST
A review is given here b y scientific field of specialization and by industry. This should be of interest for the specialist and for the reader interested in frequency of occurrence of papers in particular fields, keeping in mind the limitations of coverage expressed earlier. This bibliography lists some 68 papers in chemical and spectrographic analysis. Papers designed for training purposes are listed (11, 18, 22, 61, 65, 73, 74,81-84,95,102,105,112,116, 139, 162, 153, 166, 158, 191, 195, 198, 223, 229, 231 , 232, 234, 238, 239) as well as applications (7,12, 32, 33, 38, 48, 49, 59, 64, 67, 69, 70, 76, 79, 80, 85, 86, 92, 100,
( 8 ) A S T M Bull., NO. 231, 92-6 (1958). Melting range of rubber chemicals. (9) Baker, A. G., A p p l . Statistics 6, 45-55 I n t h e field of chemical processing and (1957). Analysis and presentation of engineering some 26 papers are listed, of results of factorial experunents. (10) Bearman, R. J., J. C h m . Phys. 29, which certain ones are for training pur1278-86 (1958). Statistical mechanical poses (4,6,14,16,31,34, 47,58,60,114, theory of thermal conductivity of 122, lS4, 160, 162, 186, 20s) 217, 233, binary liquid solutions. 236), and others in applications (50,62, (11) Behnken, D. W., Znd. Quality Control 14, 7-11 (Nov. 1957). Moving trend 77, 119, 168, 194, 209). I n physical control charts. testing, there are some 11papers listed(12) Bicking, C. A., ,Tappi 42, 89Ar91A for training purposes (10s)1SO) and the (Jan. 1959). Precision data available balance in applications (1.2,88, 36, 37, on commonly used TAPPI standard methods of test. 72, 124, 137, 169, 171). In organic (13) Bingham, R. S., Jr., I n d . Quality chemistry, there are some 10 papers Control 13, 46-58 (May 1957). Praclisted-ne for training purposes (227) tical chemical process control. and the others in applications (8, 46, (14) I b i d : , 13, 21-6 (June 1957). Control charts in multistage batch processes. 71,77,119,120, 142, 168, 194). There (15) Ibid., 15, 14-18 (Sept. 1958). Guide are application papers in chemical physto use of statistics in chemical industry. ics ( I O , 109, 141, 207, 208)) in radio(16) Zbid., 15,29-34 (May 1959). Design and in bacteriology chemistry (116,1SS), of experiments from statistical view-
118,125-127,136,140,149,161,164,165, 167, 178, 180, 186, 197, 204, 206, 260).
(49,861. 80,91,9S,94,1S7,1S8, 143-147,164,169, 170,179, 180,206,220). I n the rubber industry, there is one training series (29,SO, 61-53, 201, 202, 218,219) and another by Wescott (211-216). I n the paper industry, there is one training paper (189) and several application papers (2SJS6, 36,43,66,6S,64,68,72). There are training papers in the metals industry (11s)136, 164, 165, 226, 228, 236), in the petroleum industry (2.2, 73, 87, 192, 19S, 221,222), and application papers in the field of food (SS, 59, 79, 92,96,118, 121). I n agriculture, there is a training paper (74)and several application papers (48,67,75,127, 197). Finally, there is a training paper in ceramics (104)and one application of a new concept in a plastics problem (224). ACKNOWLEDGMENT
The author wishes to acknowledge gratefully the help of one of the previous authors of this review series, Frederic J. Linnig, in providing background, advice, and references of much value in the present review. Thanks are also due to the other past authors of this review for their suggestions and information.’ LITERATURE CITED (1) Acton, F,., S., “Analysis of Straightc
Line Data, Wiley, New York, 1959. (2) Am. SOC. Testing Materials, Philadelphia, Pa., Spec. Tech. Publ. 229 (1958). Symposium on effect of ozone on rubber.
242 (1958). Symposium on - bulk sampling.. (4) Andersen, S. L., Cham. Eng. Progr. 55, 61-7 (A ril 1959). Statistics in strategy of ctemical experimentation. (5) Anderson, H., J. SOC.Leather Trades’ C + m i s k 41, 2-7 (1957). Influence of acids and salts in sole leather tannage. Pt. 111, Acids and salta in early tan liquors. (6) Anderson, T. W., “Introduction to Multivariate Statistical Analysis,” Wiley, New York, 1958. (7) A p p l . Spectroscopy 13, 3-7 (1959). Re& on first x-rav fluorescence roundrobin test. (3) Ibid.,
nnint,. Pt,. - -. T-. (17) Zbid., 15, 12-15 (June 1959). Pt. 11. (18) Blaedel, W. *J., Meloche. V. W., c - - - - - 1
Listing by industry gives 24 papers in applications in leather (6, 37, S9-42,
“Elementary Quantitative Analysis: Theory and Practice,” Chap. 5, Row, Peterson, and Co , Evanston, Ill., 1957. (19) Bowker, A. B., Lieberman, G. J., “Engineering Statistics,’’ Prentice-Hall, Englewood Cliffs, N. J , 1959. (20) Bradley, R. M., I n d . Qualzty Control 14, 16-20 (July 1958). Determination of optimum operating conditions by experimental methods. Pt. I, Mathematics and statistics fundamental to fitting of response surfaces. (21) Brit. Chem. Eng. 3,96 (1958). Nomoof gram, statistical tests. comDarison . . 5wo ’variances, comI;arisori of two means. (22) Brooke, M., Petrol. Refiner 38,139-42 (July 1959). How statistics imwove lab efficiency. Brown, D. S.. S., Turner. Turner, W. R.. R., Smit,h. Smith, (23) Brown. A. C., Jr., Tappi T a p p i 41, 295-300’(1958): 295-300 (1958). Sealing strength of wax-polyethylene blends. (24) Bidne, T. A , , I n d . Quality Control 15, 5-10 (April 1959). Random balance. Pt. I, Missing statistical link in fact finding techni ue. (25) Ibid., 15, 11-16 ?May 1959). Pt. 11, Techniques of analysis. (26) Ibid., 15, 16-19 (June 1959). Pt. 111, Case histories. (27) Budne, T. A., Technometrics 1 , 139-93 (1959). Application of random balance designs. (28) Buist, J. M., Rubber J. 132, 714-16 (1957). Polymer testing and its contribution to developments in industry. (29) Buist, J. M., Rubber J. and Intern. Plastzcs 134, 979-83 (1958). Statistics a8 applied in rubber industry. Comarison of variations within and between rabs (30) Zbid., 135, 5-7 (1958). (31) Burrows, G. L., I n d . Quality Control 15, 15-18 (Oct. 1958). Interpreting straight-line relationships. (32) Butters, B., Chenery, E. M., A d y s t 84, 239-45 (1959). Rapid method for determination of total sulfur ia s o h and plants. (33) Calhoun, W. K., Bechtel, W. G., Bradley, W. B., Cereal C h a . 35, 350-9 (1958). Vitamin content of wheat, flour, and bread. (34) Caplin, D. A., T r a n s . Inst. C h a . Engrs. (London)36, 311-20 (1958). Use of digital computers in planning plant operations. (35) Capott, R. J , Newman, S. B., Mandel, J., T a p p i 42, 480-5 (1959). Paper test data from pendulum and inertialess testers.
(36) Carlson, J. A., Zbid., 42, 196-202 (1959). Bacterial deterioration of woolen felts. (37) Carter, T. J., J. Am: Leather Chemisk’ Assoc. 53, 250-8 (1958). Impact test for leather based on alternating punching force. (38) Chalmers, R. A., Page, E. S., Gwchim. et Cosmochim. Acta 11, 247-51 (1957) (in English). Reporting of chemical analvses of silicate rock. (39) chambers, M. F., Dem sey, M., Mittow, R. G., Nattrass, E. J. SOC. Leather Trades’ Chemisk 41, 7-13 (1957). Vegetable tannage of sheep skins: systematic study of influence of various factors in tanning process. I, Plan of the experimental and details of
b.,
tannamn _ _.. (40) Chlmbers, M. F., Mittow, R. G., Nattrass. E. F., Ibid., 41, 48-59 (1957).
Chemical analysis, manual assessnients, and darkening of leathers on exposure to light. (41) Ibid., 41, 6l?-9 J1957). Influence of duration of pickling and methods of chrome tanning- on -properties of chrome side leather. (42) Zbid.. 41. 94-105 (1957). Veeetable ‘ tannage 02 sheepskins: ’ sysGmatic study of influence of various factors in tanning process. 111, Physical properties of leathers. (43) ChGne, M., Sangeris, S., Ayroud, A. M.. Assoc. tech. i n d . waweti2re. Bull. 4, ‘119-24 (1957) (in ‘ Frenchj. A p plications of statistical analysis of variance to straw soda cooks. (44) Chew, V., ed .,,,“Experimental Designs in Industry, Wiley, New York, ~~
1958. (45) Chew, V., ed., “Proceedings of Sym-
posium on Design of Industrial Experiments, Nov. 5-9, 1956,” Institute of Statistics of the Consolidated University of North Carolina. (46) Chinai, S. N., Znd. Eng. C h m . 49, 303-4 (1957). Calculation of molecular weight and dimensions of polymers from viscosity interaction parameter. (47) Chou, Chan-Hui, Ibid., 50, 799-802 (1958). Least squares. (48) Clark, K. G., Yee, J. Y., Lundstrom, F. O., Lamont, T. G., J . Assoc. Ofic. A g r . Chemists 42, 592-7 (1959). Modified activity index procedure for dekrmining quality of water-insoluble nitrogen in mixed fertilizers containing urea-formaldehyde compounds. (49) Clarke, M. F., ANAL. CHEM. 31, 1243-6 (1959). Comparison of two species of lactobacilli as test organisms in pantothenic acid estimation. (50) Comstock, A. J., Jurnack, S. J., Mooney, R. W., I n d . Eng. Chem. 51, 325-8 (1959). Precipitation of calcium hydrogen orthophosphate. Four-variable statistical design. (51) Coutie, 0. A., Rubber .I. and Intern. Plustics 134, 328-32 (1958). Statistics as applied in rubber industry. I, Need ‘for statistical experimentation. (52) Zbid., 134, 439-44 (1958). 11, Stac tistical tests and their application. (53) Zbid., 134, 510, 599-600, 608 (1958). 111, Design of experiments. (54) Cox, D. R., “Planning of Experiments,’’ Wiley, New York, 1958. (55) Crain, C. N., Am. Scientist 46, 212A220A (1958). Some comments on statistical method. (56) Crane, F. G., Jr., Tappi 42, 5-9 (1959). Sampling in lieu of sheet sorb ing. (57) Crow, E. L., Davis, F-A., Maxfield, M. W., U. S. Naval Ordnance Test Station, China Lake, Cal., “Statistics Manual,” NAVORD Rept. 3369, NOTs 948 (1955). (58) Cullo, L. A., Colman, W. P., Z n d . VOL. 32, NO. 5, APRIL 1960
165 R
Eng. Chem. 50, 1529-30 (1958). Development of a process, case history. (59) Daum, W. F., DeVries, J. E., Miller. B. S.. Cereal Chem. 36, 176-85 (1959): Polarographic determination of cystine in wheat and flour hydrolyzates. (60) Davidson, H., Chem. Eng. Progr. 53, 395-6 (1957). Statistical quality control in chemical plant. (61) Davies, 0. L., Technometrics 1, 49-61 (1959). Some statistical aspects of economics of analytical testing (62) Delate, E. J., Walz, R. N., Znd. ET,^. Chem. 50, 309-12. (1958). Evaluating alternative ingredients in manufacture of a product. (63) Dickens, W. A,, Plummer, A. W., Tappi 40, 895-9 (1957). Correlation of equilibrium data for S02-H20-Ca0 system. (64) Dillingham, E. O., Knuth, D. T., Wessman, G. E., Zbid., 41, 321-33 (1958). Manometric method for rapid, Dractical determination of biochemical bxygen demand. (65) Doerffel, K., Chem. Tech. (Berlin) 10 (3), 151-2 (1958) (in German). Error control in routine analysis. (66) Duncan, A. J., “Quality Control and Applied Statistics,” rev. ed., R. D. Irwin, Homewood, Ill., 1959. (67) El Gindy, M. M., Lamb, C. A., Burrell, R. C., Cereal Chem. 34, 185-95 (1957). Influence of variety, fertilizer treatment, and soil on protein content and mineral composition of wheat, flour, and flour fractions. (68) Ernst, A. J., Clark, T. F., Finkner, M. D., Dutton, H. J., Tappi42,23543 (1959). Critical study of short-period caustic pulping of wheat straw. (69) Farrell, R. F., Harter, G. J., Jacobs, R. M., AXAL.CHEM.31, 1550-4 (1959). Determination of trace impurities in zirconium metal by point-to-planes spectrochemical procedure. (70) Fauth, M. I., Stalcup, H., l b i d . , 30, 1670-2 (1958). Evaluation of six methods for determination of nitrogen in nitroguanidine. (71) Feigen, G. A., Trapani, I . L., Hurd, M. S., J . Assoc. O$lc. Agr. Chemists 41, 693-703 (1958). Physicochemical properties of glycerol pectate. (72) Fetsko, J. hI:, l’appi 42, 110-21 (19591. Evaluation of smoothness and ibsorbency tests for white patentr coated paperboard. (73) Field, E. A,, J . Inst. Petrol.. 43, 233-4 (1957). Measuring- the precision . of test‘methods. (74) Finkner, hl. D., J . Assoc. O&. Agr. Chemists 40, 882-92 (1957). R e liability of collaborative testing for A.O.A.C. methods. (75) Fiekell, J. G. A., Kelley, W. H., Myers, J. M., J . Assoc. O$lc. Agr. Chemists 41, 640-9 (1958). Variability in individual core analysis of samples of a 10-0-10 and 8-0-8 fertilizer drawn by several sampling devices. (76) Frank, R. C., Dallemand, J. F., Fry, D. L., Spectrochim. Acta 9, 323-31 (1957). Statistical study of homogeneity of zinc-base spectrographic - . itandards. (77) Franklin, S. L., Pinchbeck, P. H., Popper, F., Trans. Inst: Chem. Engrs. (London)36,259-69 (1958). Statistical approach to catalyst development. Pt. 11, Integration of process and catalyst variables in vapor phase oxidation of naphthalene. (78) Fraser, D. A. S., “Sonparametric Methods in Statistics,” Tf‘iley, New York, 1957. (791 Gordon. J.. Noble. I.. Food Research ‘ 24, 6-13 ’(1959). Suitability of 2 , P dinitrophenyl hydrazine methods for determining dihydro- and total ascorbic
166 R
ANALYTICAL CHEMISTRY
acid in raw and cooked vegetables. 11, Total ascnrhic acid. .~. ....(SO) Goris, P., D u f f ~ ,W. E., Tingey, F. H., ANAL.CHEM.29, 1590-2 (1957). Uranium determination by isotope dilution technique. (81) Gottschalk, G., Dehmel, P., 2. anal. Chqz. .lB,,81-94, (1957) (in German). Statistics in chemical analysis. I, Discovery of large systematic errors and importance of chance errors applied to gravimetric and volumetric analyses. (82) Ibid., 160, 161-9 (1958) (in German). 11, Evaluation of practical analyses. (83) Ibid., 163,273-81 (1958)(inGerman). 111, Testing of results of photometric calibration measurements for LambertBeer law. (84) Zbid., 163,330-8 (1958) (in German). IV. Evaluation of Dhotometric standard measurements and fulfillment of BeerLambert law. (85) Graff, Y., Bull. S O C . chim. France 1957, 436-40. Calculation of correlation in analytical results. (86) Greenstreet, J . E. S., Norris, K. P., Spectrochim. Acta 9, 177-98 (1957). Existence of differences between infrared absorption spectra for bacteria. (87) Greyson, &I., Cheasley, J., Petrol. Refiner 38, 135-42 (Aug. 1959). Get better curves from your data. (88) Gumbel, E. J., “Statistics of Extremes,” Columbia University Press, New York, 1958. (89) Hader, R. J., Youden, W. J., ANAL. CHEY.24, 120-4 (1952). Experimental statistics. (90) Hawk, R.,, Lollar, R. M., J . Am. Leather Chemzsts’ Assoc. 52, 361-95 (1957). Some notes upon hide and leather quality interrelationships. (91) Ibid., 53, 149-61 (1958). Influence of temperature upon uptake of salt by cattlehide during brining. (92) Hepburn, F. N., Lewis, E. W., Jr., Elvenhjem, C. A., Cereal Chem. 34. 312-22 (1957). Amino acid content oE wheat, flour, and bread. (93) Hodus, H. J., ptubbings, R., J . Am. Leather Chemzsts Assoc. 52. 411-28 (1957). Laboratory method fo; measuring effects of process variables on leather quality. I, Selection of samples for laboratory experiments. (94) Ibid., 52, 429-49 (1957). 11, Laboratorv-scale exDeriments and assessment i f leather duality. (95) Hopp, H. F., Kertzler, R., ANAL. CHEX 30, 877-9 (1958). Machine computation of mass spectrometer analyses-triangular inverse method. (96) Horn, M. J., Fifield, C. C., Blum, A. E., Warren, H. W., Cereal Chem. 35, 411-21 (1958). Distribution of amino acids in wheat and certain wheat products. (97) Hunter, J . S., Ind. Quality Control 15, 16-24 (Dec. 1958). Determination of optimum operating conditions by experimental methods. Pt. 11-1, Models and methods. (98) Ibid., 15, 7-15 (Jan. 1959). Pt. 11-2, Models and methods. (99) Ibid., 15, 6-14 (Feb. 1959). Pt. 11-3, Models and methods. (100) Hurwitz, J. K., Spectrochim. Acta 12, 211-23 (1958). Spectrographic ana$& of segregates with dynamic microvolume technique. 11, Experimental verification of theory. (101) International Journal of AbstractsStatistical Theory and Method, International Statistical Institute, Oliver and Boyd, Edinburgh, England. First issue dated July 1959. (102) Ionova, K. I., Nalimov, V. V., Zavodskaya Lab. 23, 586-91 (1957). Magnitude of errors in spectrographic analysis of steels. ~
~
’
(103) Jackson, J. E., J. Opt. SOC.Am. 49, 585-92 (1959). Some multivariate statistical techniques used in colormatching data. (104) James, G. D., Am. Ceram. SOC.Bull. 36, 310-12 (1957). Some applications of statistics in ceramic field. (105) Johnson, H. V., Stross, F. H., ANAL. CHEM. 31, 1206-11 (1959). Gas and liquid elution chromatography. Quantitative detector evaluations. (106) Johnson, N. L., compiler, “Statistical Exercises,” Pt. 11, Dept. of Stat., Univ. College, London, 1957. (107) Johnson, R. C., Burggrabe, W. F., Jr., Heiny, R. L., Russell, J. L., Sweeny, R. F., Ind. Eng. Chem. 51, 422-31 (1959). (108) Juran, J. M., “Quality Control Handbook,” McGraw-Hill, New York, 1957. (109) Keck, J. C., J. Chem. Phys. 29, 410-15 (1958). Statistical theory of chemical reaction rates. (110) Kendall, M. G., “Course in hlultivariate Analysis,” . Charles Griffin, London, 1957. (111) Kendall, M. G., Buckland, W. R., “Dictionary of Statistical Terms,” Hafner, New York, 1957. (112) Kienitz, H., Z. anal. Chem. 164, 80-99 (1958) (in German). Evaluations, errors, and criticisms of errors in physical analysis. (113) Knotek, M., Janko, J., Huntick4 Zisty 12, 423-30 (1957) ( i n . Czech). Application of analysls of variance for evaluating experimental results. (114) Koehler, T.. L., Tappi 42, 261-4 (1959). Evolutionary operation: Its method and operation. (115) Kroll, F. W.,Jr., Ind. Quality Control 14, 9-13 (Dec. 1957). Effective quality control program for industrial control laboratory. (116) Kuyper, -4.C., J..Chem. Educ. 36, 128-32 (1959). Statistics of radioactivity measurement. (117) Lacey, 0. L., “Statistical Methods in Experimentation,” hlacmillan, Kew York, 1957. (118) Lamb, F. C., J . Assoc. O$c. Agr. Chemists 42, 278-85 (1959). Rept. on total solids in concentrated tomato products. (119) Lange, H. B., Chem. Eng. Progr. 53, 304-7 (1957). Investigating chemical plant process variables. (120) Lazor, J. T., J . Appl. Polymer Sci. 1, 11-16 (1959). Study. of particle initiation in vinyl chloride emulsion using response surface experimentation. (121) Lee, J. W.,Geddes, W. F.,,Cereal Chem. 36, 1-18 (1959). Studies on Brew process of bread manufacture: Effect of sugar and other notrients on baking quality and yeast properties. (122) Li, C. H., Chem. Eng. 65, 151-6 (April 7, 1959). Worksheet gives optimum conditions. (123) Li, J. C. R., “Intrpduction to Statistical Interference,’ Edwards Brothers, Ann Arbor, Nich., 1957. (124) Loebenstein, W. V., J . Research Natl. Bur. Standards 60. 105-8 (1958). Precise evaluation of &face area with indirectly calculated dead space. (125) Loscalzo, .4. G., Benedetti-Pichler, A. A., ANAL.CHEM.30, 2018-19 (1958). Methods of De Saint Venant and hlohr and use of statistical criteria. (126) Mack, M., Spielbery, K.,Spectrochim. Acta 12, 169-78 (1958). Statistical factors in x-ray intensity measurements. (127) McFarren, E. F., J . Assoc. O&. Agr. Chemists 42, 263-71 (1959). Rept. on collriborative studies of bioassay for paralytic shellfish poison. (128) Mandel, J., ASTM Bull., No. 236,
29-30 (1959). Theory of extreme values. (129) Mandel, J., Technometrics 1, 251-67 (1959). Measuring process. (130) Pvlandel, J., Lashof, T. W., A S T M Bull.. No. 239, 53-61 (1959). Interlaboratory evaluation . of testing methods. (131) Mandel, J., Linnig, F. J., AN.&. CHEM. 28, 770-7 (1956). Statistical methods in chemistry. (132) Ibid., 30, 739-47 (1958). (13?) Mann, W. B., Stockmann, L. L., louden. W.J.. Schwabel. A.. Mullen, P. A,, ’Garhkel, S. B.,’ J . ’Research Natl. B u r . Standards 62,. 21-6 (1959). Preparation of new solution standards of radium. (134) Marquardt, D. W.,Chenl. Eng. Progr. 5 5 , 65-70 (June 1959). Solutjon of nonlinear chemical engineerlng models. (135) hlarshall, C. E., I n d . Quality Control 14, 28-32 (May 1958). Teaching of statistics and training of statisticians for industral employment. (136) hIartin, A. R., J . Znst. Metals 85, 209-26 (1957). Role of statistical methods in controlling quality of nonferrous castings. (137) Mattei, V. R., Williams, T., J . Am. Leather Chemists’ Assoc., 52, 110-23 (1957). Physical properties of leather fatiiquored a t different oil levels. (138) Ibzd., 54, 12-24 (1959). Cse of fish oils for fatliquoring leather. I, Menhaden oil and cod oil fatliquore. (139) Maurice, hI. J., 2. anal. Chem. 158, 271-9 (1957) (in German). Cse of statistical methods in analytical chemistry. (140) Maurice, M. J., Veen, R., Ibid., 163, 13-21 (1958) (in German). Contribution to spectrophotometric determinativn of methanols with chromotropic acid. Example of application of variance arialysis to solution of analytical chrmistry problems. (141, l l a z o , R. M., J . Chem. P h y s . 29, 111’2-8 (1958). Statistical mechanical t h r o r y o i solutions. hleiners, A. F.: Bolze, C., Scherer, L., IIorriss, F.i-., J . Org. Chem. 23, 1 122-5 (1956 j. Application of statistical dcsign to organic synthesis. Reductive alkylation of tert-butylamine (Leuvk:irt reaction!. (i43) Mittow, R. G., J . Soc. Leather 7‘rade.s’ Chemiafs 41, 424-30 (1957). lIigratioii of salts in leather. (144) hlittvw. R. C . , Ponkhurst, K . G. A . , lbid.? 41, 128-55 (1957). Study of mechanism of fatty spue formation, I. Ibid., 41, 161-83 (1957). 11. AIittow, R. ti., Pomeroy, C. D., ., 41, 109-14 (1957). Damage,to pelt c:i.used kjy overheating during drying. (147; Ibid., 41, 402-17 (1957). Wettability of chrome-tanned leather. (148) hIoroiiey, hI., ‘,Factsfrom Figures,” 3rd and revised ed., Penguin Books, Baltimore, 1957. (149) Morris. -1.G . C.: Bozalek, S. J., d 7 ~ u l . Chim. Acta 21, 2i5-?1 (1959). lktermination of si.:!face in super~
existing computer techniques in statistics and qualitv control. ( I 51 j llulligan, B: FV., Haught, A. F., J . Research S a t l . B u r . Standards 61, 49!!-501 (1958). Correction for inbtrumental drift in flame photometry. (1.52) Saliniov: V. I*.,J . A n a l . Chem. C y fiS.8S.K. 12, 155-62 (1957) (English triimlation:. Use of analysis of vari:mces for eva1uat ion of analytical result5.
(153) Nalimov, F’. V., Nedler, V. V., Zbid., 13, 433-9 (1958). Evaluation of results of semiquantitative spectrographic analysis with aid of Poisson’s distribution. (154) Nathan, C. C., Eisner, E., Corrosion 14, 193-9t (1958). Statistical concepts in testing of corrosion inhibitors. (155) Xatl. Bur. of Standards, Znd. Quality Control 14, 5-6 (Jan. 1958). Economy in the planning of experiments. (156) Natl. B u r . Standards, Tech. News Bull. 43, 16-18 (1959). Graphical diagnosis of interlaboratory tests. (157) Ibid., 43, 143 (1959). Fitting straight lines to cumulative data. (158) Nelson, B. N., A p p l . Spectroscopy 11, 123-7 (1957). Statistical evaluation of spectrographic methods. (159) Xelson, L. S., Znd. Quality Control 14, 26-7 (Oct. 1957). Nomograph of Student’s t . (160) Neuwirth, S. I., Naphtali, L. M., Chetn. E n g . 64, 238-46 (June 1957). S e w statistical method rapidly determines optimum process condition. (161) Operations Research Group, Case Institute of Technology, “A Comprehensive Bibliography on Operations Research,’’ Publications in Operations Research No. .4, Wiley, New York, 1958. (162) Opfell, J. B., Sage, H. H., I n d . Ens. Chem. 50, 803-6 (1958). Applications of least squares methctls. (163) Orchard-Ham, W., Ibid.,50,16546 (1958). Computers as tools for synthesis, experimentation, and informstion handling. (164) Ornes, C. L., J . Am. Leather Chmntists’ Assoc. 54,452-66 (1959). Studies of fatliauorinn. I , Develonment and evaluatidn of liborat,ory methods. (165) Orths, B., ,Stahl u. Eisen 78, No. 1, 14-21 (1958): .\EC-Tr-3225 (translation). Contriaution to use of mathematical statistics in steel industry. (166) Ostle, B., J . E n g . Educ. 47, 410-14 (1957). Statistics in enginewing. (167) Ottolini, A . C., ASAL. CHEM.31, 447-9 (1959). Porous cup technique in determination of magnesium in cast iron, statistical study. (168) Pasteelnick, L. A , Leder, W. R., Chenz. E n a . Prour. 53, 392-5 (1957). Statistical- anal:,& i i i ’ polymerizatioh process. (1869) Pofneroy, C’. D., J . SOC.Leather Trades Chemists 41, 106-8 (1957). Influence of small quantities of oil on tensile properties of vegetable-tanned spiit hides. ( l i 0 ) Ibid., 42, (301 (1958). Stitch tear tests on Iight leather. (1’71) Roberts, A . G., Pizer, R.S., A S T M Bull., No. 221, 53-88 (1957). I’rotective coating adhesion measurement using eiectronic aver:iging tievice for adherorneter. (172) Rose. &I., Heiny, R . L.. Johnson, K. C., Schilk, .I. A , , Znd. Eng. Chem. ‘46, 916-22 (1954). Computer*, statistics. and mat,hematics. (1’73) Rose, rl., Jtihnson, R. C., Heiny, R. I,.. Ibzd., 47, 626-32 (1955). Computers, statistic:, and mathematics. (1’Y-l) Hose, A. Johnson, It. C.,’ Heiny, R . L., \Viiliams, T. J., Ibid., 48, 622-32 :195S). Computers,. mathematics, statistics, and automation. (175) Ibid., 49, 55l-61 (1957). (176) Rose, A., Schilk, J. A , , Johnson, .R. C., Ibid., 45, 933--40 (1953). Computers, statistics, and mathematics. (197) Rose, -4., Svreeny, It. F., Johnson, .R. C., Burggra‘-Je, tY. F., Heiny, R. L., Ibid., 50, 512-19 (1958). Computers, mathematics. 3116 statistics. (178) R.oth, &I.,iVegnian, II., ”Quality Control and Statistical Methods,” 2nd ed., Reinhold, New York, 1957. 88) Seligsberger, L., Sadlier, C., J. ;1m. Leather Chemists’ A 4 s ~ o c52, . 2-16 (185i). Xew developments ;ii tanning -sith aldehydes. 89) Sheidon, F. R., I n d . Labs. 8 , 4-10 (Fet). 1957). Taking the guesswork out of R/L)-Statist,ics. 90) Sheppard, H. It., Ginshurg, H., AST.Zl B u ! ~ . ,S O . 236, 31--4 (1959). Should statist.ical methods IK used to prepare materials specificntions. (191) Shewell, C. T., ASAL. (?HEM. 31, S o . 5,211b27.4 (1959). Statistic.s--IYSY t,ool for 1959 analytical probLrins. (192) Smith, H. F., Pctrcri. fi‘r-tiner 37, 201-8 (March 1958). Phn ! d o r e you start experimenting. (193) Smith, H., Jr., Waters, T. F., .l. i l m . Oil Chemists’ Soc. 35, 246-9 f.1958). Certitin uses of analysis of variance with standard product specscations. (194) Smolin, E. >I., Beegle, I,. C., I d . Enri. :‘hem. 50, 1115-18 !1958). Con+irlLoti~high pressure syntbeeis of 3amir?cproprioxiitrile. (195) dpecker, H., Jackworth, E., Hartkami), ?I,.! Z ,qnni. Cherr,. 158, 165-82 (1958: i i l i (xermanj, Comparative studies cn eu:ic~tness of photometric and chemical analytical procedures. (196) Sreedharari Pillai, K. (2.; “Concise t Tables for SLqtisticiens,” Y:ie S t a t i s tical Center, Univ. of the ?*fanila, 1957. (197) Steyermark, A,, J . Assoc. Ogic. Agr. Chemists 42, 319-25 (1959). Rept. on niicruandytical determination of oxygen. Pt. 11. (1 98) Strasheim, A., Xeddy, It. J., l i p p l . Spectroscopy 12, No. 2, 29-32 (1958). hlathematical method of comparing spectrochemical results. (199) Tate, &I..K., Clelland, R. C., “Nonparametric and Shortcut Statistics,” Interstate Printers and Publishers, Danville, Ill., 1957. (200) Technoinetrics, Anitxricar, Society for Quality Control i n d American Statistical Association, Issued i n 1959. (201: Thornley, E. H . , Ruhber J . and VOL. 32, NO. 5 , APRIL 1960
* 167 R
Intern. Plastics 134, 827-9 (1958). Statistics as applied in rubber industry. Application of statistics in rubber factory. (202) Ibid., 134, 903-4, 906 (1958). (203) Tingev, F. H., I n d . Eng. Chem. 50, 1017-20 (1958). Identification of estimation of variation in process measurements. (204) Tingle, \V. H., Matocha, C. I4-78.4 (Dec. 1958). Circumstances alter cases. (234) Ibid., 51, 81.1-82A (Fell. 1959). What is a measurement. (235) Zbid., 51, 81.4-824 (April 1959). I
,
Sampling study-Complex
processing
of alloys poses special sampling prob-
lem. (236) Ibid., 51, 79.4-80.4 (June 1959). Evolutionary operation. (237) Zbid., 51, 65‘4-66A (-4iig. 1959). Experiments in experimentation. (238) Touden, W. J., Znd. Qualitu Control 14, 24-8 (5Iav 1959). Graphical diagnosis of interinborntorv test reSllltl.
(239) Touden, JV. J., Condor, JV. S., Severo, S. C., Technoinetrzcs 1, 101-9 (1959). Measurements made by niatching with knon-n standards. (240) Tuker, H. E., “ A Guide to Statistical Calculations,” Putnam’s Sons, Kew York, 1958. (241) Zelen, hl., Connor, JV. S., Znd. Quality Control 15, 14-17 (llarch 1959). Multifactor experiments. Additional References (242) .4nscombe, F. J., Technometrics 1, 195-209 (1959). Quick analysis methods for random balance screening experiments. (243) Batchelor, J. H., “Operations Research-.innotated Bibliography,’J 2nd ed., St. Louis Univ. Press, St. Louis, hlo., 1959. (244) Box, G. E. P., i l p p l . Statistics 6, 81-101 (1957). Evolutionary operation, Vethod for increasing - industrial productivity. (245) Box, G. E. P., Hunter, J. S., Technometrics 1, 77-95 (1959). Condensed calculations for evolutionary operation programs. (246) Lurie, W.,A m . Scientist 46, 57-61 (March 1958). Impertinent questioner: Scientists guide to statistician’s mind. (247) hlandelsohn, J., I n d . Quality Control 13, 31-4 (hlay 1957). Relation between engineer and statistician. (248) hlayne, J. W.,Scz. Monthly 84, 26-33 (Jan. 1957). Role of statistics in scientific research.
Review of Fundamental Developments in Analysis
Differential Thermal Analysis C. B. Murphy Generol Electric Co., Schenecfady,
T
N. Y.
review covers the period from the last review on differential thermal analysis (45)until the end of August 1959. It is again impossible to include all published works in this field. Accordingly, many worth-while papers are not included. It is intended t h a t significant trends be covered. Smothers and Chiang (70) have published a book on the theory and practice of differential thermal analysis. Mackenzie has published the “Differential Thermal Investigation of Clays” (49). A well documented review has been written by West et al. (79). Application of the technique in mineralogy and geological prospecting has HIS
168 R
ANALYTICAL CHEMISTRY
been the subject of a review by FoldvariVogl (21). Papers b y Sauman (67) and Lehman (41) discuss experiences with differential thermal analysis encountered in their respective laboratories. INSTRUMENTATION
Several new concepts in differential thermal analysis instrumentation have occurred. Although equipment capable of producing thermograms in the range -150” t o 300” C. has been reported (%’), other types have been developed for low temperature work of the more conventional type. Haighton
and Hannewijk (27) have employed a n electrically heated copper block, coppercylinder sample holders, and ironconstantan thermocouples mounted in a Dewar flask for excellent results below 100” C. Hill and Murphy (SO) have devised a simple, inexpensive system employing an infrared lamp as a heating element with conventional detection equipment for low temperature organic studies, Pakulak and Leonard ( 5 3 ) have employed matched thermistors for measuring and controlling temperature in the 20’ t o 300’ C. range. Lloyd and Murray (42) have automatic equipment for high vacuum (10-6 mm.) heated by a molybdenum
