the variance of the Automated I1 system to be significantly less than that of the reference method. The data from samples analyzed with the Automated I and Automated I1 systems are plotted in Figure 5. The slope of 0.97 f 0.02 is not significantly different from 1.0 and the correlation coefficient is 0.998. Duplicate analyses with the Automated I system yielded a standard deviation of 0.023 pg SOz/l. (N = 17, mean concentration = 0.50 pg SOz/ml) which is significantly greater than that of the Automated I1 procedure. This precision is, however, better than the reference method. The results from the Automated I1 system were on the average 0.012 pg SOz/ml (2 pg/m3) higher than the reference method results. The Automated I system results were on the average 0.006 pg SOJml (1 pg/m3) higher than the Automated I1 system results or 0.018 pg SOz/ml (3 pg/m3) higher than the reference method. These biases are not significant when considering the variations in duplicate analysis by each method. References ~utomrtrdII system [ut S O ~ / m e t d j
Figure 5. Plot of comparison data for the Automated II and Automated I methods Equation of the line: y = 0.972 X
+ 0.013
Each method contains sufficient formaldehyde and pararosaniline and gives linear plots of absorbance vs. concentration for samples in the range of 0-1 pg SOz/ml (0-167 1g/m3). Any samples higher than 1 pglml are routinely diluted with absorbing reagent. Figure 4 shows the data plot from the reference method and the Automated I1 procedure comparison. The slope was calculated to be 0.96 f 0.03, which is not significantly different from 1.0 (t-test a = 0.05); the correlation coefficient was 0.992. The analysis of duplicate samples yielded standard deviations of 0.037 pug SOz/ml ( N = 7, mean concentration = 0.26 pg SOz/ml) and 0.013 pg SOz/ml (N = 16, mean concentration = 0.42 pg SOz/ml) for the reference and Automated I1 methods respectively. An F-test showed
(1) Fed. Regist., 36,22385 (1971). (2) “Users Manual: SAROAD (Storage And Retrieval of Areometric Data)”, U.S. Environmental Protection Agency, Office of Air Programs, No. ADTD-0663, November 1971. (3) West, P. W., Gaeke, G. A., Anal. Chem., 28,1816 (1956). (4) “Determination of Sulfur Dioxide: Automated Para-Rosaniline Method As Used by AQALB, 1972” from Analytical Quality Assurance Laboratory Branch, 1972 (Environmental Protection Agency). (5) “Industrial Method No. 169-72AP”, Technicon Instruments Corp., Tarrytown, N.Y., 1973. (6) Scaringelli, F. P., Elfers, L., Norris, D., Hochheiser, S., Anal. Chem., 42,1818 (1970). ( 7 ) Scaringelli, F. P., Saltzman, B. E., Frey, S. A,, ibid., 39, 170918 (1967). (8) Cedergren, A., Wikby, A., Bergner, K., ibid., 47,100-6 (1975). (9) Morgan, G. B., Golden, G., Tabor, E k . in “Automation in Analvtical Chemistrv. Technicon Svmuosia “ . 1966”. Mediad. Inc.. Tarrytown, N.Y., i967. (10) O’Keefee, A. E., Ortman, G. C., Anal. Chem., 38,760 (1966).
Received for review May 5, 1975. Accepted October 1, 1975. Mention of commercial products is for identification only and does not constitute endorsement by the Environmental Protection Agency of the U.S. Government.
Correction B. J. Dowty, D. R. Carlisle, and J. L. Laseter point out an error in their paper, “New Orleans Drinking Water Sources Tested by Gas Chromatography-Mass Spectrometry. Occurrence and Origin of Aromatics and Halogenated Aliphatic Hydrocarbons” [Environ. Sci. Technol., 9 (8), 762-5 (1975)) “Closer examination of the spectrum of compound number 58, reported in Table I as (2- or 1-Napthy1)dichloromethane, reveals this compound to be dichloroiodomethane. Although both of these compounds share common spectral features such as their isotope ratios of the molecular ion and the presence of the -CHC12 fragment, comparison of the spectrum of our compound with that of the recently synthesized dichloroiodomethane reveals our initial interpretation to be in error. Spectrum of the dichloroiodomethane was furnished by Robert Kleopfer, USEPA, Region VII.” 1174
Environmental Science 8 Technology