662 NICHOLAS AND POLLAK: THE ISOLATION OF THE LINES [Vol. 76 The Isolation of the Lines of the Mercury Arc by Filters With Especial Reference to Photo-electric Absorptiometrj- BY J. W. NICHOLAS . ~ N D F. I;. POLLAK SYNoPsIs-The light filters normally used or recommended for use in England with the mercury vapour lamp photo-electric absorptiometer do not isolate monochromatic light, with the exception of the 577/9 nip. lines. An improved series is described, which give monochromatic light in the more frequently used regions of thc mercury spectrum, and an approxima tion to it in some regions which ha1.e not hitherto been used, i.e., infra-red, red, orange and blue-green. THE use of filter photometers in place of spectrophotometers for absorptiometric determina- tions has certain advantages such as lower cost, greater speed and greater suitability for routine use by relatively unskilled operators, but normally there are several disadvantages.Filter photometers are not suitable for measurements of more than one unknown in the same coloured solution and no specific extinctions can. be determined and published for use by other workers on other instruments. Sensitivity and accuracy are also reduced by the use of mixed light, and apparent deviations from Beer’s law occur when dealing with substances having sharp absorption bands (cf. Miillerl) . States and Anderson2 find that with a substance obeying Beer’s law a curved calibration graph is to be expected when there is stray light, with the exception of the special case where the solution has equal densities for the stray light and the desired wavelength.Lothian3 has shown that when a finite waveband is used instead of monochromatic light the effective wavelength varies with the absorption curve of the solution being examined, and that the measured density bears no easily predictable relationship to the true density. Some of these difficulties can be obviated by the use of a mercury vapour lamp as light source, and this was available in the Zeiss Pulfrich photometer (Heilmeyefl) with filters said to give isolation of the 436, 546 and 577 mp. 1ines.l Vaughan5 used the mercurv arc in conjunction with the Hilger Spekker photo-electric absorptiometer. Mullerl points out that with a mercury arc and a highly selective filter, a filter photometer need not give results any less accurate than a spectrophotometer.In applying a photo-electric absorptiorneter to spectrophotometric procedures we found certain discrepancies which led us to believe that, with the exception of Ilford No. 606 for the 577/579 mp. lines, truly monochromatic light was not obtained with the filters normally advised or supplied for the purpose. It has been suggested (e.g., Stross6) that interference filters might have an application in absorptiometry because of their greater transmission. It should be noted, however, that although their transmission maxima are sharp, the positions of these maxima are subjectDec., 19501 OF THE MERCURY ARC BY FILTERS 663 to a tolerance of A10 mp., both between different filters and between different places on the same filter (Miiller7).This tolerance makes it impossible for them to be used for spectro- photometric procedures with a continuous source, and may well lead to considerable difficulties with a line source. At their present stage of development it would appear that they present no advantages for the 365, 436, 546 and 577 mp. lines. S t a a t ~ ~ , ~ has described the design of filters for the isolation of all the main lines of the mercury arc. She employed an exclusively mathematical treatment based on published data for light source, filters and receivers. Her figures take no account of background spectrum immediately adjacent to the required line, and the region 532 to 5614mp. is considered as identical with the 546mp. line, and 561-5 to 595 mp.with the 577/9 mp. lines. She also fails to mention the 3906.4 A. line, which would be passed by both her 365 and 405mp. filters. Unfortunately, Corning and Jena glasses only are used, and neither of these is available to the English analyst. In the course of his development of absorptiometric methods for metallurgical analysis, Vaughan5 has also examined a small number of filters spectrographically with the mercury arc. He gives a plate in which only Ilford 606 and Wratten 74 appear to transmit an adequate quantity of monochromatic light, but in the text gives Wratten 62 and Ilford 604 (omitted from the plate) as suitable alternatives to Wratten 74, and also suggests the use of Ilford 601 in the violet region, although his own plate shows this to be far from mono- chromatic.He mentions a suggested use of Wratten 36 + Wratten 2A for the isolation of the 436 mp. line, but does not record any experiments with this combination. Unfor- tunately his plate does not include the red end of the spectrum, although he states that red and orange filters can also be used. He gives no details of spectrographic or photographic technique, but it appears from the absence of background and the suppression of the line at 4960-3 A. that his spectra are under exposed and that his filters have therefore been tested too leniently. PRESENT PRACTICE FOR THE ISOLATION OF LINES For convenience, the groups of closely adjacent lines that would not normally be resolved by a monochromator are considered as single lines and given a single nominal wavelength.Thus we shall refer to the 3650, 3654, 3662 and 3663 A. group as the 366 rnp. line. 365 mp.- Messrs. Hilger and Watts Ltd.l0J1J2 and Haywood and Wood13 recommend Wood's glass (supplied by Hilger as H556) without any supplementary filter. This glass is apparently identical with Chance 0 x 1 , as is Ilford 828; the material appears to vary considerably from melt to melt. Our spectrogram, Fig. l b , shows that it is not monochromatic in the ultra- violet when used alone, and in addition it transmits red and infra-red. 405 mp.- Hilger and also Haywood and Wood advise Chance 8 (OV1) +- Wratten 2 for isolating the 406 mp. line. Haywood and Wood describe OV1 as a heat absorbing filter, which it certainly is not, having a transmission a t 750 mp. of 46 per cent.14 Gentry and Sherrington,l51ls by using this combination, obtained curved calibration graphs for tungsten and therefore substituted Ilford 601 for Wratten 2.Although this combination is a great improvement, we found appreciable transmission of the 365, 391 and 436mp. lines. 436 mp.- For 436 mp., Hilger and also Haywood and Wood recommend Chance 6 (OB2) and IYratten 50. Hadleyl' used Calorex (ON3) and Wratten 50, but found fluctuations due to heat effects on the gelatin filter. For this reason he preferred Calorex and Chance OBI." As an alternative he suggests modifying the Spekker so that the right-hand filter is on the right-hand side of the drum. This difficulty has been overcome in the new model H760 of the Spekker (cf. Isbellla). Rogers,lg Harrison,20 de Lippa21 and Lennardn used Ilford 601, as did Davis,Z3 who obtained a curved calibration graph with the silicomolybdate colour (cf.Fig. 11). 546 mp.- Calorex and Ilford 605, as advised by Hilger and by Haywood and Wood, were used by Kogers,lg British Iron and Steel Research AssociationM and Parker.% Calorex and Ilford 604, recommended by Vaughan, have been used by Edwards and Gailer,26 Edwards and Robinson2? * Manufacture now discontinued-Chance Bros. (personal communication)664 NICHOLAS AND POLLAK: THE ISOLATION OF THE LINES KEY TO SPECTOGRAMS Fig. 1. 366mp. (a) Unfiltered light source (control) , . ,. .. . . .. .. (b) Wood's glass, 2 mm. (Hilger, H556) . . .. .. .. .. .. combination) . . . . . . . . .. . . .. .. . . (c) Chance OB2, 1 mm. + Wratten 17 4- Chance 0x1, 3 mm.(recommended Fig. 2. 405mp. (a) Control . . ,. .. .. .. . . . . . . .. . . (b) Chance OV1, 2mm. + Wratten 2 . . . . .. .. .. . . (c) Chance OV1, 2 mm. + Ilford 601 . . .. .. .. .. . . (d) Chance OV1, 4 mm. + Wratten 86C + Chance OB2, 1 mm. (the recom- mended combination (see Table 11) gives an identical spectrogram, but is preferred because of lower red and infra-red transmission) . . Pig. 3. 436mp. (a) Control . . . . .. .. . . . . . . .. . . (b) Chance OB2, 2mm. + Wratten 50 . . . . . . . . . . (c) Chance ON13, 2mm. + Ilford 601 . . .. . . .. .. (recommended combination) . . .. . . . . .. . . (d) Chance OB2, 2 mm. + Wratten 36 -t Wratten 2A (two thicknesses) (e) Chance OB2, 2mm. + Wratten 2A (three thicknesses) + Wratten Fig. 4. 492mp. (a) Control .. . . . . . . . . . . . . . . . . (6) Ilford 603 + 804 . . .. .. .. .. . . . . . . (c) Ilford 603 + Wratten 5 + Ilford 302 + Ilford 804 (recommended com- bination) . . .. . . . . . . . . . . . . .. . . .. .. .. .. .. 36 .. .. .. Fig. 5. 546mp. (a) Control . . a . . . .. . . .. . . . . . . . . (b) Chance ON13, 2mm. + Ilford 605 . . .. . . .. . . .. (c) Chance ON13 + Ilford 604 . . .. .. .. .. .. . . (d) Wratten 16 + 74 + Ilford 504 (recommended combination) . . . . [Vol. 75 Relative" exposure, sec. 1 3 24 1 11 48 252 1 19 8 21 28 16 216 684 8 256 1420 260 * Exposures are calculated for a slit width of 0.04mm., although for the longer exposures a slit of 0.08mm. and half the exposure time was actually used. It should be noted that with intense lines necessitating a control exposure of 1 sec.there was some suppression of the background and the 4960-3 A. line. Spectrograms have also been taken with ten times the exposure and they do not show any foreign line.Fig. 1 Fig. 2 Fig. 3 Fig. 4 Fig. 5 Figs. 1-5. Spectrograms showing the isolating properties of various filter combinations (see Table 11, p. 660, and Key on p. 664)Fig. 6 Fig. 7 Fig. 8 Fig. 9 Fig. 10 Figs. 6-10. Spectrograms showing the isolating properties of various filter combinations (see Table 11, p. 669, and Key on p. 665)Dec., 19503 OF THE MERCURY ARC BY FILTERS 666 KEY TO SPECTOGRAMS-CO~~~~~C~~. Fig. 8. 577 mp. (a) Control . . . . .. . . . . . . . . . . (b) Chance ON13, 2mm. + Ilford 606 . . . . . . . . (c) Ilford 812 -+ 804 (recommended combination) .. . . Fig. 7. 607mp. (a) Control . . . . .. .. . . ,. . . . . (b) Chance 0x13, 2 mm. + Ilford 607 . . .. . . . . (c) Ilford 202 + Wratten 25 + 804 . . . . . . . . (a) Wratten 26 +- Ilford 804 (recommended combination) . . (e) Wratten 25 $. Ilford 801 (“orange-red” combination) . . Fig. 8. 691 mp. (u) Control . . .. .. .. .. .. .. .. . . . . . . . . . . . . . . . . .. . . .. .. .. .. .. . . .. .. ( b ) Chance ON13, 2mm. f Ilford 206 + Chance ON13, 2.6mm. (recom- mended combination) . . . . .. .. .. .. ,. .. Fig. 9. 735 + mp. .. .. .. .. . . .. . . .. (a) Control . . .. (b) Wratten 88A (recommended filter) . . .. . . . . . . .. Fig. 10. Uniformity of filters. Spectrograms of a number of finished filters intended to isolate the 365 mp. Control exposure, 1 sec.Relative exposure for filters, 25 to 28sec. line. Relative* exposure, sec. 8 182 270 30 288 1494 1614 66 30 514 30 48 * Exposures are calculated for a slit width of 0.04 mm., although for the longer exposures a slit of 0.08mm. and half the exposure time was actually used. I t should be noted that with intense lines necessitating a control exposure of 1 sec. there was some suppression of the background and the 4960.3 -4. line. Spectrograms have also been taken with ten times the exposure and they do not show any foreign line.666 and Lennard.22 Harrison28 abandoned 604 in favour of 605 because the latter was not sensitive to a yellow interfering colour. This suggests that some light of shorter wavelength than desired was passed by 604 (cj. Fig. 5). The possibility of using Wratten 74 has been mentioned by Vaughan and by Haywood and Wood.Fig. 5b shows the considerable yellow transmission of Ilford 605, and Fig. 5c shows the high transmission of blue-green background by Ilford 604. NICHOLAS AND POLLAK: THE ISOLATION OF THE LINES [Vol. 75 SILICON, ?:, Fig. 11. Calibration curves for the estimation of silicon in aluminium as silicomolybdate. Curve A, with traditional combination of Chance 6 and Wratten 50; curve R, with proposed 436 mp. combination. Final concentrations were : aluminium (containing 0 to 14-4 per cent. of silicon added as sodium silicate), 0.32 g. per litre; nitric acid, 0.128 X; ammonium molybdate, 10 g, per litre 557 mp.- For this h e , workers have been almost unanimous in using Ilford 606, usually with Calorex as recommended by Hilger, by Vaughan and by Haywood and Wood, e.g., Harrison,20 British Aluminium C O . , ~ ~ Bairstow, Francis and Wyatt." Most workers report straight line graphs. It is difficult to see the advantage of using such a combination (see Fig. 74. Payneal has used Ilford 607 with Calorex. GENERAL CONSIDER4TIONS It has been suggested that glass filters are more satisfactory than gelatin. The usual cause of failure of filters appears to be heat by conduction and convection, and with careful use the effect of radiation is small. Filters are not exposed to heat in the H760 Spekker absorptiometer, and we assume that any user of the H560 instrument will have modified it as recommended by Hadley.1' Some gelatin filters are markedly fluorescent, and if placed immediately in front of the photo-cell large errors occur.When placed in front of the solution under test the error will be very small, but we have nevertheless tried to eliminate the effects of fluorescence (8.g. , by absorbing the exciting or fluorescent radiation, or both, by non-fluorescent filters) and the filters recommended will give consistent results irrespective of their distance from the photo-cell. Fluorescent elements are marked ( f ) in Table 11. It has been usual to place a heat-absorbing filter between the lamp house and the gelatin filter to protect the latter from radiant heat. Since a filter can onl57 be heated by radiation being absorbed, the dyed gelatin films will scarcely be heated at all by the infra-red, although they will be heated by the strong visible lines.We have therefore allowed other considerations to determine the position of the heat absorber, and our filter combinations are intended for use in one direction only. The components in Table I1 are listed in the order in which light passes through them.Dec., 19501 OF THE MERCURY ARC BY FILTERS 667 EXPERIMENTAL METHODS Probable filter combinations were selected from the publications of Messrs. Kodak,32 IlfordS and Chancel4 to ensure that filters commercially available in England were used, The following filters were examined in various combinations- Wratten: 2, 2A, 4, 5, 8, 9, 16, 17, 21, 32, 25, 26, 29, 32A, 35, 36, 44, 44A, 45, 50, 62, 70. Ilford: 201, 202, 206, 302, 601, 603, 602, 606, 606, 607, 608, 801, 802, 803, 804, 805, 806, Chance: OB2, OB10, ON12, ON13, OX1 and OV1.Densities of individual filters were measured with a double-cell photo-electric absorptio- meter that will be described elsewhere.34 The filter combinations used on this instrument were improved as we went along, and for the final measurements were those we recommend later. The exposure required for any filter combination was determined from the sum of the densities of its elements at the required line, the aim bGing to make the photographic density the same as in a spectrum of the unfiltered light source in which the region of the required line showed considerable background. Spectrograms were taken of the mercury arc (Siemens lamp MB/D, 125 watt, as used in the Spekker) in the Hilger medium quartz spectrograph E498.The light source at a distance of 38 cm. from the slit was focussed on the collimating lens by the Hilger quartz condensing lens F1026. Slit widths of 0.04 and 0-08 mm. were used. A two-step sputtered metal filter (Hilger F1219), having densities at 4Fio mp. of 0.0 and 0.98, was placed over the slit. Spectrograms were taken on Kodak 111-L plates, which have a fairly uniform sensitisation extending as far as 900 mp. Development was in Kodak D 19b developer, undiluted. Messrs. Kodak’s statement32 that “all dyed gelatine film filters” transmit freely in the infra-red” indicated that this region required full investigation. In view of the difference between the spectral sensitivity of the plate and photo-cell, and in view also of the relatively wide range over which an integrated value was required, measurements were made on the absorptiometer already mentioned rather than on the plate.The absorptiometer was used as a single-cell instrument to measure the fraction of the light transmitted by the filter that would pass through Ilford 608, this value being corrected for the density of Ilford 608 itself. This method takes into accdunt the spectral response curve of the photo-cell and the result is referred to as the efective red and infra-red transmission. The effect of this transmission can best be shown by considering a hypothetical extreme case of a substance with a density of 0.801 (15.812 per cent. transmission) at the required n-avelength and a density of zero in the red and infra-red. Let us assume that a density reading of 04300 is actually obtained, corresponding to a transmission of 15.849 per cent.Let the effective percentage of red and infra-red light be R and let the intensity of the required line on the photo-cell at the zero setting be I ; then at the zero setting the red and infra-red intensity will be IR/100, and the total intensity will be I ( 1 1 R/100). During the reading the 73, 74, 75, 86C, 87, 88, 88A and 89. 807, 808, 809 and 812. intensity of the line on the photo-cell will be m9 I , and the intensity of red light will be ( 15*81 ”> ;’ RI :,. m). Assthe totals at the zero setting and during the reading will be equal, and therefore R = 0.044 per cent. We have therefore tried to reduce R below this level by the incorporation of one or other of the following heat-absorbing elements: Chance 0x13, OB2, Ilford 801, 802, 803 or 804.In the H76O model of the Spekker absorptiometer there is a built-in heat absorbing filter. This will normally be unnecessary, and in the case of the 691 mp. and infra-red filters will interfere. As an example of the effect of this stray radiation we show in Table I the density readings of a 1 in 100 dilution of blood as oxyhaemoglobin with various filter combinations intended * Ilford Ltd. manufacture a series of gelatin filters absorbing in the near infra-red (Sos. 801 to 804). These contain a copper salt, not an organic dye.668 NICHOLAS -4ND POLLAK: THE ISOLATION OF THE LINES [Vol. 75 to isolate the 577 mp. line. low red and infra-red absorption. The solution has a sharp absorption band near 577 mp.and very TABLE I EFFECT OF VARIOUS FILTER COMBINATIONS INTENDED TO ISOLATE THE 577 ~ p . LINE Optical density Alone . . .. . . + Chance ON13, 2mm. -+ Chance OB2, 2mm. + Chance OB2, 4mm. + Ilford 801 . . . . + Ilford 802 . . . . + Ilford 803 . . . . + Ilford 804 . . . . . . . . , * . . . . . . . . . . . . . . . . . . Ilford 812 0.950 0.960 0.980 0.980 0.959 0.971 0.972 0.980 Ilford 868 0.727 0.763 0-917 0.964 0.785 0.865 0.884 0.937 A water-cell only absorbs radiation of wavelength greater than 1 . 4 ~ . and its effect on the efective infra-red transmission was found to be negligible. It was also found that a thermometer placed in a position corresponding to the left-hand photo-cell of the Spekker absorptiometer showed no significant rise in temperature.Moreover, Messrs. Evans Electro- selenium Ltd. (the makers) state that infra-red is not harmful to their photo-cells. The left- hand water-cell is placed very near the lamphouse and its rise in temperature owing to conduction and convection causes the formation of air-bubbles with consequent drift in the zero setting. We have therefore abandoned the use of this cell, and where difference methods are employed the right-hand water-cell has also been dispensed with. In the estimation of the sensitivity of the Spekker absorptiometer with our recommended filters, the left-hand aperture was completely closed by a shutter and the drum was opened to give a full-scale deflection (about 0.5 p-amp.) on the Cambridge spot galvanometer at maximum sensitivity.Following Vaughan’s nomenclature this gives the air-to-air setting that should give full sensitivity. RESULTS- In general, we used one basic filter to absorb as many foreign lines as possible, and one or more supplementary filters to obtain complete isolation. In addition, heat and red- absorbing filters were incorporated. The Cambridge spot galvanometer normally used with the Spekker absorptiometer has a sensitivity of 170mm. per p-amp., and will not give a satisfactory response with some combinations. For these the substitution of a more sensitive galvanometer, e.g., Tinsley type VS6/45, is recommended. In view of the importance of the red region, which has not previously been used, an alternative filter combination of higher transmission but wider waveband is given.In all the red combinations the lines are much less intense in relation to the background and the filters do not give such strictly mono- chromatic light as the others. Following the practice of many workers with the tungsten lamp we shall refer to each region by its dominant wavelength, which in both cases is a line (607 and 691 mp). There is no commercially available filter that will cut off the infra-red on the long wave side, so it is necessary to rely on the falling off in sensitivity of the photo-cell with increasing wavelength. Unfortunately different photo-cells vary in their response to infra-red, but no doubt the makers would be able to select cells having the desired characteristics. It should be noted that the response in the infra-red is more sluggish than in the visible and ultra-violet regions, but not enough to interfere unduly with the convenience of taking readings.PREPARATION 01; FILTERS- For convenience and to reduce the density by the avoidance of surface reflections we made compound filters containing all the gelatin films between glass plates, which were clear glass or glass filters according to requirements. Filters are usually cemented with Canada Balsam.32 The yellowing of this with age may interfere with the transmission of wavelengths below 492 mp., and it is possible that other media such as damrnar or plastics may be preferable here (see Nicholas and Pollak=). I t is not essential for colourless glass t o be optically worked, as the For some wavelengths as many as five constituent filters were used.Dec., 19501 OF THE MERCURY ARC BY FILTERS 669 light already has to traverse the optically imperfect silica and glass walls of the lamp, and we have found the glass of photographic plates perfectly satisfactory.Glass filters must be plane-parallel to avoid the introduction of density gradients. TABLE I1 RECOMMENDED FILTER COMBINATIONS I\’ave- Recommended Individual length, combination densities* mb. 365 405 43 6 492 546 577 “607” “891” 735 + ‘ ‘Orange red” OB2, lmm. Wratten 17 (f) 0x1, 3mm. OV1, 4mm. Wratten 86Ct OB10, 1-43 mm. OB2, 2mm. Wratten 2A (treble thick- ness) (f) Wratten 36 Ilford 603 Wratten 5 Ilford 302 Ilford 804 Wratten 16 Wratten 74 Ilford 804 Ilford 812 Ilford 804 Wratten 26 (f) Ilford 804 ON13, 2 m n ~ Ilford 206 ON13, 2.5mm.Wratten 88A Wratten 25 (f) Ilford 801 0.627 0.607 0.221) 1.180 0.902 0.242 0.369 0.354 0-852 1.030 0.343 0.110 0.122 0-170 1 -085 0.239 0.967 0.627 0.290 1.119 0.333 0.365 0.499 Effective red and Over-all infra-red density transmis- (cemented) sion, O/ / O 1.411 < 0.03 2-272 0.2 1.410 <0*03 1.496 0.1 1-411 .:0*03 1.527 <0.03 1.374 __ 1.239 0.197 - 0.357 .- Air-to-air setting on Spekker for full deflection 0.985 1.28 1.52 1.30 1.46 Remarks Requires sensitive galvano- Combination amended meter November, 1950; spectro- gram of this new combina- tion is identical with Fig. 2d. See Fig. 11 Requires sensitive galvano- meter Satisfactory results also ob- tained with Ilford 606, 812 or Wratten 73 with any heat absorber Includes 607 to 623 mp. lines.Requires sensitive galvano- meter Transmits 672 to 709 mp. Requires sensitive galvano- meter Requires sensitive galvano- meter Density measured at 607 mp. Not monochromatic ; in- tended for use when only Cambridge spot galvano- meter available XoTEs-Filters marked (f) are fluorescent. Filters not marked Wratten or Ilford are made by Chance Bros. * The density of many glass filters differs from melt to melt and the thickness is not therefore a complete We have therefore given the density of elements in proved satisfactory combinations in specification. addition. t Messrs. Kodak have recently changed the composition of Wratten 86C without changing its designa- It is therefore probable that recent batches will not give Suitable batches should correspond with the published absorption curve,32 This combination is recorded in the hope that the original tion or giving any other indication to the user. results identical with ours.and have a density at 405 nip. of about 0.9. filter will again be made available.6iO 1. 3 . 4. * L . r ) . ti. 7. 8. 9. 10. 11. 12’. 13. 14. 15. 16. 17. 1s. l!). 20. 21. 23. 2s. 23. 2’6. 27. 28. 29. 30. 31. 32. 33. 34. 35. .) 8 -“. SILVER NICHOLAS AND POLLAK KEFEKENCES [Vol. 75 hliiller, It. H., f l i t ? . Iztig. Chein., .-lizal. Ed., 1941, 13, 667. States, 31. S., and Anderson, J . C., J . Opt. SOC. Amer., 1942, 32, 659. Lothian, G. F., “Absorption Spcctropliotometry,” Hilger and \Vatts Ltd., London, 1949. Heilmcyer, Ludwig, “Medizijzischc .S~ektl.ophotonietrie; aiasgewtihlte Methoden und neure r l ~ t t t ’ ~ , - szrc~~u~tgse~,ueb?~isse am Korperfavbstofjee?r ztnd Iiorpevfliissi~:keiiE,I,“ C.Fischer, Jena, 1933 ; EngliGh translation by Jordan and Tippell, Hilger, I-ondon, 1943. Vaughan, I3. J., “Further Advances in the Use of the Spekker Photo-electric Absorptiometer in Rletallurgical halysis,” Institute of Chemistry, London, 1942. Stross, \Y., Aiiulyst, 1949, 74, 622. Miiller, 1s. H., I n d . Eng. Chenz., Anal. Ed., 1946, 18, So. 11, p. 21 of advertising section. Staats, E. M., J . Opt. SOC. Auier., 1938, 28, 112. -, Ibid., 193‘3, 29, 221. Hilgcr and IVatts Ltd., “The Hilger Photo-electric Absorptiometer,” London, 1947. .~ , “Instructions for Use of Absorptiometer,” London, 1944. .- , “Hilger Spekker Xbsorptiometer Type H.760,” London, 1949. Haywood, F. IV., and iVood, A. A. It., “Metallurgical Analysis by Means of the Spekker Photo- Chancc Rros. Ltd., “Chance Coloured Glass for Scientific l’urposes,” Smethwick, not dateci. Gentry, C. H. It., and Sherrington, I,. G., .41ztclyst, l94ri, 71, 432. -- __ , Ibid., 1948, 73, 57. Hadiey, \V. H., Ibid., 1945, 70, 43. Isbell, R. A. L., Ibid., 1949, 74, 618. Rogers, B., ,’l.letallui~gia, 1945, 33, 13. Harrison, T. S., .4nal*yst, 1945, 70, 362. de Lippa, 31. Z., Ibid., 1946, 71, 34. Lcnnard, G. J., Ibid., 1949, 74, 253. Davis, 13. C., “The Photometric Analysis of Steel and Aluminium Alloys,” Royal Aircraft Estab- British Iron and Steel Research Association, Il.letnlLzcrgin, 1948, 39, 105. Parker, C. ;I., -4wdyst, 1949, 74, 112. Edwards, F. H., and Gailer, J . W., Ibid., 194.5, 70, 365. Edwards, F. H., and Robinson, A. hl., Ibid., 1946, 71, 379. Harrison, T. S., J . SOC. Cheiti. Iizd., 1944, 63, 347. British Aluniinium Co. Ltd., “Chemical Analysis of :Iluminium and its Alloys,” London, 1947. Hairstow, S., Francis, J . , and Wyatt, G. H., Analyst, 1947, 72, 340 Payne, S. T., “Colorimetric and Polarographic ’Inalysis of Some Non-Ferrous Metals and Alloys, ‘’ Eastman ICodak Co., “\Vratten Light Filters,” Rochester, N.Y., 1945. Ilford Ltd., “Ilford Colour Filters,” London, not dated. Nicholas, J. W., Metallurgin, in the Press. Nicholas, J . W., and Pollak, F. F., J . Sci. Ins&., 1951, 28, 23. electric Xbsorptiometer,” Hilger, London, 1944. lishment, Farnborough, Report No. M.7880, not dated. British Non-Ferrous Metals Research Association, Report No. 729, London, 1947. I h D 1;irst submitted, Febvuarv, 1950 1f-1 r H AM , ESSE s Amended, Jz&; 1950