July, 1917
T H E J O U R X A L OF IiVDUSTRIAL A N D ENGINEERING CHEMISTRY
AMERICAN ZINC, LEAD SMELTING CO. METHOD (COWteSy Of D r . John Johnston)-Weigh 0 . j gram into a flask of :zoo cc. capacity. Add I O cc. hydrochloric acid and boil nearly tci dryness. Add Overheating or baking j cc. nitric acid and take to dryness. should be avoided. The best time t o remove t h e flask from the hot plate is when it is stained red all over. Care should be taken t o avoid spurting. Cool, and add I O cc. hydrochloric acid and boil to half its volume. Add 50 cc. hot water and I O cc. ammonia. If manganese is present add 2 0 cc. saturated solution of bromine water. Boil and filter through I I cm. filter, which should leave about one-third of the funnel exposed, into a beaker 375 cc. capacity. m’ash three times with hot water. Redissolve precipitate into original flask with hot dilute hydrochloric acid, I part acid to 3 parts water. Wash well with hot water. Add I O cc. ammonia as before, and the same amount of bromine water, if manganese is present. Boil and filter through the same paper into the same beaker. Wash three times with hot water. Make the filtrate neutral with hydrochloric acid, using litmus paper as indicator, and add 6 cc. hydrochloric acid in excess. If copper is present add 2 0 g. test lead and boil until all copper is thrown down. Heat to i o o C. and titrate with a standard solution of potassium ferrocyanide. Run the solution in rather slowly and stir constantly, A slight color change will be noted in the beaker when the precipitation is almost complete. This should not be ignored, since if i t does not occur there is a possibility of error. Use a I per cent solution of ammonium molybdate for outside indicator.
GEOPHYSICAL LABORATORY, \VASHIKGTOS.D. C.
A NEW METHOD OF SEPARATING ZINC FROM CADMIUM
AND THE LATTER’S DETERMINATION IODOMETRICALLY1 By ERICJ O H N ERICSOS
Further research in t h e development cf t h e writer’s method for complete spelter analysis has demonstrated t h e possibility of separating t h e bulk of t h e remaining zinc from cadmium b y crystal1ii:ation as zinc sulfate. T h e exact composition of t h e latter has not yet been ascertained. I n one analysis of t h e crystals only ~ j j o. per cent was found, while t h e formula Z n S 0 4 . 7 H 2 0calls for 2 2 . 7 3 per cent. rllthough a small trace of cadmium is entrained in t h e zinc sulfate or zinc-ammonium sulfate, for technical purposes only one crystallization is deemed necessary, in view of t h e large sample ( 1 9 . z grams) of spelter taken originally. The procedure in spelter analysis is as follows: Referring t o earlier publications2 for ?.etails for removing a n d determining lead, t h e filtrate from t h e latter is boiled until nearly neutral and a white precipit a t e appears. Then add jo cc. dilute sulfuric acid ( I : 3 ) , boil down. t o about 80 t o IOO cc. volume and allow t o stand over night. I n t h e morning t h e bulk of t h e zinc will be found crystallized out as sulfate. Decant t h e clear solution and wash three times with cold water, allowing each washing t o drain. Dilute t h e filtrate t o zoo cc‘. and pass in hydrogen sulfide until all cadmium is precipitated: usually I j t o 2 0 minutes is sufficient with a fairly rapid evolution of gas. Allo.iv t h e precipitated cadmium sulfide several hours time t o settle before filtering off. Determine 1 Presented a t the S y m p o s i u m on t h e Chemistry and d.letallurgy of Z i n c , 54th Meeting American Chemical Society. Kansas City. April 10 t o 14, 1917. Eighth I n t e r n . Connr. AoP1. 5 ~. Chem.. 1. 183. and ‘Tms TOURNAL. (1913’1, 401. ~
67 1
cadmium b y a n y of t h e methods mentioned in the above-mentioned papers. Cadmium may also be determined iodometrically according t o von Berg’s‘ method (modified), by transferring t o a n Erlenmeyer flask, adding about 1 2 5 cc. of distilled water, a measured excess of N / I Oiodine solution and then 30 t o jo cc. dilute hydrochloric acid. Shake and titrate with sodium hyposulfite until slight iodine excess is indicated; t h e n add a few cc. starch solution and finish titration until disappearance of the blue color. T h e difference is due t o cadmium: I cc. N,’Io iodine solution = 0 . 0 0 5 6 2 gram cadmium. T h e new zinc-cadmium separation has been applied successfully t o zinc ores also. Before its introduction i t was a difficult matter t o detect and determine accurately t h e small amounts of cadmium usually occurring in these ores. PROCEDURE-DiSSOlve j grams ore in nitric or hydrochloric acid, according t o the nature of t h e ore, fume off with 2 0 cc. sulfuric acid, add water, boil and filter. T o t h e filtrate add a n excess of ammonia, boil and precipitate iron and alumina. Dilute t o 500 cc., filter and pipette off an aliquot portion representing 3 or 4 grams; evaporate t o l o r bulk a n d until small white precipitate appears; then add sulfuric acid and boil down t o 80-100 cc. Remove from hot plate and allow zinc t o crystallize out. Decant solution and precipitate cadmium as previously directed. Redissolve on filter with hot hydrochloric acid, neutralize filtrate with ammonia and add about I O grams trichloracetic acid: dilute t o 2 0 0 cc. and precipitate once more with hydrogen sulfide. A pure cadmium sulfide is now obtained, which may be determined by a n y of thi: methods previously described, b u t t h e most accurate is undoubtedly t h e gravimetric determination as sulfate. I think t h a t the above separation can be applied in brass analysis in detecting small amounts of cadmium. Recently, attention was called t o the difficulty of detecting 2nd determining small amounts of cadmium,2 and t h e need of more accurate methods. This applies particularly t o methods of separation, since t h e actual determination does not offer any unusual difficulties. C. H. LORDCORPORATION 1910 CALVMET 41.E , CHIC.%r.t8
THE CHEMICAL EXAMINATION OF NATURAL BRINESs By @. R . SWEE’JEY ANI)
J.¶?rfES
R . WITHROW
The proper analysis of natural brines has aliT-ays been important. They are used b y chemical manufacturers t o make comparisons m-ith a view t o reaching decisions as t o prospectire yields of salt, bromine and other products. The war-time elevation of t h e price of bromine from 3 0 cents t o as high as $ 6 . j o per lb., as well as a similar elevation of other products derived from natural brine, has given rise t o search for additional sources of these products and a careful scrutiny of many of t h e brines encountered in oil, gas and coal Sutton’s “Volumetric Analysis,” 10th Ed., p . 172. Eng. Mining J . , Mar. 3, 1917, p . 392. a Read before t h e Industrial Division, Kansas City Meeting of the -4merican Chemical Society, April I ? . 191:. 1
1
671
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
development, a n d hitherto wasted. As a result many analyses have been made in t h e last three years as a basis for manufacturing consideration. Some of these were made in t h e laboratories of manufacturers themselves a n d some b y consulting chemists. Analyses from both sources have come into t h e hands of t h e authors as t h e basis for report upon prospective manufacturing values. We, also, have had occasion t o make check or confirmatory analyses. It early became evident t h a t there was no standard or uniform procedure being followed by t h e different workers. To this fact may be due a large part of t h e non-agreement encountered from time t o time, though inexperience with this t y p e of analysis is also a factor. Few chemists, even water analysts, are experienced in such a t y p e of work as bromine determination in brine. This may be shown best b y citing a report t o its president by t h e laboratory of t h e chemical company. T h e letter from t h e president of one chemical company t o another stated: “The analysis of t h e two samples of brine which you sent us has been completed a n d i t was some job. T h e results are as follows:
........................... ........................ ........... ............................ ................................ .......... ............................
Depth.. Diameter.. Specific Gravity of Brine. CaO.. Br Halides (as sodium halide). Iodine
WELL A WELL B 1200 ft. 1200 ft. 8 in. 4 in. 1 .07 1 1.070 0.69 per cent 0 . 6 8 per cent 0.42 0.22 9.21 9.24, None None
These samples were from a brine whose composition was well known t o us. Furthermore, they were on t h e same property. It will be noticed t h a t they are of t h e same depth a n d also t h e same specific gravity, CaO content, a n d halide content. n’evertheless, they aFe reported of different bromine content-a divergence of nearly IOO per cent. Such a divergence would be a very important matter industrially, for one of these wells would give nearly twice t h e yield of bromine for t h e same turnover of salt and calcium chloride a n d at t h e same fuel cost as t h e other well. Considering t h e difficulty of bromine determination, by t h e usual methods, t h e infrequency of demand for i t a n d t h e concordance of all other determinations on these t w o brines one is tempted t o suspect t h e accuracy of t h e bromine determinations. As a matter of fact even t h e lower value is over twice t h e bromine content of t h e field in question as shown by both analyses b y various chemists a n d experience of all t h e plants operating on this particular brine. Such situations give rise t o controversy and discredit analytical chemistry. An examination of t h e literature for a basis for standard or uniform procedure disclosed no exact one which could be recommended. T h e procedures described for t h e examination of “mineral water” are not applicable directly. Certain modifications which our experience has introduced are recorded here. Not all of t h e procedures described have been exhaustively studied as yet. T h e purpose of this paper is t o make a beginning with t h e hope t h a t others, who have had experience in this work, will contribute their experiences, or will criticize these procedures. I n this way a procedure may be developed which may be accepted as standard.
Vol. 9, No. 7
T h e object is t o develop a method which will meet t h e needs of t h e manufacturing chemist rather t h a n a method of exhaustive analysis. Brevity and speed of manipulation, with reasonable accuracy, are, therefore, t h e requirements. ANALYTICAL PROCEDURES
SAmxE-The sample when pumped from t h e earth will generally be clear, b u t on standing i t becomes turbid due t o t h e separation of a brown precipitate. This precipitate is mainly iron, b u t may contain silica a n d alumina. It is probably caused by oxidation a n d hydrolysis of ferrous bicarbonate. Generally by t h e time t h e sample will have reached t h e chemist t h e iron will have separated. T h e scheme of agitation t o suspend t h e deposit uniformly through t h e liquid before taking a p a r t for analysis is inaccurate, as experiments have shown. Furthermore, t h e specific gravity is changed and this will affect t h e entire percentage composition. Consideration of this point has led us t o conclude t h a t t h e best procedure would be t o collect a sample of about one liter, allow i t t o oxidize and settle completely, determine t h e amount of deposit, a n d then make analyses on t h e filtered sample. T h e analysis would not be exactly t h a t of t h e original brine, b u t t h e difference will be very slight, and, since this procedure gives more nearly t h e thing t h a t t h e manufacturer wants, it is best t o proceed in this manner. D E P O S I T ON S T A N D I N G (AERATION)-The Sample Of about one liter, which will usually contain some deposit, is allowed t o stand, with occasional shaking, a n d removing of t h e stopper, for two or three days, or until deposition is complete and t h e precipitate settles well. T h e height of t h e liquid is carefully marked on t h e outside of t h e bottle, and t h e entire sample is t h e n filtered, rejecting t h e first I O O cc. T h e precipit a t e is well washed, ashed a n d ignited t o constant weight. The bottle is dried and t h e amount of water which i t contained t o t h e mark is determined. With these d a t a t h e grams per liter are calculated, using t h e specific gravity of t h e filtered sample, a n d t h e result recorded as “Deposit on Aeration.” The errors will not be large if percentages be calculated, using this figure. Since in t h e industries all natural brines are exposed t o air a n d allowed t o settle before they are further treated, this value is just what is wanted by t h e manufacturer. Further examination of t h e precipitate is not necessary. It is a question whether or not i t would be fair t o assume t h e precipitate t o be iron oxide (FenOs) a n d calculate i t to, a n d report i t as, ferrous bicarbonate. SPECIFIC GRAVITY-The specific gravity is obtained by t h e Westphal balance, and is taken a t I j o C., although perhaps i t would be better t o use zoo C., since this is more nearly t h e average temperature. T h e specific gravity of t h e fresh brine will be different from t h a t of t h e sample through which t h e precipitate is suspended, a n d this in t u r n will be different from t h e filtered sample. On one brine, for example, t h e specific gravity of t h e filtered sample was I . 2 3 0 7 , while t h a t of t h e sample in which t h e precipitate was suspended was 1.2342. If t h e chemist could t a k e
July. 1917
T H E JOURNAL OF I N D U S T R I A L A N D ENGINEERING CHEMISTRY
t h e specific gravity of t h e clear brine as soon as pumped from t h e well i t would no doubt be best, from t h e point of view of the original brine b u t this will generally be impractical. Even if t h e specific gravity could be obtained on t h e fresh brine there would be some volume change after t h e precipitate settled a n d a small error would be introduced when taking a filtered sample for later analyses. For these reasons, a n d also because t h e manufacturer is interested in t h e settled brine, i t is believed t h a t t h e best procedure is t o use t h e filtered, aerated brine, a n d t o determine t h e specific gravity with a Westphal balance a t 15 O C ( ? ) . This value is used in calculating percentages. T O T A L soLIDs---Many chemists omit this determination because of its questionable accuracy, b u t its value in calculating total water content for “evaporation fuel” comparisons makes it important. Brines rich in CaClz require a rather high temperature, above 160° C., t o expel t h e water completely. At this temperature t h e magnesium a n d calcium salts lose a part of their acid constituents, a n d a wide range of values will be obtained depending on t h e temperature a n d duration of heating. T h e total solids can be calculated from t h e complete analysis, b u t this value should be checked by t h e total solids as obtained b y evaporation. This point, is being studied in this laboratory a t t h e present time, a n d i t is hoped t h a t b y a suitable arrangement t h e volatilized acids may be collected, titrated a n d then be added t o t h e residual weight of t h e total solids. I t may be t h a t a weighed excess of some base may be added t o retain t h e acid which is otherwise volatilized. T h e constituents likely t o volatilize are chlorine, bromine, iodine (slight), sulfur trioxide a n d carbon dioxide (slight). T h e total solids are determined in t h e filtered sample, using 25-cc. portions a n d should be reported as Total Solids b y Evaporation. This gives t h e manufacturer a basis for a reasonable estimation of t h e water t o be evaporated in working t h e brine. SILICA-A 25-cc. portion of t h e filtered brine is acidulated with j cc. of concentrated hydrochloric acid a n d is evaporated t o dryness. It is then dried at 1 2 0 ’ C., or higher if necessary, €or a n hour. Five cc. of hydrochloric acid are then added, t h e vessel is warmed, 2 0 cc. of water are added, and, after warming, t h e whole is filtered a n d washed free from chlorides. T h e filtrate is evaporated a n d treated as just described, a n d t h e operation is repeated on t h e second filtrate. T h e combined precipitates are ignited in a platinum crucible, and weighed. The residue is treated with sulfuric a n d hydrofluoric acids. T h e loss in weight is reported as silica, a n d t h e residue is added t o t h e iron a n d alumina. I R O N A K D ALumxmf-The filtrate a n d washing, which should contain j cc. of concentrated hydrochloric acid, are treated with a few drops of nitric acid, boiled a few minutes, a n d then made alkaline with ammonia. It is then boiled until all t h e ammonia is expelled, a n d after standing, is filtered, washed and ignited in t h e crucible from which t h e silica was expelled. T h e residue is reported as Iropa and Aluminum Oxides. A separation of t h e iron a n d aluminum is not necessary.
673
T h e results should be reported separately from t h e iron which separated on aeration. It should be remembered t h a t t h e iron, aluminum a n d silica are not in solution as oxides, b u t as salts. For this reason there will be a slight difference between t h e total solids on evaporation a n d t h e calculated total solids. Much time is saved if t h e iron, alumina a n d silica are all precipitated together, with ammonia. T h e amount of silica remaining in solution is very small. These constituents have no commercial value, a n d need not be reported separately. It should be remembered in this latter case t h a t ammonium chloride must be added. CALcIux-The filtrate from t h e iron a n d aluminum is diluted t o 2 5 0 cc. a n d 2 j cc. are taken; this is diluted t o I j o cc., heated t o boiling a n d a hot I O per cent solution of ammonium oxalate is added in excess. After standing for some time (15 minutes), i t is filtered a n d washed with hot water. T h e precipitate is dissolved in warm dilute hydrochloric acid, a little ammonium oxalate solution added, and ammonia then slowly added t o complete the precipitation. The precipitate is filtered out, after standing one-half hour, and is ignited t o t h e oxide a n d weighed, or is dissolved in dilute sulfuric acid a n d titrated with permanganate. Our experience seemed t o show t h a t i t was not necessary t o allow t h e precipitate t o s t a n d 12 hours as is recommended in some of t h e books on water analysis. T h e calcium should be reported as sulfate a n d chloride. MAGNESIUM-The combined filtrates a n d washings from t h e calcium are acidified with hydrochloric acid, a large excess of sodium hydrogen phosphate is added, a n d t h e n ammonium hydroxide with constant stirring until t h e liquid smells of ammonia. Ten cc. of strong ammonia are added in excess and t h e whole is allowed t o s t a n d 1 2 hours. T h e precipitate is filtered out, washed with dilute ammonia a n d redissolved in hydrochloric acid ( I : 5 ) . T h e volume is made u p t o 7j cc., a little sodium hydrogen phosphate added, a n d then ammonia, drop b y drop, with constant stirring until t h e solution smells strongly. After 4 hours t h e magnesium is filtered out on a n alundum or Gooch crucible, washed with 2 per cent ammonia water, dried a n d ignited t o Mg2P207. If a n alundum crucible is used i t should be heated within a glazed crucible. T h e magnesium should be calculated t o t h e bromide a n d chloride. The above procedure gave very good results. I n t h e usual procedure t h e filtrate from calcium is evaporated t o dryness, a n d t h e ammonium salts are volatilized. This requires great care, and much time, a n d did not give a n y better results t h a n t h e procedure described. The reprecipitation must be carried out, even on very small amounts. There seems t o be good authority, however, for t h e evaporation a n d ignition which we have omitted, a n d t h e point should be investigated further.’ 1 In the discussion of this paper in the Industrial Division i t was suggested t h a t the speedier method for magnesium, used in Low’s “Technical Methods of Ore Analysis.” p. 159, might be of service here but we have not yet investigated this method, or the method of precipitation with sodium hydroxide sometimes used on traces of magnesium.
674
T H E JOURNAL OF I N D U S T R I A L A N D ENGINEERING CHEMISTRY
BARIUM A N D STRONTIUM-If no sulfates are present, barium a n d strontium must be looked for. Indeed B case is on record where barium, lead a n d sulfuric acid were present simultaneously in a natural mineral water.’ From work in progress in this laboratory i t seems, however, t h a t in the case of brines, where very little COZ is present, t h a t sulfates preclude t h e presence of barium or strontium. Barium oxalate is sparingly soluble, and a strontium oxalate is insoluble in water. When these metals are present t h e y will be partially precipitated along with t h e calcium. This point seems t o have been overlooked hitherto. In cases where the barium a n d strontium amount t o 0 . 2 per cent t h e error introduced cannot be neglected. T h e magnesium results may also be affected, It may be possible t o precipitate t h e barium a n d strontium with ammonium sulfate before precipitating t h e calcium, but no work has been done on this phase. The fact t h a t t h e barium is not completely precipitated b y ammonium oxalate makes it impossible t o apply a correction t o t h e calcium precipitate. The determination of t h e barium a n d strontium in t h e calcium precipitate is too time-consuming t o be practical for t h e ordinary technical analysis. When t h e barium a n d strontium content is small the error can be ignored. T h e procedure used was identical with t h e one described in t h e Department of Agriculture Bull., 91, “Mineral Waters of t h e United States,” b y J. K. Haywood a n d B. H. Smith; a simpler method has not yet been found. AMMONIA-Traces of ammonia have been reported in some brines, but t h e amount is generally so small as t o be of no commercial importance. It may be, however, t h a t its significance is greater t h a n we now know, especially in brines for electrolysis. The suggestion has been made by cell operators t h a t nitrogen chloride may be connected with the explosions which occur from time t o time in electrolytic chlorine apparatus. If this should prove true the determination of ammonia will be important. S U L P U R I C A C I D , S O D I U M A N D POTASSIUX-Fifty CC. of t h e filtrate from t h e iron a n d alumina are diluted t o I O O cc. a n d treated, while boiling hot. with I O per cent BaC12 solution, adding it slowly a n d with constant stirring. T h e B a S 0 4 is filtered off, t h e paper burned off in a porcelain crucible, and t h e precipitate dissolved in a few cc. of warm, concentrated sulfuric acid. T h e solution is now carefully poured into 2 5 0 cc. of water, and, after standing some time, is filtered, washed a n d ignited. It should be reported as calcium sulfate. This method of freeing the Bas04 f r o m iron a n d other absorbed matter is very effective. It is essentially t h a t taught for decades a t t h e J o h n Harrison Laboratory, University of Pennsylvania, Philadelphia. T h e filtrate from the sulfuric acid is used for sodium a n d potassium. From this point, t h e procedure we have been using is t h e same as described in “Mineral Waters in t h e United States,” Department of Agriculture Bull., 91,LOG.cit. I t is difficult, however, t o determine small amounts of potassium in the presence 1
Carles, Ann. chim. anal., 1902, 91.
Vol. 9, No. 7
of large amounts of sodium chloride, a n d it is believed t h a t some procedure’ should be used which will precipitate most of t h e sodium first. CHLORINE-The brine should be tested with phenolphthalein. I t will usually be neutral, b u t if it is not, it should be made so with NaHSO4 solution. I O cc. of t h e filtered sample are diluted t o a liter a n d I O CC. used for titration. This is diluted t o zoo cc., z cc. of KZCrO4solution are added and t h e mixture is titrated t o t h e end-point. NapCr04 would probably be a satisfactory indicator here b u t we have not yet proved this t o be true. T a k e a n amount of s t a n d a r d sodium chloride solution equivalent t o t h e a m o u n t of silver nitrate used, dilute t o zoo cc. and titrate as before. The difference represents the amount necessary t o affect t h e indicator a n d should be subtracted. This procedure is accurate enough since the chlorine i s used only as a check on the analytical work. T h e bromine value must be deducted. BROMINE-The colorimetric procedure, as given for ordinary waters, is not usable with brines. Experiments showed t h a t after repeated extraction with 90 per cent alcohol t h e residue still contained bromine. T h e distillation methods are time-consuming and not very easily manipulated. For these reasons a colorimetric method was developed. Procedure--Io0 cc. of t h e brine are made alkaline with N a 2 C 0 3 a n d are evaporated t o dryness. It is t h e n t a k e n u p in water a n d filtered into a 250-cc. flask. It is made distinctly acid with HzS04 a n d is diluted t o the mark: z j cc. are pipetted into a jo-cc. Nessler tube and chlorine is added until the maximum color has developed: I O cc. of carbon tetrachloride are then added, a n d t h e mixture is shaken a n d compared with a set of standards made u p from NaBr solution in the same way. By this rough check t h e approx’imate amount of bromine will be discovered, a n d a set of standard solutions are then prepared which are very close above a n d below the unknown solution. Again 2 5 cc. are taken, chlorine water is added t o a maximum color and t h e same amount is added t o the standards; t h e sample is then shaken with I O CC. of CCll and poured into a wet filter; when t h e water has drained off t h e filter should be punctured a n d t h e liquid caught in a zj-cc. Nessler tube (this is best done in a darkened room b u t darkness is not essential). If a sample does not exactly match t h e standard t h e colors can be compared b y diluting with CC14; or since t h e operation is so simple, a new set of standards can be made up, and t h e n a new determination made. If a test shows t h a t all bromine was not extracted by I O cc. of CCla a second extraction should be made. This is generally not necessary. Traces of iodine which are present in most brines will not interfere. T h e iodine need not be reported. I t is difficult t o appreciate the unreliability of published statements regarding t h e occurrence of bromine. For instance, although t h e State of Michigan reports, a n d t h e most reliable information states, t h a t Midland, Mich., brine contains 0.1 per cent of bromine, yet the I Professor C. W. Foulk, of the Division of -4nalytical Chemistry of the Ohio State university laboratory, is now investigating this matter.
July, 1917
T H E J O U R N A L OF I N D U S T I U A L A N D E N G I N E E R I N G C H E M I S T R Y
most exhaustive German work on bromine' states on page 3 t h a t t h e brine from Midland, Ohio, sic., contains 4 . I 8 per cent magnesium bromide which is equivalent t o 3 . 6 3 per cent bromine. or 36 times stronger t h a n those who operate on it cls.im it t o be. REPORTING RESULTS
T h e results should be reported in such a manner as t o give t h e manufacturer t h e thing w h k h he wants. The reporting of t h e constituents as ions, while strictly scientific, is of no value t o t h e manufacturer. A11 of t h e sodium and potassium should be calculated to chloride. Since t h e CaSOl separates on t h e copper tubes in t h e evaporators t h e H2S04 should be reported as calcium sulfate. The bromine should be calculated as magnesium bromide, since i t has long been so considered in t h e t r a d e ; b u t bromine as free bromine should also be reported. The residual calcium a n d all t h e magnesium are calculated t o chlorides since t h e y go on t h e market as such. Strontium and barium should be given as chlorides. The silicon should be reported as t h e oxide since t h e form in which i t is combined is not known. Iron a n d aluminum are reported together as oxides since their :separation is too time-consuming. The residue which separates on standing should also be given. Results are preferably reported in percentages though some manufacturers are accustomed t o grams per l i t e r . The specific gravity and temperature should always be reported; for this reason also, a standard temperat u r e should be used so t h a t results would be really comparable. When t h e positive and negative ions are calculated t o compounds they should nearly satisfj. each other. It should be borne in mind, however, t h a t t h e iron, aluminum and silicon are given as oxides, a n d not as salts, in which form they usually occur in t h e brine. There may also be small amounts of COZ a n d iodine which are not included. If, however, the check is not reasonably close, i t indicates a n error, or else some undetermined constituent is present. As a n illustration of t h e extremes in composition which t h e analyst must expect t o meet, t.wo examples from Ohio brines will serve. BRINESOURCE: Eastern Ohio Coal Mine Specific gravity. . . . . . . . . . . . . 1 , 0 3 4 Baume equivalent.. . . . . . . . . . 4.8' Sodium chloride. 3.26 per cent Magnesium bromide 0.007 Bromine . . . . . . . . . . . . 0.006 1.63 Calcium Magnesium chloride., . . . . . . . 0.05 Calcium sulfate.. . . . . . . . . . . 0.001
............
............
Southern Ohio Driven Well 1.1;o 22.2 12.08 per cent 0.124 0.1Oi 10.81 2.61 0.03 0.04 0.002 29.00
.
While this work considers primarily t h e commercial natural brines, t h e same procedure will doubtless a p ply t o t h e analysis of artificial brines, such as used in soda ash manufacture and in electrolytic cells, although t h e amounts of calcium a n d magnesium will be much less in these solutions. LABORATORY O F I I D U S T R I A L CHEMISTRS OHIO STATEUNIVERSITY,COLUMBUS 1 "Monographien u. angewandte Electrochemie. ifber d. elektrolytische C.ewionung von Brom." by Max Schlotter.
675
SEASONAL DISTRIBUTION OF SOX AND FECAL STRAINS OF THE COLON-AEROGENES GROUP IN SURFACE WATERS 5 y M Y R T L EGREEXPIELD AND
W. N. SKOURUP
Received April 30, 1917
When this work was started, i t was with t h e object of determining the variation of t h e organisms of t h e colon-aerogenes group in t h e surface water supplies of Kansas, during wet and dry weather, and their response t o treatment. Three supplies on t h e Verdigris River were chosenCherryvale, Independence and Coffeyville, a n d two on t h e Neosho River-Humboldt and Chanute. A11 t h e towns have rapid sand filters a n d coagulate with alum. Independence uses l5me in addition, a part of t h e time. Cherryvale pumps t h e water from t h e Verdigris River into a storage basin, holding 1,288,000 gallons, which is about five days' supply. From this i t flows by gravity t o t h e city four miles distant. Table I shows t h e pollution at t h e raw water intakes of t h e cities in question. SEWAGEPOLLCTIONAT R A W WATER INTAKES WATERPOLLUTED BY SEWAGE FROMDistant PopuConSewage. CITY Town Miles lation nections Purification Cherryvale Neodesha 20 3,011 876 None Independence hleodesha 24 3,011 876 S-one Independence1 24 12,144 2600 Septic tank Sr Coffeyville Cherryvale2 30 4,235 8001 Contact bed Humboldt Iola 8 7,866 2350 Septic tank Iola & 16 7,866 2350 Septic tank Chanute 8 2,131 150 Septic tank Humboldt 1 Contact bed for one-third of the sewage. 2 €ontact bed in poor condition. T A B L EI-STATISTICS -RAW
OF
I& 1
UETHOD
O F ISOLATIOX
portable laboratory was set up a t Independence and the collection of all samples was personally supervised. Samples were iced a n d delivered t o t h e laboratory a few hours after collection. Upon receipt a t t h e laboratory, they were plated in standard agar and incubated 2 1 hrs. a t 37'. Various dilutions were planted a t t h e same time in lactose peptone bile. Streaks from t h e positive fermentation tubes were made on Endo plates. Three coli-like colonies were picked from each Endo plate and grown on a n agar slant. These slants were sent t o t h e main laboratory and purified again by streaking on Endo plates and picking a characteristic colony. From these, transfers were made t o lactose, dextrose, saccharose, and dulcite broth tubes and t o t h e di-potassium acid-phosphate media of Clark & Lubs.' The broth tubes were incubated 48 hrs. at 37' a n d t h e di-potassium-acid-phosphate tubes 7 2 hrs. at 3 7 '. One-half t h e latter was treated with methyl red t o indicate the H+ ion concentration a n d t h e other half with I O per cent KOH to obtain t h e Voges-Proskauer reaction. All cultures t h a t did not ferment lactose and dextrose were discarded. It was demonstrated by Rogers, Clark and El-ans? that organisms of t h e colon-aerogenes group occurring on grains may be differentiated from those of fecal origin by t h e gas ratio. Clark & Lubs2 showed t h a t there is a complete correlation between t h e gas ratio and t h e Hf ion concentration, the fecal strains in their media being characterized by a high Hfion Infec. Dis.. 17 (1915). 13: I b i d . . p. 160.
:J . 2
