grams of potassium sulphate it was practically complete in 1 hour. Hydrolysis with 0.7 gram of mercuric oxid, 12.2 grams of sodium sulphate and 15 cc. of sulphuric acid was incomplete in 12 hours, but complete in 24 hours. In a number of experiments conducted with open flasks in the presence of 0.7 gram of mercuric oxid and 10 grams of potassium sulphate, hydrolysis with 25 cc. of sulphuric acid required 2 hours, while with 20 cc. of acid only 1 hours were required. With 8.2 grams of sodium sulphate, instead of 10 grams of potassium sulphate, and 15 or 20 cc. of sulphuric acid, 2 hours were required, while 11⁄2 hours showed complete hydrolysis. It is to be noted that with the use of the open flasks another factor, the rate of volatilization of acid, is introduced. This effect depends mainly on the intensity of heating and to a less extent on the time of boiling. The hydrolysis of certain organic compounds of various constitutions was reported in 19161. In the presence of 0.7 gram of mercuric oxid, 10 grams of potassium sulphate and 25 cc. of sulphuric acid, weights of the compounds varying from 0.2 to 0.4 gram were hydrolyzed completely by heating in open flasks at the boiling point for 2 hours. The use of sodium sulphate in the place of potassium sulphate has been applied to the compounds given below. Reliable results were obtained when these compounds were boiled for 2 hours with mixtures of 0.7 gram of mercuric oxid, 8.2 grams of sodium sulphate and 20 cc. of sulphuric acid. Below are grouped the compounds studied. When 25 cc. of acid were used with 8.2 grams of sodium sulphate and 0.7 gram of mercuric oxid, the hydrolysis in many cases was not complete 1 J. Assoc. Official Agr. Chemists, 1920, 3: 306. as with nicotinic acid, nicotin zinc chlorid and hydroxyquinolin. With this mixture, the completeness of hydrolysis depends very markedly on the amount of acid volatilized. The influence of reagents and apparatus was also investigated. Ammonia free water, redistilled from alkaline permanganate solution through a metal Kjeldahl connecting bulb soldered to a block tin condensing tube, was used in this work. The bulb was connected with the flask by means of a cork stopper entirely covered with tin foil. In all experiments where it was desired to avoid the influence of glass or of rubber stoppers, the above apparatus was used. It was noted that the pure reagents of commerce for the estimation of nitrogen by the Kjeldahl hydrolysis contribute small amounts of ammonia reacting substances. Rubber stoppers in the Kjeldahl flask during hydrolysis' are believed to contribute ammonia reacting substances. Again, rubber stoppers used in connecting the Kjeldahl flask to the condenser contribute ammonia reacting substances. The ammonia reacting substances contributed by a rubber stopper held in the neck of a flask during the acid hydrolysis are sufficiently large to be appreciable in ordinary routine Kjeldahl determinations, unless very closely defined conditions are followed. The error is, furthermore, variable and the use of stoppers unnecessary. Even the especially purified reagents contain traces of such substances. The glass, also, contributes a small amount of alkaline reacting substances. The magnitude of the error due to the alkaline reacting substances is, however, so small that all of these may be neglected except in work in which high precision is necessary. The conclusion is obvious that, in all routine work involving determinations by the Kjeldahl method, it is necessary to deduct from the result obtained the amount corresponding to the ammonia reacting substances contributed by reagents and apparatus in use in the particular experiments. The results recorded in this investigation were obtained even when greater precautions were taken than is customary in routine work. It is very obvious that under less carefully controlled conditions in routine work the errors, which are here called inappreciable, will become large enough to affect seriously the accuracy of the results obtained. When hydrolysis is made in open flasks, the proportions of sulphuric acid and potassium or sodium sulphate which may cause loss of ammonia by volatilization have been indicated. The proportions of sulphuric acid, mercuric oxid and potassium or sodium sulphate giving complete hydrolysis of a refractory compound (pyridin zinc chlorid) have been shown. These proportions contain slightly more acid than those which may cause loss of ammonia by volatilization. This difference in proportions is so slight, however, that hydrolysis in open flasks of 1 J. Ind. Eng. Chem., 1916, 8: 639. refractory compounds must be conducted with closely controlled conditions. This is particularly true in the case of sodium sulphate because amounts of sodium sulphate in excess of those proportionate to the suitable amounts of potassium sulphate are necessary. The differences in the behavior of the two sulphates is probably due to differences in the tendencies of the acid sulphates to retain water. Consequently, the hydrolysis of very refractory compounds with sodium sulphate is not at present recommended, although with very closely controlled conditions excellent results can be obtained. The influence of the reagents and the apparatus on the accuracy of the modified Kjeldahl method has been indicated. REPORT ON POTASH. By T. D. JARRELL1, (State College of Agriculture, College Park, Md.), Referee. The following samples and instructions were sent to collaborators: Sample No. 1.-Commercial potassium chlorid. Sample No. 2.- A mixture of acid phosphate and sulphate of potash-magnesia (containing about 9% of potassium oxid). Sample No. 3.-A mixture of acidulated garbage tankage, acid phosphate and commercial muriate of potash (containing about 7% of potassium oxid). SAMPLE No. 1. Determine the potash by the official2 and the following method: PERCHLORATE METHOD. Dissolve the potash as in the official method. Acidify with about 5 cc. of hydrochloric acid. While the solution is hot, precipitate the sulphate by adding, drop by drop, in slight excess normal barium chlorid solution acidified with hydrochloric acid. Cool, make to volume and shake. Allow the precipitate to settle and filter. Transfer an aliquot, corresponding to 0.25 gram of sample, to an evaporating dish, add 5 cc. of perchloric acid (sp. gr. 1.12); evaporate on a steam or sand bath until it fumes strongly, take up the residue with 5 cc. of water, add a second 5 cc. of perchloric acid and again evaporate the solution carefully until all free hydrochloric acid is driven off and dense white fumes of perchloric acid appear. If the solution goes to dryness and a hard mass remains, take up with a few drops of perchloric acid. When a water bath is used for the evaporation, finally place the dish on a hot plate and heat carefully until hydrochloric acid is driven off. After cooling, add 20 cc. of 95% alcohol and stir well. Allow to stand for 30 minutes. Decant the alcohol through a Gooch crucible having a fairly thick pad (about } inch thick) of asbestos and wash twice by decantation with 95% alcohol containing 0.2% perchloric acid, made by adding 1 cc. of perchloric acid (sp. gr. 1.12=20%) to 100 cc. of 95% alcohol. Transfer the precipitate to a Gooch crucible with the 95% alcohol containing perchloric acid and wash until the entire filtrate amounts to 75 cc. Finally wash twice 1 Present address, Bureau of Chemistry, Washington, D. C. 2 Assoc. Official Agr. Chemists, Methods, 1916, 12. with alcohol-ether (1 part 95% alcohol to 1 part ethyl ether), using 3-5 cc. each time to wash out all perchloric acid. Dry for 30 minutes at 120-130°C., then weigh. Dissolve the potassium perchlorate from the Gooch crucible with about 200 cc. of hot water and dry to constant weight in an air oven. Allow to cool and weigh. The loss in weight is potassium perchlorate (KClO). SAMPLE NO. 2 AND NO. 3. Determine the potash by the official1 and the following methods: MODIFIED OFFICIAL METHOD. This is the same as the official method except that the addition of 2 cc. of concentrated hydrochloric acid to the potash solution is omitted. After washing 2.5 grams on the filter paper with boiling water, add directly to the hot solution ammonium hydroxid and ammonium oxalate and proceed as in the official method. PERCHLORATE METHOD. Weigh 2.5 grams of the sample upon a 12.5 cm. filter paper and wash with successive small portions of boiling water into a 250 cc. graduated flask to a volume of about 200 cc. Allow to cool, make to the volume and shake. Do not add ammonium hydroxid or ammonium oxalate. Transfer 50 cc. of the solution to a porcelain or silica dish (do not use platinum), add an excess of a 3% solution of barium hydroxid and without filtering evaporate to dryness over a sand bath. Gently ignite the residue over a Bunsen burner below redness for about 5 minutes. Extract the residue with 25 cc. of boiling water, breaking up the material as much as possible. Filter into an evaporating dish of about 175 cc. capacity, and wash with boiling water until the filtrate amounts to about 150 cc. Add 5 cc. of perchloric acid, evaporate carefully on the sand bath until it fumes strongly, take up with 5 cc. of water, add a second 5 cc. of perchloric acid, evaporate, cool, and proceed as already outlined. It is requested that the modified official method be tested thoroughly (i. e. omitting the addition of 2 cc. of concentrated hydrochloric acid to the potash solution) on some high potash content samples, using acid phosphate, kainit, commercial potassium chlorid, commercial potassium sulphate, and manure salts. Since these mixtures can not be obtained on the market at present, it is suggested that the collaborator prepare them in his laboratory. TABLE 1. Collaborators' results* of potash determination expressed as potassium oxid. |