When the decomposition is ended the crucible is strongly heated so as to hold the platinum sponge, which is produced, together sufficiently to prevent its being removed in the subsequent washing of the crucible by decantation. By the ignition at a high temperature necessary to secure this, the greater part of the calcium chlorid is volatilized. After cooling, a few drops of concentrated hydrochloric acid are placed in the crucible and if the slightest yellow color be shown the acid is evaporated and the ignition repeated, with the addition of a little oxalic acid. In most cases the slight yellow color produced comes from a trace of iron and will therefore appear again after the second ignition. The crucible is subsequently washed by repeated decantations, finally with boiling water, and after drying is ignited and weighed. The advantage of this process is that without further trouble the reduced metal is completely freed of any salts of the alkaline earths, etc., which have been carried down with it and also from any of the uncombined sodium chlorid which may not have been washed out by the alcohol. In fact, the results obtained in this way are nearly always lower than those obtained through the direct weighing of the double salt, and the wash water which is first poured off contains, as a rule, traces of the alkaline earths and almost without exception some sodium chlorid. Correction for the filter ash is unnecessary because the ash is completely dissolved by the treatment received. The platinum sponge which is collected in the crucible in this way is removed in case it does not adhere to the sides and the crucible is then ready for the next operation. =265. Sources of Error in the Platinum Method.=—In the comparative work done in the determination of potash by the members of the Association of Official Agricultural Chemists there has been noted, from year to year, marked differences in the data obtained by different analysts. Such differences often are due to personal errors, or a failure to accurately follow the directions for manipulation. Sometimes, however, they are due to sources of error in the processes employed. In the platinum method these sources of error have been long known to exist. Chief among these is the remarkable facility with which potash becomes incorporated with the precipitates of other bodies. The character and magnitude of some of these errors have lately been studied by Robinson.[214]