162 



neutral or acid to sensitive neutral litmus paper, when made neutral to litmus were, 

 in every case, still acid toward phenolphthalein. 

 Osborne further says: 



To render gram portions of these several protein preparations neutral to litmus 

 required in a few cases not any, in most cases from 0.1 to 1.5 cc of decinormal alkali;, 

 while to make the same gram portions neutral to phenolphthalein required the further 

 addition of from 0.7 to 1 cc of decinormal alkali, except for legumin, which required 

 2 cc. Edestin made neutral to phenolphthalein and dissolved in sodium chlorid solu- 

 tion reacts distinctly alkaline toward litmus. This Alkaline reaction is caused by the 

 edestin itself and not by organic salts of the alkali, since such preparations yield a 

 very small amount of ash, less than 0.05 per cent, which is neutral to both litmus and 

 phenolphthalein. Solutions of all the other protein bodies I have exam- 



ined, when similarly made neutral to phenolphthalein, react decidedly alkaline with 

 litmus. 



In the investigation which the referee reports at this time no attempt was made to 

 determine total acidity, only those acids being taken into account which were extracted 

 by a rather prolonged treatment with water. Sixteen samples of gluten feed, repre- 

 senting five brands, two of wheat bran and one each of wheat middlings, wheat feed 

 and cottonseed meal, were examined. 



Ten grams of the feed were weighed into a beaker and stirred with 50 cc of water 

 for ten minutes, then transferred to a plain wet filter and washed with successive 

 small portions of water, until the washings amounted to 150 cc; the extract was 

 then made up to 200 cc with water and 20 cc portions (=1 gram feed) used in the 

 subsequent titrations. A blank determination was also made with 200 cc of water 

 run through a filter paper as before, and the washings were found to be neutral to 

 methyl orange, phenolphthalein, and litmus. 



The following indicators were used: 



Phenolphthalein: One gram dissolved in 100 cc of 50 per cent alcohol. 



Litmus paper: Very sensitive neutral paper. 



Methyl orange: One gram dissolved in 1,000 cc of water. 



Congo red: One gram dissolved in 100 cc of 30 per cent alcohol. 



Gunzburg's reagent: Two grams of phloroglucin and 1 gram of vanillin dissolved in 

 30 grams of alcohol. 



Toepfer's reagent: One per cent solution of phenolphthalein in alcohol and 0.5 per 

 cent solution of dimethylamidoazobenzol. 



The alkali used was approximately decinormal sodium hydroxid, 1 cc being equal 

 to 0.003996 gram sodium hydroxid. 



Twenty cubic centimeter portions of the watery extract, equal to 1 gram of feed, 

 were taken for each test. Owing to the usually highly colored solutions, the aliquot 

 was diluted with 500 cc of water for the test with methyl orange, and with 50 cc for 

 the other indicators, except in the Giinzburg test, where 0.5 cc of the extract and 

 the same quantity of the reagent were used. Three hydrochloric acid solutions were 

 also prepared, N/14, N/28, and N/56, respectively. 



The Giinzburg and Toepfer tests and Congo Red are recommended as reliable in 

 determining whether or not free mineral acid is present. These tests were applied 

 first to the three hydrochloric acid solutions and unmistakable positive results were 

 secured with all, the most dilute acid used, N/56, equivalent to about 0.065 per cent 

 of hydrochloric acid, responding perfectly to the reactions indicated. A mixture of 

 one of the aqueous feed extracts and dilute hydrochloric acid, the total mixture con- 

 taining 0.065 per cent of free hydrochloric acid, was likewise subjected to these tests 

 and positive proof was secured that nothing present in the feed extract in any way 

 interfered with the delicacy of the reactions. However, following the suggestions of 

 Osborne's work, when the salt of a weak acid, for instance, sodium acetate, was added 

 to these same test solutions, none of the above prescribed tests for free mineral acidity 

 responded, although hydrochloric acid had been added in every case. This experi- 

 ment shows quite conclusively the danger and inaccuracy of asserting either the 

 presence or absence of mineral acids from data obtained by these tests. 



