614 STATE BOARD OP AGmCULTURE. 



color cliaiige occurs at approximately this point, i. e., its effective range 

 must include a Ph range of 7.0-7.4. In the snlphonphthalin series this 

 range is best covered by phenolsulphonphthalin or phenol red. Hence 

 tliis indicator was tested as a substitute for rosolic acid to the use of 

 which there is tlie very weighty objection that the reaction which is to 

 be measured lies very close to the upper limit of its range of usefulness. 

 It is diflficult to observe a difference in the color of rosolic acid in buffer 

 solutions of 7.4 and 7.G wlien they are not placed in a series with other 

 solutions Avith wliich a comparison serves to accentuate the differences 

 in the colors of the individuals. And even between 7.4 and 7.8 the 

 distinction is one of intensity rather than of character, both being pinks, 

 the latter somewhat deeper than the former. 



Witli plienol red, however, there is still a decided change in the 

 character of the color to take place in passing towards the alkaline 

 limit of its range. The color of this indicator at 7.G contains much 

 more red and decidedly less brown than at 7.4. Hence in titrating to 

 tlie latter reaction any overstepping of the mark is at once indicated 

 by the increase in the red element in the tube. 



The point of weakness in the coaorimeter method lies in the prepara- 

 tion of the standard solutions. Unless these are accurately prepared or 

 can be checked electrometrically the results are open to question and 

 in order that tliis may be avoided every precaution must be taken to 

 secure (a) chemicals of absolute purity and of known composition and 

 (b) to make up all solutions of them with the utmost care. To test 

 the accuracy with which the standard solution used in the Eastman and 

 Hildebrand method could be duplicated, four such solutions were pre- 

 pared using the same sample of phosphate. They all checked the 7.4 

 standard of Clark and Lubs (See Table I) as closely as could be read. 

 To 20 cc. of one of them an extra cc. of N/IO HCl was added (4.2 

 cc. had been added to this amount to bring its reaction to the proper 

 point as prescribed in the directions for tlie preparation of the 

 standard solution) and to tlie same quantity of another, one cc. of 

 N/10 NaOH was added. In each case the reaction was changed only 

 about O.IPh. This shows that the standard is quite a reliable one 

 when prepared according to directions. As has been pointed out by 

 others, Na.^PO^ is not an ' ideal substance to use for this purpose. 

 The variation in its water of crystallization makes it difficult to prepare 

 an accurate!}' quantitative solution but in the present case the varia- 

 tions resulting from this source of error would probably not be suflficient 

 to cause any significant variations in the result obtained with it. As 

 will be pointed out later, however, there is one source of error Avliich 

 sliould be avoided. 



Several titrations of an acid ammonium citrate solution were made 

 using this standard and rosolic acid. IJJ/IO NH^OH was used to titrate 

 10 cc. samples. 10 cc. jiortions of the standard solutions were taken and 

 the pT'opcr fjuantities of acid and indicator added. As tlie addition of 

 tlie ammonia solution to the samples almost doubled their volum&s, 

 water was added to the standard so that the concentration of the indi- 

 cator would be aj)]iroximateily the same in each tube. The first two 

 titralions were made comjiaiiiig the colors with that of the standard 

 by merely holding (hem in the hand. For (he remaining three the 

 comparator was used. The following amounts of N/10 NH^OH were 

 required : 



