Febefary 12, 1897.] 



SGIENGE. 



247 



the different chemists is very good. The 

 exceptions are the No. 3 steel, which con- 

 tains arsenic in considerable amount, and 

 where the discrepancy is 0.009 per cent., 

 and in the No. 4 steel, where the discrep- 

 ancy is 0. 007 per cent. Considerable work 

 was done on the No. 4 sample, in an ef- 

 fort to reconcile discrepancies, and it was 

 found that the turnings from this sample 

 were irregular, and that two different bot- 

 tles of the sample gave different results. 

 The average of six determinations from one 

 bottle was 0.1057, and the average of five 

 determinations from another bottle was 

 0.0964 per cent. Furthermore, siftings 

 from quite an amount of the turnings gave 

 0.140 per cent." 



But these methods are long and laborious. 

 It would be impossible with the most rapid 

 of them to get a result in much less than a 

 day, while two days would certainly be re- 

 quired for some of the others . Accordingly , 

 since the demand for rapid phosphorus de- 

 terminations during the last ten or fifteen 

 years has been very great, an enormous 

 amount of work has been done in trying to 

 meet this demand . Modification after mod- 

 ification has been introduced, and paper 

 after paper published on the subject. It is, 

 perhaps, not too much to say that few chem- 

 ical journals that publish any original work 

 at all have escaped three or four articles 

 per year, on the determination of phosphorus 

 in iron and steel, or on some phase of a 

 rapid method for such determination. The 

 result of all this work has apparently been 

 constantly increased rapidity, with con- 

 stantly greater approximations to accuracy. 

 The present state of the matter is, perhaps, 

 best shown by Thackray* in his paper, ' A 

 Comparison of Recent Phosphorus Determi- 

 nations in Steel.' This writer sent to some 

 twenty- three different chemists borings from 

 two different samples of steel, with a request 

 to have the phosphorus determined in each 



* Trans. Am. Inst. 3Iin. Eng., 25, 370. 



sample, and a description of the method 

 used sent with the results. Each chemist 

 was told that samples had been sent to 

 others, but no attempt was made to have 

 any special method used. The chemists 

 embraced a professor in a technical school, 

 the chemist of a large consumer, a number 

 of commercial chemists, and a number of 

 chemists employed by steel and iron works. 

 Oaone sample thirty-six different results 

 were sent in, and on the other thirty-eight. 

 Twenty-seven different methods were em- 

 ployed, some of the chemists sending in re- 

 sults by two and even three methods, and 

 some sending duplicate determinations. The 

 results obtained were obtained as follows, 

 the figures being percentages of phosphorus 

 in the steels : 



Average of all determinations 0.0496 0.0835 



Highest result 0.055 0.091 



Lowestresult 0.045 0.076 



Maximum diflerence 0.010 0.015 



The methods employed may be divided 

 on the basis of time required into three 

 classes: 



1st. Those which may be called rapid, 

 and which give a result in two hours or 

 less. 



2d. Those which may be called slow, 

 and which require considerably more than 

 two hours, but still give a result the same 

 day. 



3d. Those which may be called very 

 slow, and which do not give a result until 

 the second day or later. 



Thirteen results on each sample were 

 given by ' rapid ' methods, eleven on the 

 No. 1 sample and twelve on the No. 2 

 sample by ' slow ' methods, and twelve on 

 the No. 1 and thirteen on the No. 2 by 

 ' very slow ' methods. Arranging the re- 

 sults in accordance with this classification 

 of the methods, we have some very inter- 

 esting data, the figures being as before, 

 the percentages of phosphorus in the two 

 steels: 



