248 



SCIENCE. 



[N. S. Vol. V. No. 111.. 



R. method. S. method. V'y S. Method 

 1. 2. 1. 2. 1. 2. 

 Average of all deter- 



minatiojQs 0.0499 0.0840 0.0490 0.0826 0.049G 0.0837 



Highest result 0.054 0.091 0.052 0.086 0.055 0.089 



Lowestresult 0.045 0.078 0.046 076 0.046 0.078 



Maximum difference... 0.009 0.013 0.006 0.010 0.009 0.011 



To our minds these figures are very im- 

 pressive. It is worthy of note : 



1st. That the average results given by 

 the ' rapid ' methods only differ on either 

 steel from the averages given by the ' slow ' 

 or ' very slow ' methods, by a little over 

 0.001 of a per cent. 



2d. That the maximum difference be- 

 tween the highest and lowest results given 

 by the ' rapid ' methods on either steel is 

 but a trifle greater than is shown by the 

 ' slow ' or ' very slow ' methods. 



In other words, if we interpret these 

 results correctly they show that the rapid 

 methods for determining phosphorus in 

 steel now known and used in many labora- 

 tories give results that are well-nigh as ac- 

 curate and reliable as those yielded by the 

 longer and more laborious methods, and it 

 must not be forgotten that, although we 

 have placed two hours as the time char- 

 acterizing a rapid method, a number of the 

 results given above were obtained by the 

 use of methods which give a single de- 

 termination in forty-five minutes, and en- 

 able one operator to make twenty phos- 

 phorus determinations in a day. We are 

 free to say we do not believe such a show- 

 ing would have been possible five years 

 ago. 



But these results still leave something to 

 be desired. The discrepancy between the 

 highest and the lowest result is still too 

 great. It is, perhaps, a little hazardous to 

 place limits, but we do not think the chem- 

 ists of the country should be satisfied until 

 they are in possession of a method or 

 methods which are so carefully worked 

 out and so well described that in the hands 

 of different chemists of fair ability and 

 experience results will be obtained by all, 



when working on the same steel, that will 

 not differ from each other more than 0.003 

 per cent. The Sub-Committee on Methods of 

 the International Committee on Standards 

 for the Analysis of Iron and Steel, before re- 

 ferred to, have had in hand, now for some 

 two years, studies on a rapid and accurate 

 method for the determination of phospho- 

 rus in steel. It has been the hope of the 

 Sub- Committee that the ideal above given 

 would be attainable by this method. In 

 reality, the work of the Sub-Committee has 

 embraced an examination of almost every 

 chemical point involved, taking very little, 

 if anything, for granted, and checking and 

 proving every step. The work is not yet 

 quite ready for publication, one or two 

 points remaining which are not entirely 

 settled, and it has been deemed advisable to 

 withhold the method until these are com- 

 pletely cleared up. 



Some years ago, with the publication* of 

 what is commonly known as Ford's method, 

 the determination of manganese took a de- 

 cided step forward, at least in this country, 

 so far as speed is concerned. Previous to 

 that time the long and laborious acetate 

 method, which involved the separation of 

 the iron from the manganese as basic ace- 

 tate and subsequent precipitation of the 

 manganese by means of bromine or as py- 

 rophosphate, had held full sway. Ford's 

 contribution consisted, as is well known, in 

 separating the manganese from hot nitric 

 acid solution of the iron or steel, by means 

 of potassium chlorate, and Williamsf added 

 the modification, now in common use, of 

 determining the separated oxide of manga- 

 nese, by its action on a standard solution 

 of ferrous sulphate or oxalic acid. This 

 method, as now worked in many laborato- 

 ries, gives a single result in forty minutes- 

 and two in an hour, and enables one opera- 

 tor to turn out twenty to twenty-five deter- 



* Trans. Am. Inst. Mining Eng., 9, 397. 

 t Trans. Am. Inst. Mining Eng., 10, 100. 



