354 TOOL STEEL FOR THE UNITED STATES NAVY. 



three-fourths of an inch of the shank into brine for a few seconds, and then 

 the entire tool. The cutting end is treated at the same temperature as the 

 hammer end and the temper of both ends drawn by submerging the chisel 

 in a molten lead bath at the desired temperature. One of the samples 

 submitted under this class contained chromium, and the heating and cooling 

 curve showed that a higher temperature was needed for treating, and the 

 test showed that the temper should be drawn at a higher temperature than 

 for the carbon tool steels. 



It has been noted when making tests of the chisels that the heat treat- 

 ment does not extend back far from the cutting edge, and there is only a 

 short distance on the tool where the maximum cutting life can be obtained. 

 If the chisel proved to be brittle on the first test, indicating that the temper 

 had not been drawn sufficiently, in all probability the second test would be 

 more satisfactory, while a third test might show that the chisel was too soft. 



The button-rivet set used for the class 4 carbon tool steel is shown on 

 sheets Nos. 1058 and 1521, Plates 142 and 144. The set was heated to the 

 desired temperature and quenched in brine, after which the temper was 

 drawn in a lead bath. Each set drove a certain number of hot rivets, and 

 an observation of its condition was made after the test, but this is not a test 

 to destruction as are the preceding tests, and is, therefore, not so decisive. 



GENERAI, NOTBS. 



Earlier in this paper the statement was made that the observations of 

 the elapsed time of run for the selective test of tungsten tool steel, class i, 

 data for which are shown on sheet No. 1058, Plate 142, were adjusted by the 

 principles of least squares. By referring to that sheet it will be noted that 

 four observations were rejected, and by referring to sheet No. 1520, Plate 

 143, it will be noted that only one observation was rejected. The terms 

 "residual" and "probable error" used on sheet No. 1058, Plate 142, were 

 changed to "deviation" and "probable deviation" on sheet No. 1520, Plate 

 143. The observations of the carbon tool steels do not agree as closely as 

 those for tungsten tool steel, class i, and the causes for the variations are 

 not very easily determined. In the case of the milling cutters the cutting 

 life is considerably reduced if there is very much vibration, and in order to 

 overcome this as much as possible a heavy, rigidly constructed milling 

 machine was used. If the cutters were not exactly central in the arbor so 

 that they did not rotate around their geometric axis, vibrations would be 

 set up which would increase in violence until the cutter finally broke, and 

 it was therefore necessary to fit very carefully the cutters into the arbor. 



