156 Wisconsin Academy of Sciences, Arts, and Letters. 



holes is a rather haphazard one at best, so far as accuracy is con- 

 cerned. 



There are two chances for error bj the time the plates are 

 rolled. First. The holes will not all be made exactly where 

 marked ; if one whole is punched slightly one side of its mark, 

 and the one which it should match the other way, the error is 



Fig. 1. 



multiplied. Moreover, it is quite impossible to produce these 

 plates and have them perfectly homogeneous. There will be hard 

 and soft places. The great pressure from the rolls in making the 

 plates cylindrical will cause changes in distance between some 

 of the holes, as the temper of the plate varies. When the cylin- 

 ders are placed together for riveting, many holes will shut past 



Fig. 2. 



Fig. 3. 



Fig. 4. 



one another from 1-16 of an inch to 1-2 or 2-3 their diameter. 

 This, in itself, is objectionable enough, but the case is aggravated. 

 The overlapping metal should all be removed by the reamer and 

 the hole filled by a suitable rivet. If the overlapping of the holes 

 is not such as to compel the use of the reamer, a most objection- 

 able resort is the tool known in shop parlance as a " drift pin," 

 which is nothing more than a steel pin, slightly tapering, and when 

 well oiled can be driven in with such force that the solid iron is 



