398 S, ./. SHAND 



section of the cross-wires); screw L is now turned until the grain 



travels past and its other margin lies exactly beneath the cross; 

 then screw A' is turned till the next augite grain comes into position; 

 then screw /. till the grain has passed, and so on. When the 

 traverse has been completed, which with a section of ordinary size 

 may take from one to three minutes, the readings oi the microm- 

 eter screws are again noted ami written down below the former 

 readings. It is obvious that the difference between the two read- 

 ings oi screw /. gives the sum of the diameters of all the augite 

 grains which were intersected during the traverse, while the differ- 

 ence for screw R gives the sum for all other minerals in the rock. 

 Without stopping to make these subtractions, however, sledge (.' 

 is pushed forward into a new position and a second traverse is 

 made in the return direction; at its completion the micrometers 

 are again read and the readings jotted down beneath the previous 

 ones. Sledge C is again pushed forward, and another traverse 

 made, and so on until the number of traverses is considered sutli- 

 cient. One may now proceed to the subtraction of the successive 

 readings, and the calculation of the percentage of augite. which is 

 obviously 



sum of successive differences of L N/ 



sum .•! successive differences of L + sum of successive differences of /C 



The most expeditious manner of recording the readings is to write 

 them down in parallel, vertical columns in the middle oi the page; 

 then, when all the measurements have been made, the differences 

 of /. are quickly tilled in to the left and the differences of A' to the 

 right, as in Table 1. on p. 399, which is part of an actual 

 estimation. 



It will be seen that by this method only two numbers have 

 to be recorded for each traverse after the first, while by any other 

 method some twenty or more may be necessary. 



It may be advisable at this stage to recall the conditions which. 

 as Rosiwal has pointed out. must be observed if the linear traversing 

 method i- to give reliable results. 



1. The length measured must be at least one hundred times the 

 average grain of the rock. (With the additional facility afforded 

 In- the recording micrometer, it would involve little extra labor 



