ASSOCIATED METAMOKPHIC KOCKS OF THE LAKE-DISTETCT. 



573 



number of cases to work upon, it seems almost hopeless to get more 

 than a very rough general result. 



By a, perfect case, the measurement of which may he considered 

 fairly reliable, I mean therefore one in which the cavity is of uni- 

 form depth and of just sufficient depth to enable the vacuity to 

 move freely to every part without once going out of focus ; in such 

 cases every thing will be sharp and clear in one focus, and, unless 

 the shape of the cavity be very irregular, the relative size of the 

 vacuity to the whole cavity may be approximately ascertained. Many 

 of the measurements to be brought forward presently were made 

 from such as these, though in others the vacuity has not been 

 seen to move to every part of the cavity. From this and similar 

 sources of error it is impossible to expect any thing like an absolute 

 uniformity in the measurements ; but when, all extreme cases being 

 discarded, an average is struck, the mean result may be fairly taken 

 to represent the probable truth. The following are examples of 

 this : — 



■166 



No„ 1. < 



154^ 

 154 

 154 

 166 



180 



180 



154 



l; 154j 



y -162 



•125 

 •200 

 •166 

 •200 

 •180 

 No. 2. ^ -166 y 

 •166 

 •166 

 •154 

 •166 

 •142 

 L-142 



•164 



. The above are consecutive measurements from different rock-slices. 

 In No. 1 there are five cases in which the whole liquid-cavity is 6£ 

 tiraes the size of the vacuity, one case in which it is 6 times the 

 size, and two cases in which it is only 5^ times. The mean, "162, is 

 probably not far from the real truth. 



In No. 2 there is one case in which the liquid- cavity is 8 times 

 the size of the vacuity, two cases in which it is 7 times, one case in 

 which it is 6% times, six cases in which it is 6 times, one in which 

 it is 5|- times, and two in which it is 5 times. Now the mean of 

 these is -164 ; but probably the measurements *200 and *180 should 

 be discarded, seeing that the majority are considerably below them, 

 and the error is more likely to be on this side than the other ; this 

 would reduce the mean to *155. 



The following examples, Nos. 3 and 4, are from slices of Skiddaw 

 Granite, the specimens being taken from spots two miles apart; — 



