FOYAITE-IJOLITE SERIES OF MAGNET COVE 649 



because, owing to the elliptical shape of the area, a specimen 

 {e. g., covite) from near the end of the major axis may be at an 

 actually greater distance from the center than one {e.g., foyaite) 

 from the end of the minor axis, though genetically inside the 

 latter. 



It would, of course be best to have several analyses of each 

 type from different parts of the zones, both radially and circu- 

 larly, so as to get the mean composition of each. But as that 

 involves the making of very many analyses, we must be content 

 at present with selecting what seem to be representative speci- 

 mens, and assume that their analyses correspond to the average 

 composition of each type. 



Assuming this, two courses are open to us. We can either 

 measure the distances from the center along a radial line on 

 which all occur, or average the distances of the various occur- 

 rences of each type. The latter has been the process adopted 

 here, since it seemed more likely to eliminate errors due to local 

 conditions. 



For each type measurements were made on Williams's map in 

 many directions from the Baptist church to the middle point of 

 each zone exposed, and the mean of these taken in each case. 

 On the diagrams the abscissal positions of III, IV, V and VI, 

 from the origin at II represent these relatively, as it is not nec- 

 essary that the diagrams be of the same scale as the map. The 

 position of the foyaite (VI) is not fixed as accurately as those of 

 the others, since, being at the periphery, it is in great part over- 

 laid by the surrounding shales. Small outcrops outside the 

 main area, however, allow a rough estimate of its average dis- 

 tance, though it undoubtedly extends farther away from the 

 exposed area than the few outcrops indicate. 



In my former paper I assumed that the petrographic center 

 of the area was in the " magnetite bed," and that this was the 

 result of the decomposition of underlying jacupirangite. As» 

 however, this is quite uncertain, it seems best for the purpose in 

 view to disregard this area. For the present then the anal3^sis 

 (and the diagram position) of the jacupirangite may be neg- 



