MancJiestcr Memoirs, Vol. hi. (19 12), No. S. 9 



position (2), the total number of annuli on the long axis 

 varies from 96-101, at 30"^ to the left of the long axis 

 from 97-107, and at 30" to the right of the long 

 axis from 89-101. The extent of variation in the 

 length of the scales on the long axis is -61 mm. Taking 

 another example in the scales from position (3), the 

 total number of annuli varies from 105-118 on the long 

 axis and from 109-118 on the left of the long axis, 

 and from 108-122 on the right of the long axis, while 

 the maximum and minimum lengths of the scales differ 

 as much as 103 mm. 



Turning again to the examination of the averages of 

 the results, it is to be noticed that the total number of 

 annuli on the scales, with few exceptions, increases from 

 position (i) to (3), and decreases from (3) to (4) — that is, 

 from the head to the adipose fin, and from the adipose fin 

 to the tail. The Table indicates that a similar variation 

 is found also in the total lengths of the scales and in the 

 lengths per period. As has been previously pointed out, 

 this variation occurs in each peronidium and not in any 

 particular one. The scales taken from the ventral side 

 of the fish also have an increasing number of annuli per 

 period, from the head to a position on a level with the 

 adipose fin ; and in the same way this increase is in each 

 period, and the length of each period is greater from the head 

 to the adipose fin. There are irregularities in some cases, 

 but it can be said that there is some definite arrangement. 



A further comparison of the results obtained indicates 

 that there is another method in which the scales are 

 arranged. It will be noticed that positions (i), (2) and (3) 

 are exactly above positions (5), (6) and (7) ; and comparing 

 the figures obtained from positions at the same distance 

 irom the head, it is clear that in nearly all cases the scales 

 on the ventral side of the fish are longer and have more 



