218 



Fart III. — Eighteenth Annual Report 



length, bat as this decreases with increase of body-length, the observed 

 decrease in skull-breadth relative to skull-length must be all the greater 

 when considered relative to the body-length. If the groups of mature 

 specimens alone be considered, it is seen that the probability of the 

 observed decrease being real is much less than it was in the case of the 

 skull-length. In the first two cases — those from the southerly North 

 Sea — the probability is about 19 to 1 in the first case and about 4 to 1 

 in the second in favour of the decrease. In the third case — the group of 

 male specimens from Aberdeen — the probability is again about 4 to 1. 

 The amounts of this decrease for the mature specimens are respectively 

 *73, '42, and *60 in the three cases, which gives '58 of a decrease on the 

 average. If this amount of decrease, which is relative to skull-length, be 

 added to the decrease of skull-length relative to the body-length, *75 

 per cent., and divided by 2 as before, then we find that an approximate 

 decrease of *66 per cent, takes place in the skull-breadth relative to the 

 body-length as the total length of the fish increases by 100mm. 



If the total length of the fish, including tail and skull-length, had been 

 used as the standard and not the body-length, this result, along with that 

 showing the relative decrease in skull-length, would then perhaps have 

 been easier to grasp. A short and simple calculation from the values 

 given will show approximately this decrease relative to the total length of 

 the fish. The length of the tail relative to the body length has been 

 shown to be constant during growth, and practically 30°/ 00 . The 

 relative length of the skull at 300mm. is — for the southerly JN r orth Sea 

 specimen — practically 25°/ 0 o- Hence, since the body-length has been 

 taken as 100, the total length of the fish is 155. This gives the skull- 

 length as 16*l°/ ()0 of the total length of the fish at 300mm. At 

 400mm., again, the relative tail-length has not altered, but the relative 

 skull-length, for the same groups, has become approximately 24°/ 00 . 

 Hence, the total length of the fish is now represented by 154, and the 

 skull-length relative to this is 15"6 0 / 00 . The decrease in skull-length, 

 therefore, is '5°/ 00 where the increase of total length is 100mm., or, in 

 other words, at 400mm. the skull is 2mm., on an average, shorter than 

 it would have been if there had been no growth-variability. If a similar 

 calculation is gone through for the skull-breadth, it is seen that , 45 0 / 00 

 of a decrease occurs, and hence that the skull-breadth is narrower at 

 400mm. by l*8mm. than it would have been if there had been no growth- 

 variability. 



To return again to the Table, a comparison may be drawn between the 

 immature and mature specimens at the same size. It will be remarked 

 that there is a decrease in each case, but the probability that this observed 

 decrease is real is only 4 to 1. The amount of the decrease is, on the 

 average, -42 0 / O0 . As was remarked in the case of the former dimen- 

 sions, maturity, or the ripening of the reproductive organs, seems to have 

 little effect during the first period on the structures observed. It might 

 have been thought that the maturity of the reproductive organs would 

 lead to the withdrawal of nourishment from other parts. This is pro- 

 bably the case, but from what has just been shown its effect seems to be 

 less during the first than during the second period. The difference 

 between the immature specimens with average at 300mm. and the 

 mature specimens with average at 400mm. is, on the average, r00°/ 00 , 

 whilst between the former and the mature specimens of the same size it 

 is only -42°/ 00 . 



Further remarks on the relation of "maturity" to "growth-varia- 

 bility" will be made at the end of next section. 



Skull-depth. — This dimension was measured from the commencement 

 of the supraoccipital spine to the middle of the base of the basioccipital 



