SCALE OBSERVATIONS OF THE SQUETEAGUE AND PIGFISH. 305 



When viewed under low magnification, these alternate bands appear, in most cases, 

 clearly; and if they really represent winters, it is a simple matter to determine age by 

 counting them. Hoffbauer seems to have demonstrated that these zones do represent 

 winters by observations on the scales of carp kept in aquaria and under known conditions. 



One is, however, confronted with many obstacles in relying entirely on this means 

 of age determination. In very old fishes, as Tims points out, the annuli, through wear 

 and diminished growth, become so indistinct and close together that it is almost, if not 

 quite, impossible to arrive at a satisfactory conclusion as to age. 



Investigating the scales of Cynoscion regalis and Orthopristis chrysopterus, the writer 

 has not been able to verify all these observations, for the reason, possibly, that these inves- 

 tigations were made on English brook trout, salmon, and carp, the scales of which he has 

 not had the opportunity to examine. 



Miss Esdaile (191 2) in the following language accepts and states very clearly the 

 fundamental assumption of all the workers on age determination: 



Examination * * * shows that the annuli [circuli] are arranged in a definite manner, some 

 far apart and others closer together. Those far apart are, according to Mr. Johnston, formed during the 

 rapid growth of the fish in the summer, and those closer together during a time of slow increase in the 

 winter. 



This implies clearly that bands representing equal lengths of time ought to exhibit 

 at least approximately equal numbers of circuli, and that scales of the same size ought 

 to be sculptured with a similar number of circuli. But later in her paper she states 

 that "there is no constant variation in the number of annuli in the different periods of 

 the scales from the same position." In her table no. 2, scales from positions 4 and 5 

 have the third peronidia of the same width — 0.48 mm. — yet one has 10 circuli, the other 

 7.8 circuli on the long axis. Again, scales from positions 6 and 7 have total lengths of 

 6.07 and 6.09 mm., respectively; yet one has a total of 107.8 circuli, the other of 118.2 

 circuli, a difference of 10.4 circuli, or 9.2 per cent variation from the mean. In no case 

 where the widths of any two peronidia were the same, were the corresponding numbers 

 of circuli identical. 



A similar criticism can be made from the photographs in Gilbert's paper (1913); 

 take, for example, plate vi, figure 10, of his paper: Judging from the width of the 

 summer bands, the growth each year is less than that of the preceding, yet the sepa- 

 ration of the circuli is greater each year than that of the preceding. There are more 

 circuli per linear centimeter in the third band than in the fourth, yet the former is 

 wider and is supposed to have grown more rapidly. 



The following observations bear on the nature of the annuli: 



1. The circuli on the scales of the C. regalis are almost equidistant. Figure i is a 

 graph correlating the number of circuli and distance apart measured in tenths of a mil- 

 limeter. Each ordinate unit represents one circulus and the units of the abcissa are 

 tenths of a millimeter. If the annuli are uniformly one-tenth millimeter apart a 45° 

 straight line would result. Barring very small fluctuations, this is true. Their sep- 

 aration does not vary in the vicinity of the annuli, nor does their separation vary 

 with different distances from the periphery. 



2. The direction of the annuli is not necessarily coincident with that of the circuli. 

 This is more or less apparent on all the scales examined but is most strikingly demon- 



