306 BULLETIN OF THE BUREAU OF FISHERIES. 



strated on the scales of the Clupeidae. On the scales of Brevoortia tyrannus and Pomo- 

 lobus mediocris and others of the Clupeidae the annuli cross the circuli at more or less 

 acute angles — laterally at almost right angles — the annuli being coincident in direction 

 with the scale contour while the circuli are arcs of concentric circles whose center is 

 posterior to the scale, and are not coincident in direction with the contour. (PI. lvii, 

 fig. 22.) In these cases there seems to be no more than an accidental relation between 

 the annuli and the circuli. It may be seen that this is also true for C. rcgalis in plate lvi, 

 figure 19. Whether the annuli in these different genera are homologous characters is 

 open to question, but their number, disposition, etc., suggest that they are. 



3. On the scales of Cynoscion regalis the number of circuli between the last annulus 

 and the periphery, is, in July and August " much less than half the number of circuli 

 between any two adjacent annuli. The number of circuli between any two annuli is 

 from 30 to 100. Between the last annulus and the periphery the number of circuli 

 varied from 4 to 8 in July and August. 



4. Measurements calculated from the annuli considered as summer bands agree 

 with the length groups actually measured. The following method was employed: 

 Three hundred and eighty-two specimens were measured at random.* Their lengths 

 were found to fall in modes of 19.91, 26.31, etc., cm. (Tables i, 2, 3.) Then the 

 lengths of 28 specimens were divided into parts proportionate to the distances between 

 the several annuli, and these lengths entered as the respective first, second, etc., years 

 of the fish. (Table i.) The averages of these lengths were then compared with the 

 modal lengths of the 382 measured fish. (Table 2.) It will be seen that the averages 

 agree remarkably. 



5. Annuli are narrow areas parallel with the contour of the scale, in which the 

 regularity of the circuli is interrupted, manifested as branches, breaks, or terminations. 



6. The scale is separable into laminae, the edges of which coincide with the annuli. 



7. Annuli stain pink with picrocarmine. 



8. Annuli have a refractive index different from that of the spaces between. "^ So 

 far as the writer has been able to determine, the refractive index of scales has never 

 been actually measured. 



The conclusion of previous investigators that annuli are approximations of circuli 

 and are caused by retarded growth is rendered questionable by the foregoing observa- 

 tions. If the annuli were approximations of circuli, the expected curve would be the 

 dotted lines in figure i, showing retarded growth at the time the annuli were formed. 

 In the second place, if retardation of growth brought the circuli closer together, then in 

 the fifth or sixth year of the life of the fish, when growth is much slower than in the 

 earlier years (indicated by the narrowness of the bands), the circuli would be closer 

 together, giving the scale the appearance to the unaided eye, or under low magnification, 

 of having at the center widelj' separated circuli while, approaching the periphery, the cir- 

 culi would appear closer together. The scale would then have a light inner part, growing 

 darker toward the periphery. The writer did not find this to be the case on the scales 



o- This character is much less constant on the scale of the pigfish, Orthopristis chrysopterus, Linnsus, due, perhaps, to the 

 more variable spawning time of the latter fish. Specimens 2 cm. long were taken as early as June 15 and as late as Sept, i, 1913. 



6 The detailed measurements in the table cover only 65 specimens. However, in addition to these the lengths of 382 fishes 

 measured by Hccht and Crozier were used in constructing the modes. 



c Carlet (1878): DahUi9io). 



