312 BULLETIN OF THE BUREAU OF FISHERIES. 



They then increase in number as the periphery is approached, the radii in most cases 

 appearing late, beginning at the lamina edges. This is explained by the supporting 

 effect of each lamina on the next following lamina. On the under side of the scale 

 the radii do not appear, for the lower or uncalcified layer is flexible and does not break 

 until it becomes calcified, when it leaves a fissure showing, in stained scales, the uncalci- 

 fied, red-stained layer below. 



All the other factors being constant, then, the radii might be expected to increase 

 in a definite proportion as the periphery is approached. The number of radii varies 

 with (i) the activity of the fish, (2) the size of the scale, (3) its shape, (4) its thick- 

 ness, (5) its degree of calcification, (6) its curvature, and (7) the position of the fulcrum 

 of the scale. 



(i) Variations of activity may mean relative activity of the different parts of the 

 body, or of the same parts of the body at different seasons of the life of the fish. By 

 examining scales from all parts of the body of the same fish (pi. lii, Liti) it will be seen 

 that there are no radii on the inflexible parts ; on the very slightly movable parts very few 

 are found ; and on very movable parts the whole anterior field is sculptured with radii ; 

 but on certain scales, symmetrical in shape, and on flexible parts of the body, the num- 

 ber is found to increase to a certain extent as the periphery is approached, afterwards 

 diminishing, until there are no more radii at the periphery than at the focus (pi. nv, fig. 

 15). This is explained by the relative activity of the fish at different seasons. If this 

 explanation is correct we have an index of the relative activity of the fish throughout life. 



(2) As the scale increases in size the number of radii must increase proportionately 

 if the extent of bodily movement remains constant ; but if the radii are found to increase 

 in number to a certain point, then remain the same in number where an increase would 

 be expected, or decrease, and if we assume that the radii are caused by bodily move- 

 ment, the probability is that the fish suffered a diminution in activity at this point. The 

 number of radii at the several annuli on the scales of forty specimens were counted and 

 tabulated (table 5), showing that the expected increase does not occur on Cynoscion 

 rcgalis. Plate Liv, figure 15, shows a scale on which the radii thus decrease in number 

 after the third year. 



(3) Narrow scales have fewer radii than broad ones, the reasons for which are 

 obvious. Rachyceniron canadus has long narrow scales with very few radii ; Paralichthys 

 albiguttus has scales of a similar shape with one or two radii. The scales of Istiophorus 

 nigricans, the extreme of this shape, have no radii at all. Scales around the vent of 

 Brevooriia tyranmis or Cynoscion regalis are long and narrow, and have very few radii. 

 On a scale taken from the top of the peduncle of Cynoscion regalis, one of the anterior 

 angles was prolonged, and there were many more radii on the side of the prolonged angle 

 than on the opposite side. Shape also determines the relative direction of the radii. 

 When the anterior angles are both prolonged, the radii are seen to be parallel and not diver- 

 gent, as usual. (PI. Lii, fig. 7.) The scales of Fundidus majalis normally possess this 

 character, and here the radii are uniformly parallel. 



(4) Thin scales have fewer radii than thicker ones. On the scale of Urophycis 

 earlli"' no radii were found, and the scales were extremely thin, although large. Thick- 

 ness, however, varies little in the scales of the same fish, and, so far as the writer has 

 found, in the same species. 



^ This is an interesting scale bearing evidence relating to Baudelot's theory of spines (sec Review of Literature). 



