158 University of California Publications in Zoology [Vol. 13 



the true nature of these scales would be surmised only. That the 

 triangular scales (fig. E) are immature scales of the Greek-cross type 

 is shown by the fact that in the adult all of the scales in this area are 

 of the Greek-cross form. The description, then, which we have just 

 made of this series becomes important in showing the manner in 

 which the Greek cross reaches its complexity of form. 



VI. The Form of the Caudal Scale 



Scales approaching the tip of the caudal region, while dissimilar 

 to the fully formed Greek cross, closely resemble an immature scale 

 of that type. They are unlike the latter, hoM^ever, in having the spines 

 longer and sharper and in having the lateral wings or arms narrower. 



From such a scale on an immature specimen, one would be inclined 

 to conclude that the caudal scales are like the immature scales just 

 studied, and that in time they will reach the Greek-cross form. That 

 such is not the case is shown by the fact that in the adult the scales 

 of the caudal region never attain the Greek-cross form. In the adult, 

 while they thicken up somewhat, they remain as modifications of the 

 ventral type. 



VII. Dorso-Ventral Transitional Areas 



In our study of the scales ventral to the pectoral fin we observed 

 that they grade more or less gradually into the dorsal type on the top 

 of the fin (pi. 8, fig. 4). The question arises, are the two types, which 

 we have characterized as ventral and dorsal, in reality distinct types, 

 or is one a modification of the other? To decide this we may study 

 more in detail transitional areas between the two types. 



In such a study, a strip of the integument may be taken from the 

 mid-dorsal line running over the tip of the nose (pi. 9, fig. 8). In 

 this strip the dorsal scales are typical and become modified only at 

 the place where the surface bends sharply downward. At this place 

 there is a narrow transitional area three or four scales in width ; beyond 

 this are scales which are modifications of the ventral type. 



The transition is here made so abruptly that the distinction be- 

 tween ventrals and dorsals is seen within narrow limits. At the 

 beginning of the transition the scales are clearly modifications of the 

 Greek-cross type, while those following are seemingly modifications 

 of the spade-shaped type. In a word, the two types appear to be 

 distinct. But the area of transition is here exposed, and consequently 



