124 
BULIvETIN OF THE BUREAU OF FISHERIES. 
quite regularly interrupted; in Clupea also it is common to find a pair of irregular apical radii, joining 
basally to form a sort of U. In Clupea the lateral circuli, following the trend of the transverse radii, 
reach the margin obliquely, not at right angles as in Clupanodon; this seems to be in some degree a 
matter of age, yet the large scales from Woods Hole, in which the obliquity of the circuli is very strongly 
marked, show growth lines delimiting various younger stages, and when less than half grown the circuli 
were already moderately oblique. It is a curious tiling that the scales from Sandy Island (Bureau of 
Fisheries) have the transverse radii more numerous, much less regular, and interrupted at intervals, 
the circuli more transverse, in the young verj'- little oblique, and the lower limit of the hyaline apical 
area straight or almost, whereas in the Woods Hole scales it is concave or even broad V-shaped. Are 
there two types of herrings, or is this individual variation ? 
The large scale of the shad, Alosa sapidissima, has been figured in Smithsonian Miscellaneous Col- 
lections, volume Lvi, no. i, page 2, and discussed at length in the Proceedings of the Biological Society 
of Washington, volume xXiii, pages 6i-fi2. It is shown that the transverse radii are derived from the 
apical radial system, all stages of transition being visible. The adult Alosa scale has the circuli and 
radii at the sides very oblique; in some scales the radial system is very highly developed, with very 
numerous apical radii, and those about the middle of the scale variously anastomosing, forming an 
irregular network. There is, however, no approach to the type of network seen in some Engraulididae 
and better developed in Osteoglossidas. So far as the scale goes, Alosa must stand at the base of 
Fig. 4 . — Brevooriia tyramms (Clupeidae). Apical teeth. Bureau of Fisheries. 
the clupeine series, with Clupea close to it but a little more advanced. Clupanodon then represents 
a considerable further step in the direction of clupeine specialization, with the fixation of the more 
characteristic features. 
In Sardinella humeralis (Cuvier & Valenciennes), from Tampa, Fla., the scales differ conspicu- 
ously from Clupea, Clupanodon, Alosa, etc., in being much broader than long, the lower comers obtusely 
almost rectangular. The circuli are strictly transverse, not oblique; there are two widely separated 
entire transverse radii, and part of a third one forming a small reversed broad V on the lower margin; 
the lower margin of the hyaline area is slightly concave. These scales are perhaps not adult; they 
resemble in many ways the immature scales of Alosa. 
The scales of the alewives, Pomolobus, have the lateral circuli moderately oblique, except in the 
yoimg; the undulating transverse radii distinctly rather widel)^ spaced, often broken in the middle 
but not regularly or widely so, as in Clupanodon. Tn^o species before me are rather easily distinguished : 
Pomolobus cestivalis (Mitchill). Glut herring. Scales about as long as broad; lower margin of 
hyaline area concave. 
Pomolobus pseudoharengus (Wilson). Branch herring. Scales conspicuously broader than long; 
lower margin of hyaline area straight. In a former paper I said I did not know how to distin- 
guish yoimg scales of P. pseudoharengus from young of Alosa sapidissima. However, in the 
A losa the lower margin of the hyaline area is uniformly distinctly concave ; in P. pseudoharen- 
gus it is straight or even convex, except in the very earliest stage, when the diameter is much 
less than a millimeter. Reviewing the Clupeinae, it appears that we may place Alosa at the 
base of the series, and just above it Clupea. Then Clupanodon stands at the end of a branch, 
from the side of which springs Opisthonema, leading directly toward Brevoortia. From an earlier 
part of the same branch may arise Pomolobus, and somewhere near here Sardinella. All this 
is based on the scales, and must of course be modified when the structures are considered. 
