OBSERVATIONS ON FISH SCALES. 
123 
scale is indeed curiously similar in form and structure to that of Synodus, though the nucleus is not 
so far apidad. The region above the nucleus is minutely roughened or tuberculate, and the apical 
circuli are longitudinal. The basal circuli are finelj^ beaded. 
HIODONTIDjF. Toothed herrings. 
The scales of the moon-eye Hiodon iergisus are discussed in Smithsonian Miscellaneous Collections, 
volume 56, no. 3, page 3. They have much in common with Elops, but the wholly transverse fine 
circuli of the apical field are not at all modified above the nucleus. The basal radii are numerous and 
irregular, inclined to be divided into two groups, and the basal margin is undulate, not scalloped. The 
basal circuli are without distinct beading. The scale is of a cyprinoid type. 
CHIROCENTRID.®. Long herrings. 
The long herring, or dorab, Chirocentrus dorab (Forskal), from the Philippine Islands, has transverse 
circuli and radii, es^ntially as in Clupeidae, the radii usually angled in the middle. There are also 
irregular basal longitudinal radii, few in number, much as in the anchovy Stolephorus argyrophanus. 
On page 113 of volume xxm of the Proceedings of the Biological Society, Washington, for "Jive trans- 
verse circuli” read "jine transverse circuli,” and it is the basal, not the apical field, which has these 
more widely spaced circuli. Upon minute comparison I find that the whole arrangement, both circuli 
and radii, agrees essentially with Stolephorus argyrophanus, and hence in its scales Chirocentrus is to be 
compared with the Engraulidae rather than with the Clupeidae. The scale is, in fact, wholly as in the 
engraulids. 
ENGRAULIDIDjF. Anchovies. 
I alter Jordan’s arrangement by placing these before the Clupeid®, as being more primitive, at 
least as to the scales, and also nearer the chirocentrids. Dr. Max Ellis has prepared a paper fully dis- 
cussing the available material. Some of the species show an evident reticular network, in one case so 
well developed that the scale looks like that of some osteoglossid. 
CLUPEID^. Herrings. 
I here consider only the Clupeinae, of which I have examined nearly all the American genera. 
Two tribes are indicated, Brevoortiini, for Brevoortia, and Clupeini, for the other genera. In the men- 
haden, Brervoortia tyrannus (Latrobe), the apical margin of the scale is produced into long parallel teeth 
with very slender ends. These teeth arise from a pellucid apical zone, and the intervals between them 
are prolonged basad as grooves for a distance about equal to the length of the teeth. Below these grooves 
are numerous small and short grooves looking like roots, the large grooves seeming to be the stems from 
which they arise. Below this the scale is entirely covered with very fine transverse circuli, and has 
in addition irregular pits and two transverse radii. The dentate scale margin is by no means peculiar 
to Brevoortia; it is quite distinct, for example in Alosa and Clupea, but the teeth are very irregular 
and by no means so long or tapering. In juvenile Alosa the margin is not dentate. The pitting seen 
in Brevoortia is also very strongly developed in some scales of the thread herring, Opisthonema oglinum 
(Ue Sueur), from Woods Hole; it is evidently derived from evanescent transverse radii. In some scales 
of O. oglinum there are five transverse radii, all but the uppermost broken in the middle; there is then 
no pitting, but the well-pitted scales have only the upper radius. In the pilchard, Clupanodon pil- 
chardus , from Palermo, Italy (Bureau of Fisheries), there are seven or eight transverse radii, all widely 
broken in the middle except the first, the lower ones oblique, the whole reminding one of the ribs of 
the human skeleton seen from in front. The extremely fine circuli are strictly transverse, meeting the 
lateral margins at right angles. I find essentially the same in the West Indian sardine, Clupanodon 
pseudohispanicits Poey, from Woods Hole (pi. xxxv, fig. 18), but in two scales I notice that the 
circuli cmwe upward on one side only. 
Ojmparing Clupanodon with the sea herring, Clupea harengus (from Woods Hole and Sandy Island), 
two differences are at once apparent. In Clupea the transverse radii are commonly (though not uni- 
formly) entire, or not interrupted in the middle, whereas in Clupanodon they are (except the first) 
