HORSE-MACKERELS. 
83 
ranx , apparently ending at the middle of the orbits or 
over their anterior part, but continued more or less 
distinctly by the marginal ridge of the frontal bones. 
The other, outer lateral ridge on the skull starts from 
the os squamosum, ( pteroticum ) and is generally con- 
lined to the posterior frontal bone (os sphenoticum), 
but sometimes extends to the middle of the orbits. 
Though these ridges are not wanting in the true Mac- 
kerels, they are much less highly developed and do 
not extend so far forward". This difference, however, 
like the number of the vertebrae, is an internal one 
and as a character is thus subject to practical difficul- 
ties in its employment for the determination of species 
belonging to this family. Furthermore, as it is rather 
a difference in the degree of development than in form 6 , 
its validity as a family character may be doubtful, 
though we here retain the Carangoid family in order 
particularly to distinguish the Horse-Mackerels in our 
Fauna as a connecting link between the Bramidce and 
the Mackerel group, a position clearly pointed out for 
them by the development of the above-mentioned, lon- 
gitudinal ridges on the skull. 
The correctness of this procedure is also shown 
1 >y the developmental changes due to age, which are 
better known in the Carangidce than in the Scombridce. 
The most detailed information on this point 0 is given 
by Lutken in his excellent work 1 *. It is a universal 
rule that in the Carangidce, as in the Scombridce, the 
margin of the preoperculum is furnished with teeth or 
spines during youth, there generally being a few large 
spines at the corner and smaller spines or teeth on 
each side of them 0 . In the Pilot-fish ( Natter ates ) si- 
milar spines appear not only on the preoperculum but 
also on the posttemporal bone, the supraorbital margin 
of the frontal bones and even the rostral region 7 . These 
juvenile marks bear witness to the original analogy 
between this family and the JBeryx type, as well as 
with the Cottidce , though the Carangidce are unques- 
tionably brought nearer to the latter by the compara- 
tively late appearance of the scales on the body, which 
in certain forms are completely wanting. The other 
changes due to age, as we have already seen in the 
Bramoids, consist in the elongation of the form of 
the body and of the pectoral fins — the latter becoming 
more and more pointed and directed more and more 
upwards with a more horizontal base — and also in 
the diminution of the ventral fins. We have also to 
notice in the Carangidce the reduction of the first dorsal 
fin and of the spinous rays in the anal, the changes 
in the latter causing some likeness to the family of the 
true Mackerels. In some Carangidce ( Selene vomer), as a 
compensation for this reduction, the anterior parts of 
the second dorsal and the anal fins are developed into 
high, falciform flaps and the pectoral fins are simulta- 
neously elongated. In others again (Selene setipinnis) 
age also causes a reduction in the length of the anterior 
parts of the dorsal and anal fins. 
In a family with so great changes of development 
the definition of the genera and species has naturally 
been difficult and unfixed, for the different stages of 
development, before their correct signification was known, 
were regarded even as separate genera. However, even 
after the corrections made by later writers on this 
point, the Carangoid family still contains a large num- 
ber of species. Jordan and Gilbert 51 estimate the 
number as high as 180, the great majority of which 
belong to the tropic seas of both hemispheres. Many 
are highly esteemed as an article of food, and as they 
a The difference is given by Gunther, though not as a family character, and may be found in his descriptions of the skeleton of 
Scomber scoi?ibrus {Cat., 1. c., p. 358) on the one hand, and Caranx trachurus {Cat., 1. c., p. 421), Chorinemus lysan {Cat., 1. c., p. 472) 
and Lichia glauca {Cat., 1. c., p. 478) on the other. 
b This is also the case in the large species of Tunnies, where these ridges are, it is true, comparatively much lower than in the 
Carangidce, but still high enough to cause it to be referred to as an exception when the supraocipital ridge in Orcynus tliynnus (see Cuv., 
Val., Hist. Nat. Poiss., VIII, p. G6) may be traced as far forward as the ethmoid bone. 
c There are several remarks on the subject, however, in Gill: Proc. Acad. Nat. Sc. Philad. 1862, pp. 430 and 440. 
d Spolia Atlantica, 1. c., pp. 504 — 552. 
e In an Indian species, Caranx ( Carangiclithys ) typus, these teeth are retained even by adult specimens, provided that Bleeker's de- 
scription be really based on examination of fullgrown specimens; and in Caranx Upturns from the West Indies, they occur as uniform teeth 
at the corner of the preoperculum, though they are scarcely visible except on the skeleton. On the removal of the skin both the suboper- 
culum and interoperculum also present distinct crenulations at the outer margin. 
f That this genus, too, should be regarded as representing the earlier developmental stages of the family type, appears from its colo- 
ration, from the dark transverse bands, the number of which diminishes with age, while in Caranx these bands are generally present, it is 
true, but only as a juvenile character. Caranx uraspis {Uraspis carangoides ), however, according to Bleeker, retains these transverse bands 
even when fullgrown. 
a Syn. Fish. N. Amer., Bull. U. S. Nat. Mus., No. 16, p. 431. 
