AMM0DYTID7E. 
577 
this relative length decreases with increasing age, and 
in this respect Am. lanceolatus thus retains a juvenile 
character. Nor can the character be employed as a 
specific distinction, for in a young specimen of the Sand- 
Launce 56 mm. long, taken by Fries in Bohusktn, the 
length of the head is nearly 21 % of that of the body. 
In the same specimen the length of the lower jaw is 
9‘3 % of that of the body, while in other cases the 
boundary between the species in this respect lies at 9 %. 
The length of the pectoral fins, where the line of di- 
stinction is the same, but the direction of distinction 
reversed — the percentage less in Am. Icmceolatus than 
in Am. tobianus — in a specimen of Am. tobianus 20 
cm. long is only 9'1 % of the length of the body, and 
in a specimen of Am. lanceolatus 11 cm. long 9 % of 
the same length. In this respect, too, the percentages 
decrease with increasing age, and in this respect Am. 
tobianus thus retains a juvenile character. In his de- 
scription of Ammodytes americanus “ Storer says the 
dorsal fin begins just at the end of the pectoral fin, 
and that the length of the latter fin is only 1 / 3 of that 
of the head. It was b} 7 this statement that I was in- 
duced to refer the Sand-Eels brought home by the Vega 
Expedition from Pitlekaj, north-west of Behring Strait, 
to Am. lancea, var. americancd, for in a specimen 77 
mm. long the beginning of the dorsal fin lay at a 
distance from the tip of the snout measuring 29‘3 % of 
the length of the body, and extremely little in front 
of the tip of the pectoral fin, though this fin measured 
1 1 % of the length of the body. Later investigations, 
however, especially those of Jordan and Gilbert, show 
that Storer’s description and figure must have been 
based on some exceptional specimen, for, unless this 
were so, such a character could not have escaped ob- 
servation. Still, this shows that even the characters 
most important in a systematic respect are subject to 
variation. According to Storer Am. americanus may 
also attain a size of 3 dm., though from a specimen 
which the Royal Museum has received through the 
Smithsonian Institution from Woods Hole (Mass.), it 
appears to be quite identical with Am. tobianus. The 
other two American species that have been ranged be- 
side Dekay’s Ammodytes americanus c , but which are 
regarded by Jordan and Gilbert merely as varieties 
of this species, show among their characters a variation 
of the transverse dermal folds on the sides of the body 
between 130 and 182, thus filling the gap between the 
normal numbers in Am. tobianus and Am. lanceolatus , 
and showing how the form-differentiation may bring 
about a resemblance to Am. lanceolatus, without altera- 
tion in the structure of the mouth and without the 
development of teeth on the head of the vomer. The 
development of the said dermal folds, the absence of 
which is one of the most important characters of Am. 
cicerelus, in Am. tobianus is a character of growth ex- 
tremely irregular in its appearance. Even in specimens 
7 cm. long it is sometimes impossible to discover them. 
Of the other character which should serve to distinguish 
Am. cicerelus, the undulating margins of the dorsal and 
anal fins, I have found at least a trace in a slight con- 
cavity of the margin of the anal fin in a young Sand- 
Launce. All this goes to show, not only that all the 
species are extremely closely related — so closely that 
we may well be tempted to regard them merely as 
varieties of the same species, or as species at the be- 
ginning of their differentiation from each other — but also 
that this relationship has its origin in a form essentially 
resembling Am. tobianus , or perhaps in this very species. 
At least one of the species described above, the 
Sand-Launce, must be regarded as circumpolar. The 
Royal Museum possesses examples of this species from 
Norwegian Firnnark, the Murman Coast, the White Sea, 
North-East Siberia, Greenland, and Iceland. In Spits- 
bergen, however, it is unknown. According to Machado 
(quoted by Steindaciiner) the range of this species 
extends southward to Cadiz. Day assumes that it may 
occasionally wander into the Mediterranean; but he gives 
no observation of its occurrence there. In the Baltic 
the Sand-Launce penetrates at least to the island-belt 
of Stockholm and, according to Mela, to the islands 
o o 
round Aland and Abo and into the Gulf of Finland 
up to the island of Hogland. In the first locality Sun- 
devall could not find it; but during the investigations 
which I caused to be made this summer (August, 1890) 
a Mem. Amer. Acad. Arts, Sc., n. ser., vol. VIII, p. 411, pi. XXXIII, fig. 2. 
b Gt. Intern. Fisher. Exbib. London 1883, Swed. Spec. CataL, p. 170. 
c Girard’s Ammodytes personatus and Cork’s Am. alascanus. 
