On THE STRUCTURE AND CLASSIFICATION OF THE TrEMATASPIDAE. 
21 
although the two post-orbital daslies, and the tliree pairs in tlie anterior transverse line, 
are absolutely constant; at least they were so in the eight or ten specimens perfect enough 
to show the liues clearly. 
The Circum-orbital Lines * *) consist of two semicircular Unes of dots and daslies almost 
completely surrounding the frontal eminence. The first dash is invariably a long one, lying 
just behind, and médian to, the post-orbital tubercle. Then follows a rather wide interval 
till the line begins again with a well defioed row of closely set dots and dashcs extending 
in a semicircle toward a point in front of the orbits. Неге the line forms a slight loop 
forwards and then bending backwards, maybe continued in some specimens by two small dashes 
straight across the median line. lu old specimens, ail the dashes of the same side, except 
the posterior pair, may unité to form a well defined continuons groove. In such cases, 
however, the anterior ends of each line do not become continuous across the median plane. 
Jn youug specimens, the circumorbital lines may consist of a beautifully regulär sériés 
of dots and long and short dashes. 
The Marginal Line begins with a transverse dash nearly opposite the middle of the 
post-orbital opening. The line then tends outwards and backwards in a graceful curve that 
extends the whole length of the shield. The dashes in this line are long and fairly constant 
in number and regulär in size. There is a slight change of curvature, and an increased 
space between the dashes, where this line is crossed by the anterior transverse one. I hâve 
counted five dashes in front of this interruption, and seven or eight back of it. 
The Anterior Transverse Lines. There are invariably three long dashes in each of 
these lines. They begin a little behind, and to one side of the small circular openings in 
the Shell (anterior entapophyses) and extend outwards and slightly backwards. The first two 
dashes lie on the median side of the marginal line, and the third near the posterior median 
side of the posterior lateral opening. 
The Posterial Dorsal Line begins in front of the crista occipitalis either as an irre¬ 
gulär transverse line, or as a pair of slightly bent dashes, and is continued backwards, on 
either side of the crista occipitalis, by from two to four separate dashes. 
I hâve failed to find any trace of these dot and dash markings on the ventral shield. 
In the collection of the St. Petersburg Academy, there is a triangulär plate, probably 
belonging to the oral regioD, whose surface was marked by a single bent groove like those we 
hâve just described. 
Superficial markings of the shell, comparable with those of Tremataspis, probably 
occur in most of the Ostracoderms. Lankester has sliown in several figures of Pteraspis and 
1) The délimitation of the linos described is perfectly 
clear except just behind the orbits. I could find herc no 
certain indications as to wliethcr the two post-orbital 
* dashes reprcsent the médian ends of the circum-orbital, 
or of the marginal lines. I bave assumed, mainly for con- 
venience of description, the former to be the case, and 
hence regard the first two pairs of dashes back of the 
orbits as the posterior ends of the circum-orbital lines. 
