600 
the gyroceran bend in this shell is sufficient to account for the dor- 
sal furrow in the nephritic outline of the paranepionic substage. 
In the preceding remarks I have dealt solely with those genetic 
series and forms in which the dorsal furrow appears, but there are 
many in which there is no sign of any furrow. 
As has been said above, all of the straight and arcuate forms and 
the gyroceran shells, in none of which has either a dorsal or a con- 
tact furrow been found, except in Cranoceras. Passing these by, 
one comes to the nautilian shells which are transitional between the 
gyroceran and nautilian having the whorls in closer connection than 
in the gyroceran and a slight contact furrow. ‘These, so far as 
known, have other correlative characters. The umbilical perfora- 
tions are large and open, the apex of the conch is often free, the 
contact taking place on the venter of the ana- or metanepionic sub- 
stage or later, and the whorls are subsequently never involute, or in 
other words, they are discoidal shells. 
In these shells I have in Paleozoic time found no exception to the 
rule, that the dorsum of the nepionic stage is convex, and there is 
no dorsal furrow, a furrow being formed only after contact, and in 
later stages of development... One of the best examples of these 
series is that of Barrandeoceras. Barrandeoceras Sternbergi contains, 
as has been stated in the descriptions, two distinct forms, one Bar- 
randeoceras Sternbergt, Pl. xiv, Fig. 3, has the whorls approximate,,. 
and in the other they are not in contact. The purely gyroceran 
character of these shells is apparent in the loosely uncoiled 
young as well as in the later stages. They are also valuable 
in showing that the flattening of the dorsum and a dorsal lobe may 
arise as in Pl. xiv, Fig. 5, independently of contact, and this and 
the form of the ephebic stage is precisely similar to that of the para- 
nepionic volution of Barrandeoceras tyrannum. In this last and in 
Barrandeoceras Sacheri and Bohemicum, there is no dorsal furrow, 
but these ephebic characters of Sternberg? are repeated and a contact 
furrow is produced after the whorls touch. 
Estonioceras is another series in which nearly all degrees of coil- 
ing can be studied, and here also the absence of a dorsal furrow in 
the nepionic stage isa marked characteristic. The contact furrow is 
maintained as long as the whorls are held together, showing pro- 
gressive growth, but this rapidly disappears in the gerontic stage, as 
shown by the figures given on Pl. v and vii. 
Remeléceras impressum is also a good example of the generation of 
