Hyatt, j 274 [April 4, 



chamber. The siphon becomes approximately reduced and the 

 rosettes begin to be variable with age, and finally altogether dis- 

 appear in the adults of extreme forms. Sac. (Orth.) docens, sp. 

 Barr. pi. 250, is a transition form, but we place it in this 

 genus because at an age, when an Actinoceras would have the 

 rosettes large and perfect, this species begins to lose them, and the 

 siphon decreases also. The reduction of the siphon is a degra- 

 dational senile shrinkage, and it occasions the loss of the rosettes. 

 M. Barrande views this old stage of the siphon as a return to 

 the tubular siphon, but in our opinion we cannot call this a 

 tubular siphon. As a matter of fact it is a modified nummuloidal 

 siphon, as may be seen by comparison with others. Sac. (Orth.) 

 Richteri, sp. Barr. is selected as the type and in the beautiful 

 figures of M. Barrande we may read on plates 318, 322, 323, 349, 

 that the young have an empty nummuloidal siphon, and that the 

 adults have the usual imperfect rosettes of this genus, and that 

 in the old these disappear again leaving the siphon empty. 

 M. Barrande's species with mixed elements, i. e., siphon on one 

 side tubular and on the other nummuloidal are simply species 

 of various groups with imperfectly developed siphon, or unsym- 

 metrical anomalies of development. This genus is well repre- 

 sented in the Silurian, Devonian, and Carboniferous. 



Tretoceras, Salt. Journ. Geol. Soc. Lond. Vol. 14, p. 179, 

 has according to that author, Blake's British Cephalopods, and 

 Barrande, conical prolongations, and a siphon which appears 

 in Blake's figures to be microchoanitic. The cones compare 

 closely only with Kayser's deformed Gomphoceras, " Missb. Devon. 

 Gomph., Zeit. Deutsch. Geol. Gesell. Vol. 26, pi. 16. There is a 

 similar central trace figured by Barrande on the casts of Bath- 

 moceras, but reversed in position. 



Orthoceratidae. 

 This family includes longicones with tubular siphons, and septa 

 widely separated. We do not regard the Actinoceratidae as the 

 ancestors of the tubular siphoned Orthoceratidae ; but on the 

 contrary, the Orthoceratidae as the normal form and the probable 

 ancestral type. All the nummuloidal siphons are tubular in the 

 early stages. M. Barrande in Syst. Sil. Vol. 2, Text 3, p. 748, has 

 shown conclusively the passage of the Sactoceran forms into this 



