630 THE VOYAGE OF H.M.S. CHALLENGER. 



comparison of many hundred specimens of them, and of tlieii" dimensions, lias con\'inced 

 me that this was an error, and that the small spherical or elliptical medullary shell of 

 Tetrapyle and the other Diplozonaria possesses already the same complex structure, 

 composed of a system of three girdles, as Trizonmm and Lanmcilla. Whilst in 

 the Haplozonaria probably the simple central chamber only represents the medullar}' 

 shell (enclosed in the central capsule), and the first sj^stem of girdles (complete in 

 Trizonium) the external cortical shell, with the progessive growth this latter Ijecomes 

 enclosed in the central capsule and so constitutes the " trizoual medullary shell " of the 

 Diplozonaria and Triplozonaria. 



A very difficult matter is the mode of connection between the cortical and medullary 

 shell. In most of the Pylonida it seems that the first or transverse girdle (in each 

 system) is produced by the formation of two lateral wings or chambers (one on each 

 side of the medullary shell), so that each wing (or half girdle) represents a short and 

 wide, nearly cylindrical tube, the axis of which (with free openings on both poles) is 

 jaarallel to the principal axis of the medullary shell. In this case (probably the ordinary 

 one) both principal faces of the medullary shell itself (dorsal and ventral face) c-onstitute 

 the middle part of the first girdle, whilst its lateral parts are formed by the wings 

 (comparalile to the lateral chambers of Anvpliitliohis). 



In the second case (probably a much rarer one) there is a free ring-shaped space 

 between the medullary shell and the first (transverse) girdle, and both are connected 

 by a small number of very short and small radial beams (R. Hertwig, loc. cif., p. 52, 

 line 19 to 21 from above). This mode of connection would be the same as is common 

 between the concentric shells of the Sphseroidea and P r u n o i d e a. The 

 distinction between these two diflereut modes of connection is often very difficult. 



The second or lateral girdle is commonly not in direct connection with the 

 medullary shell, or onlj^ by some scattered radial beams (mainly in the principal axis). 

 This lateral girdle arises by prolongation of Ijoth wings of the transverse girdle in the 

 lateral plane, so that from both sides (right and left) they become united on the poles 

 of the principal axis. The minor axis of the elliptical lateral ring (thus formed) is 

 therefore the major axis of the foregoing (transverse) elliptical ring; the major axes of 

 both are perjDendicular one to another. The major axis of the lateral ring is the 

 principal (or longitudinal) axis of the whole body. 



The third or sagittal girdle becomes developed from the second almost in the same 

 manner as the second from the first. On both poles of the principal axis two latticed 

 wings arise from the lateral girdle, growing further in the direction of an elliptical ring, 

 which represents the jaerimeter of the sagittal plane or median plane. These wings 

 are already mentioned by J. Miiller as " prominent roofs, protecting the gates of the 

 Tetrajn/le-shell." If these roofs grow towards the equatorial plane of the shell and 

 become united in pairs on the poles of the sagittal axis, the third girdle becomes 



