44 I. IJIMA : HEXACTINELLIDA. I. 



domiuantly sleuder triactins, which I propose to call the thctac- 

 tin in view of the T-like disposition of their rays. Occasionally 

 there occur comitalia of other forms, as the commonest of which 

 may be mentioned linear diactias and a variety of tetractins, 

 which, unlike the stauractin, consists of two rays in a straight 

 line and two lateral unpaired rays. I have called this sort of 

 tetractins the pavaietractin, being at loss for a better apellation. 

 All the comitalia lie bundled together with their prolonged 

 complete axis, while the lateral unpaired rays project forth from 

 the bundle at indefinite intervals and in various directions. 



The oblique system, which consists of two sets of right and 

 left handed spiral beams, is usually less strongly developed and 

 weaves its way mainly between the two other systems. How- 

 ever, it anastomoses not infrequently with either of these, and 

 occasionally some of its beams are seen to intersect the circular 

 beams on the inside, while a goodly number pass on to the 

 outside of the longitudinal in order to communicate with the 

 looser parenchymalia on that side. The spicular composition is 

 essentially the same as in the other systems ; only the principalia 

 are here furnished by thetactius or diactins in the absence of 

 either stauractin or peutactin j)i"iiicipalia. 



The parenchymalia of the sieve-plate are afforded by con- 

 tinuations of the longitudinal and oblique beams, especially of 

 the former (PI. IV, fig. 4). However, in most species the prin- 

 cipalia are here chiefly oxydiactins bent more or less in ac- 

 commodation with the irregularities of the sieve-plate beams ; the 

 accessoria are mainly diactins and thetactius. In some species 

 (f. i., E. marshalli), the thetactins may predominate over all 

 other forms of spicules, furnishing alike the principalia and the 

 comitalia of the sieve-plate parenchymalia. 



