Dr. H. RaufFo/i the Genus Hindia, Dune. 175 



such a union of the central points in a radial direction ; but 

 it was always doubtful whether the union was not simulated 

 by a trabecula lying- above or below : the clearest and most 

 connected images very seldom showed the beam 4 at all, and 

 then quite short and disunited ; the casts also indicate undi- 

 vided mesh-spaces. The abortion of the fourth trabecula also 

 appears quite natural if we consider that the heads of the arms 

 are often so much dilated laterally that tliey reach to the 

 central point of the spicules with which they are united, and 

 this great lateral growth must in course of time suppress the 

 fourth arm ; in connexion with this we have also the circum- 

 stance that the concave surface of the arms is smooth, as their 

 denticulation would be superfluous. 



This conception of the siieletal structure of the Himlia- 

 body seems to me to agree with all the observations, coinciding 

 with them and explaining them. 



As each individual part shows itself to be tetracladine, so 

 also the whole connexion ; tlie articulation is determined by 

 a law in conformity with this tetracladine character. Each 

 spicule has its four arms uniformly distributed in space (one 

 of them aborted) ; each pair of arms therefore (theoretically) 

 encloses an angle of 10y|°; but several sucli spicules, when 

 situated upon radial axes (each set of three arms running 

 parallel, while the longitudinal direction of the fourth indicates 

 the radial axis itself), constitute together by these three arms 

 three surfaces cutting each other at angles of 120°, canal- 

 walls. (Projections of these sets of three arms in a surface 

 perpendicular to the fourth arm = transverse sections of the 

 hexagonal column.) 



Thus the construction of the whole skeletal body appears to 

 be in the most perfect harmony with the nature of the 

 individual elements, and most perfectly adapted to their 

 character. 



Duncan states that the arms usually unite by their " frilled " 

 heads. This statement appeared to me to be contirmed here 

 and there in a radial section, a preparation which Duncan 

 also probably had. But as the specimen from which the 

 preparation was taken is distorted, and as it seems probable, 

 from a concentric fracture in the interior, that a more external 

 part of the sponge has been forcibly pushed inwards, these 

 portions, in which the arms appear to be united in an abnormal 

 fashion and which no longer show the radial streaks, but 

 present a hexagonal form of mesh, may really belong not to 

 radial but to tangential transverse sections. In these trans- 

 verse sections, however, the peculiar mode of union of the 

 spicules would of course no longer appear distinctly, for we 



