ENDOCYCLICA 53 I 



in the lower part of the radii near the peristome l)y the adhesion 

 of the pore-plates in groups of two, three, or more. Sometimes 

 as many as nine pore-plates may thus adhere. 



When adhesion takes place between the pore-plates it is of 

 course preceded by crowding, and this interferes with their 

 equal development. Some which extend so far horizontally as to 

 meet their fellows of the opposite side of the radius are called 

 primary plates ; others which are small and wedged in between 

 the larger ones are called demi-plates. Systems of classification 

 have been built up (chiefiy by palaeontologists) in which great 

 stress has been laid on how the primaries and secondaries enter 

 into the constitution of the compound plate, but it does not seem 

 to the present author as if this were at all a satisfactory basis 

 for classification. All the pore-plates are primarily equivalent, 

 and the question as to which are interfered with in their growth 

 so as to become secondary is trivial. The so-called Arbacioid 

 type consists of one primary with a secondary on each side ; the 

 Diadematoid type of three primaries, with occasionally a secondary 

 between the aboral and the middle primary ; and finally the 

 Triplechinoid type of two primaries, with one or more secondaries 

 between them. 



Aristotle's Lantern. — Under this head we may consider the 

 auriculae and gills as well as the jaws and teeth. In Cidaridae 

 external gills appear to be absent, but from the lantern coelom 

 large radial pouches project upwards into the general coelom 

 cavity. These pouches are supposed to be respiratory, and are 

 termed internal gills or Stewart's organs.^ They co-exist with 

 external gills in Echinothuriidae and in Diadematidae, though in 

 the last family they are present only in a vestigial form, two 

 being found in each radius. The auricular arch both in Cidaridae 

 and in Arbaciidae is composed of two pillars which do not meet, 

 but in the last-named family they are based, as in Echinidae, 

 generally on the ambulacral plates, whereas in Cidaridae they 

 arise from the interambulacral plates (the ambulacral plates 

 being here very narrow). The epiphyses are absent in Cidaridae 

 itnd Ar1»aciidae, and are imperfect in Diadematidae. 



Spines. — These organs are extraordinarily variable, and 



' In this case the fluid Hows from the lantern coelom into Stewart's organs ami 

 vice I'crsa. Oxygen must be absorbed through the iieristome. Tlie Cidaridae are 

 not as sensitive to want of oxygen as the other families (Uexkiill, loc. cit.). 



