Parts of the Lantern of Aristotle in Echinus miliaris. 47 1 



processes together form at this stage a comparatively flat bridge, the arch- 

 form being attained later. It is significant that this bridge was not developed 

 by the palaeozoic Echinocystoida, Perischo-echinoida, Cidaroida, Holectypina, 

 and two orders of Centrechinoida. According to Jackson (2, pp. 179 and 183) 

 this is an important character found only in the four families of the Temno- 

 pleuridae, Echinidae, Strongyloceutrotidae, and Echinoruetridse, comprising his 

 new sub-order Camerodonta. Evidently the presence of this bridge for the 

 attachment of the protractor muscles is a sign of advance belonging to a late 

 period in the phylogeny of the sea-urchins. 



The two epiphyses connected by their extensions are further in contact 

 below with the same jaw, so that -it is customary in text-books to designate 

 them as a pair of epiphyses. But, as far as one could make out from the 

 position of their rudiments, it strikes one that the two epiphyses which lie 

 one on each side of and below a rotula form a pair. This being so, one may 

 say each pair maintains its radial position in the adult, there .being no 

 diverging of the epiphyses from a radius. On the contrary, as has been 

 mentioned already, the two alveoli of each jaw, growing aborally like the 

 limbs of a V, each effect a junction with an epiphysis belonging to an adjacent 

 radius. This association, though further strengthened by the formation of a 

 bridge or an arch, is to be regarded as only of secondary importance. The 

 orientation of the ossicles of the lantern in the adult may, therefore, be as 

 follows : compasses, rotulae, and epiphyses radial ; jaws and teeth inter- 

 radial. 



5. The Eotulce. 



Each of the five rotulae is laid down in a radius of the " echinus-rudiment " 

 in the metamorphosing larva, as a tri-radiate spicule, always with two rays 

 directed towards the oesophagus and one ray turned in the opposite direction 

 (Plate 11, figs. 1 and 2,R.B.). By the usual process, a broad, fenestrated, and 

 more or less rectangular plate is formed, roofing over the two epiphyses below, 

 which are thus concealed from an aboral view of the lantern at this stage 

 (Plate 14, fig. 5). In this plate, the boundaries of the primordial tri-radiate 

 spicule may still be traced. This stage recalls to one's mind the broad rotuke 

 found in Palceodiscus ferox (6). 



6. The Compasses. 



In the absence of any evidence to the contrary, the " compasses " have been 

 usually regarded as unpaired elements. The bifurcated ends, however, were 

 supposed by Sollas (5) to indicate their paired nature. But their develop- 

 ment shows conclusively that, whilst they are duplex structures, the forked 

 ends are not indications of a paired nature. 



