ox MEELIA XOEMAXI. 661 



siliceous and calcareous skeleton had seemed to me as improb- 

 able as that of a centaur, for no distinction between siliceous 

 and calcareous sponges has been considered so profound as 

 that of the chemical constitution of the skeleton. 



After examining about 1000 specimens, and tracing the 

 development of the soft tissues and skeleton at the growing 

 edgeSj I became convinced that Merlia was, beyond doubt, 

 a sponge with a siliceous and calcareous skeleton (5). The 

 o-rounds on which this conviction is based are so firm that the 

 final proof, such as would be afforded by seeing an embi-yo 

 of Merlia settle down and develop the siliceous and 

 calcareous elements, is, in my opinion, no longer necessary. I 

 am having, however, relays of specimens sent from Porto Santo 

 every fortnight, in the hope of discovering when Merlia 

 forms its reproductive cells and embryos, and of thereby 

 being enabled to work out the development. 



Description of Merlia normani.^ 



The great majority of the specimens, of which there are 

 about 1000, were obtained from sixty fathoms off Porto Santo 

 Island. All the specimens from this depth are in the form 

 of small incrustations, on an average about a square centi- 

 metre in area and a millimeter or less in thickness. The 

 crusts grow on shells, branches of corallines, Foraminifera, 

 worm-tubes, etc., are bi-ight vermilion in colour, and have a 

 smooth surface. They conform closely to the surface of tlie 

 objects upon which they grow, creeping round the edges of 

 shells and encircling the branches of corallines (PL 32, figs. 

 1-3). They are fiat and thin wlien spread over tbe smooth 

 inner surface of a shell, but in other situations they may 

 form sliglitly thicker convex bosses. Merlia always grows 

 on hard unyielding surfaces, and never on soft objects such 

 as sponges or Ascidians. The Merlia crust is always so 

 intimately united to the surface of foreign objects that it is 

 almost impossible to flake it off without removing some of the 



' A certain amount o£ repetition of details given in the introdiictoi"y 

 account seems unavoidable. 



