56 CELLULARIID^E. 



code of the sponge, or other yielding substance, and 

 holds the polyzoon, like an anchor, to its place*. 



Dredging on one occasion in Salcombe Bay, I took up a 

 piece of some cotton material which was overgrown by 

 S. reptans. Oil the uneven fibrous surface the adhesive 

 disks would have been almost useless ; a few of them only 

 were developed ; but the toothed processes were present in 

 profusion, and had worked their way in amongst the 

 threads of the fabric, which had become entangled amongst 

 the hooks, and so anchored the tufts securely. On a 

 specimen beside me, growing on the flat surface of a 

 Laminaria-frouA, I cannot find a single grapnel ; but the 

 disks are finely developed, and of large size. 



S. reptans grows in somewhat circular tufts, the branches 

 spreading out on all sides, and being more or less de- 

 cumbent. 



This fine species is one of the most highly specialized 



* This interesting piece of structure did not escape the " lyncean " Ellis. 

 He says: "Some of these little radical tubes are discovered by the micro- 

 scope to be full of hooks, the better to secure the Coralline when it adheres 

 to soft spongy substances " (Corall. p. 38). He also figures one of these 

 modified fibres. 



Couch has also briefly noticed the occurrence of hooked fibres on the 

 present species (Cornish Faun. pt. iii. p. 127). But we are indebted to Mr. 

 Peach for the most detailed observations on this interesting point. He de- 

 scribes very graphically the grapnels of S. scruposa, " the stout hooked 

 spines " of which were buried in the sponge. These hooks " are shaped like 

 the thorn of a rose-tree, and surround the 'root-fibres;' and when dragged out 

 they hold in their grasp numbers of the sponge-spicules." He also met 

 with similar appendages on S. reptans when growing upon sponge. A 

 specimen developed on Flustrafoliacea was altogether destitute of the hooks, 

 the fibres terminating in a number of short and simple " radiating pro- 

 cesses," by which they adhered. From these processes " short disk-like 

 pieces" are given off, which are inserted into the opening of the cell of the 

 Flustra ; and in this way a firmer grip is secured. There appears in this 

 case to be another slight adaptive modification determined by the habitat. 

 On the smooth and solid surface of the Laminaria the terminal fibrils are 

 much more branched, so as to give a larger adhesive surface to the disk. 

 Mr. Peach's papers are published in the Linnean Society's Journal, Zool. 

 Yol. xiii. p. 479, pi. niii., ami the Journal Roy. Inet. Cornw. 1877, No. I'.'. 



