PORIFERA. I. 8l 



b\- spongin which seems onh', however, to unite their ends. The membrane forming the inmost Hning 

 of the canals, as before mentioned, has spicules loosely gathered into bands, the direction of wliich in 

 relation to the canals ma>- be varying (PL XV, fig. 4); no spongin is found here. When we have got 

 a clear understanding of the structure of the skeleton of the rind and the walls of the canals, the 

 skeletal structure of the fistulas is easily understood; the fact is that their external part is formed by 

 the rind, or, to put it more exactly, by the outer layer of the rind which passes into the outer layer 

 of the fistulae, and their interior layer is formed by the membrane surrounding the canals with its 

 skeletal net of loose fibres. These two parts are comparatively loosely joined, and may easily be 

 separated. The external la3er of the fistulse has a thickness of about a"""; here the fibres are com- 

 paratively powerful, and esj^ecialh' jDowerful ones appear running longitudinally; the meshes formed by 

 the fibres, are also averagely larger than in the rind-layer of the body, so that fistulae upon the wliole 

 consist of a comparativeh' open plait of fibres. Outermost the same layer with vertically projecting 

 spicules is found as on the other parts of the body. The inner layer of the fistulae is constructed in 

 quite the same way as is the case in the canals, and innermost it is lined with the same membrane as 

 these. The fistulae consist almost exclusively of fibres and membranes without any other tissue. The 

 outer and inner layers are in the upper end more firmly joined, and here the inner membrane shows 

 a great many holes, so that it is reduced to a network; by this feature the pores on the fistulse are 

 made to lead more or less directly into the inner cavity. It is not difficult to understand that these 

 fistulse can act as oscula, for the water streaming forth through the excurrent canals, is poured into 

 their inner cavity, and from the upper part of this cavity it is emptied out through a great nianv 

 pores; thus the whole collection of pores serves to remove the quantit)' of water which would else 

 have to be removed through a single osculum. 



Spictila: a. Megasclera are oxea, slighth', not rarel\- a little irregularly curved, with sharp, not 

 long points, most frequently bounded by straight lines. Their length is generally between 0-20 — 0-268""", 

 but it seems to be somewhat \-arying, as in a specimen from station 83 the length frequently is 0-298™"', 

 the most common length is 0-23—0-26""™. Ridley and Dendy, 1. c, state the length of oxea in 

 Bowerbank's original specimen to be 0-19—0-25"'™. The thickness is 0-008—0-0128™"'. Shorter and 

 finer needles occur, but only singly; in the interior tissire, however, they are in some places seen in 

 no small numbers, b. Microsclera: only one form, viz. sigmata; these are very small and of a rather 

 constant size; the length is averagely 0-014™™, sometimes up to 0-017™™, the thickness in the middle is 

 about 0001™'". Boverbauk gives the length as -j^ inch = 0-012™™. Ridley and Dendy, on the 

 other hand, give the length of the sigma to be 0-038™™, and the thickness 0-0032™™; this size of the 

 sigmata would almost seem to make it doubtful, whether Ridley and Dendy's species is Oceanapia 

 robnsta, or perhaps an allied species. Sigmata are found throughout the sponge, but are especially 

 frequent in the membranes, as well the dermal membrane as the membranes of the canals, and also in 

 the inner tissue. 



As Topsent has sent me a piece (fragment and fistulse) of the Gelliodes cavicorm's established 

 b\- him 1. c, I have been able with certainly to decide that it is Oceanapia robusta. Professor Top- 

 sent has also in a letter declared that he thinks the two species to be identical. — As I have had 

 occasion to examine a piece of the specimen mentioned by Fristedt 1. c. which, with some misgivings, 



The Ingolf-Expedition. VI. i. " 



