lO BULLETIN OF THE BUREAU OF FISHERIES. 



branches of the canals are at this stage imperfectly developed, and the flagellated cham- 

 bers have scarcely begun to differentiate. Examination shows that the radial systems 

 interconnect with one another. 



The distribution of the radial efferent systems is well shown in figure 12, plate 11, 

 and figure 13, plate in, both of them photographs of entire slide preparations, the former 

 taken with transmitted, the latter with reflected light. The preparation shown in 

 figure 12 was kept six days in the live box. The flagellated chambers and canal systems 

 are well developed. The movements of the flagella belonging to the collar cells and the 

 currents passing out of the oscula were directly observed with the microscope. 



A slightly later stage in the development of the canals is shown in figure 14, represent- 

 ing part of a slide preparation that had been kept eight days in the live box. Three 

 of the efferent radial systems appear. A higher magnification would reveal an osculura 

 at the center of each system. The oscula are sometimes mere apertures in the dermal 

 membrane, but they may also terminate short tubes (oscular tubes) which ascend 

 vertically from the incrustation. The terminal ramifications of the efferent canals 

 are well differentiated in this preparation. Pores are scattered over the dermal mem- 

 brane. The afferent canals are not conspicuous. Between the efferent canals lie 

 immense numbers of flagellated chambers. How abundant the flagellated chambers 

 are in these young sponges may be inferred from figure 19, plate iii, which represents 

 a small part of a cover-glass preparation in about the stage of figures 12 and 14. The 

 chambers are thickly crowded between the efferent canals. 



When the plasmodia have metamorphosed and the canals and chambers have 

 developed, the skeleton makes its appearance. In sponges that have been kept a few 

 days in the live box one observes spicules (styles) some of which are strewn horizon- 

 tally through the body, others of which ascend more or less vertically, projecting from 

 the surface. The latter are arranged both singly and in small tufts. The skeleton in 

 this condition is shown in figure 15, plate in, which represents part of a slide preparation 

 kept eight days in the live box. At this stage the horny columns are exceedingly small, 

 consisting of minute aggregations of spongin round the bases of the ascending spicules. 

 All the spicules characteristic of the species are present. As to the size of the spicules, 

 the chief point of difference from the adult condition lies in the slenderness of the smooth 

 styles. Actual measurements made at this stage of growth are as follows: Larger 

 smooth styles, 200-250/£ by 5/x; spinose styles, 72/i by 5/i; isochelae, 14/^ long; toxas, 

 40JJ. long. The incrustation at this time is very thin, about }i mm. thick. 



If the preparations are kept in the live box they gradually thicken, and the skel- 

 eton continues to develop. Figure 16, plate in, represents a vertical section of a prep- 

 aration that was kept six weeks in the live box. In removing the incrustation from the 

 glass plate on which it had grown, it was somewhat torn. The incrustation is about 

 ^3 mm. thick. There is a distinct basal homy plate. The vertical homy columns are 

 conspicuous. The spicules characteristic of the species are all present, and the smooth 

 styles are as large as those found in many normal incrusting specimens. The smooth 

 styles actually measured in this preparation 250-340/! by 8-io/i. On comparing figure 



