Porifera. 



the upper (formerly posterior) pole; from here the overgrowth starts, not diffusely. 

 As the flat cells advance the flagella become amoeboid processes indistinguish- 

 able from those of the covering cells , the columnar epithelium is confused from 

 agglomeration of the protoplasm about the nuclei. Axinella after 3 /4 hour shows 

 the centre of the body filled with tightly packed cells containing small nuclei (A], 

 while the peripheral parts, except the surface of fixation, are occupied by B^ and 

 B. 2 in their gelatinous matrix. 2 nd Stage. Definition of the layers (2 nd 

 half-hour) . The cells of the central mass, A. are indistinguishable except by their 

 nuclei, a few B^ amoeboid cells lie amongst them, but almost all B cells are ex- 

 ternal , together with the mostly tangential spicules. B^ on the outer surface 

 appear already in the adult condition as a flat epithelium, beneath it as similar 

 cells destined to form epithelial cells for the cavities or connective and contractile 

 cells. 3 rd Stage. Interlocking of the layers and appearance of ca- 

 vities (to end of 1 st day). Without any external changes the spicules become 

 radial and penetrate the mass of small nuclei, accompanied by definite spicule- 

 cells. The amoeboid cells also penetrate largely, they contain objects which cer- 

 tainly closely resemble the small nuclei [cf. Delage, Bericht f. 1891 Porif. p 5 

 and f. 1892 passim]. If they be so the phenomenon is without morphological im- 

 portance ; the number is small compared with the number of small nuclei lying 

 closely packed, almost like a syncytium, without any connection with amoeboid 

 cells. Delage himself finds a difference of structure and staining-power between 

 the 'nuclei of ingested cells' and those of the free cells : M. considers the former 

 to be probably undigested yolk-granules; they react like nuclei with borax- 

 carmine and other stains, but malachite green remains fixed in them though rapidly 

 washing out of the nuclei. In Gellius varians [see below] the objects in the 

 amoeboid cells , preserved in osmic acid , do not colour with paracarinin , and 

 with carmine easily lose the colour on washing, in each respect differing from the 

 groups of small nuclei. M. considers it possible (also in Spongilla] that they may 

 be nuclei of collar-cells pathologically destroyed, comparing observations of Wilson 

 and Dendy ; but there can be no question of cells so ingested ever being again set 

 free. - - The cells under the outer surface arrange themselves round the localities 

 of future subdermal cavities, early advanced in accordance with their high 

 adult development. As the larva becomes more flattened and transparent, sec- 

 tions show such spaces to have become lacunae, while between them the ectosomal 

 tissue takes its permanent character , with connective cells and radial spicules. 

 Less sharply defined lacunae appear among the small nuclei, the flagellate 

 chambers arise as groups, showing in section circles of about 10 cells; each 

 flagellate cell of the larva forms a flagellate cell of the adult. Surrounded by 

 many such groups appear the large spaces of future efferent canals, the A 

 cells bounding them show no epitheliar disposition, but are overgrown, almost as 

 in the metamorphosis, by elements with large nuclei. 4 th Stage. Completion 

 of Sponge Anatomy (to end of 3 rd day). The subdermal spaces, completely 

 lined with epithelium, are united into one subcortical cavity, traversed by radial 

 columns of spicules and tissue. The cells accompanying the spiciiles (confounded 

 by Delage with the spicule-cells , which are always B { ) are of the type B^ ; the 

 plasma is striated. In some sponges the cells alone directly bind the spicules, in 

 Axinella a cementing spongin-like substance secreted by them may be very early 

 shown with Orange G stain. B l are present in all parts of the tissue. About the 

 3 rd day pores and osculum appear, the pores open and shut and appear to be 

 perforations in single cells leading into the subdermal space. Irregular channels 

 may lead from this to the chambers, or they may be placed directly upon it ; the 

 actual prosopyle appears to be an extremely fine lacuna between two flagellate 



