312 DR. G. P. BIDDER ON 



inclusions of large size, around which vacuoles frequently make their appear- 

 ance " ; they stain like the reserve granules. Robertson and Minchin have 

 four other interesting figures of collar-cells with spherical enclosures, of 

 which fig. 32 shows an almost uniformly staining sphere of 4/i diameter 

 surrounded by a large vacuole. Dendy's fig. 66 and fig. 84 show three 

 nearly spherical cells of 7 /x diameter in phagocytes; two have a bilateral 

 symmetry or are commencing division. 



If Dendy's theory of spermatogenesis may be taken as withdrawn, these 

 appearances are all unexplained. It is therefore open to us to conjecture 

 that the monad of Syncrypta spongiarum, after ingestion by the collar-cell, 

 is not digested, but becomes surrounded by a vacuole in which it grows to a 

 diameter of 4/^, inducing, in Grantia compressa, considerable abnormality in 

 the containing collar-cell. At this size it appears to pass into the mesoglcaa, 

 where it increases to the 7 /j. diameter of Urban's symbiotic algse and Dendy's 

 yellow bodies ; but G. compressa appears to have specialised a system of 

 phagocytosis which prevents its multiplication, so that it may be only in the 

 Clathrinids and Hircinia that the parasitic (or sjnnbiotic) generation attains 

 success *. 



On this hypothetical history we should assume that conjugation occurs 

 between the monads before ingestion by the collar-cells, that the zj'gote in a 

 hospitable sponge gives rise to many generations of parasitic red cells, but 

 that their symbiosis, if it be true in Clathrinids, is rejected by- Grantia. 

 With the onset of unfavourable conditions the red cells, instead of further 

 multiplication of their like, form spheres of coherent gametes which escape 

 to seek a new host and reinforced vitality f. 



The fact that among calcareous sponges the red cells have hitherto only 

 been recorded in occupation of two individuals militates strongly against 

 Urban's conclusion that they are here symbiotic. Dendy found no collar- 

 cells in " Leucosolenia gardineri" which was possessed by them. The 

 suggestion is rather that this organism is a dangerous parasite of all sponges 

 but Hircinia, whose physiology is accommodated to it, perhaps in true 

 symbiosis. It appears to be destroyed in G. compressa by phagocytosis, and 

 the necessity of repelling this enemy must be taken into account in con- 

 sidering unexplained phenomena in the histology and anatomy of sponges. 



If the suggested hypothesis be correct, and the red, violet-brown, and 



* Dendy describes in the "ova" of Leucosolenia gardineri (p. 4) "conspicuous spherical 

 bodies .... closely resembling in appearance the yellow bodies .... but more variable in . 

 size." Their being described inside the nucleus would prevent us guessing that this could 

 be phagocytosis ; but the nucleus " has an irregular outline and no nuclear membrane." 



t The pyrenoids observed by me in A. cerebrum suggest, that the containing monads have 

 been digested ; Robertson's fig. 34 and Dendy's fig. 70 seem indicative in the same direction. 

 It is possible that the newly-formed zygote can resist digestion by the collar-cell, but that 

 unpaired gametes merely supply food for the sponge. 



