DEVELOPMENT. 119 



suffers from the cold and the scarcity of food, and dies away. But 

 throughout the moribund parent gemmules are formed. These survive 

 the winter, and in April or May they float away from the dead parent, 

 and develop into new sponges. Some become short-lived males, others 

 more stable females. The ova produced by the latter, and fertilised by 

 spermatozoa from the former, develop into a summer generation of 

 sponges, which, in turn, die away in autumn, and give rise to gemmules. 

 The life history thus illustrates what is called alternation of generations 

 (see p. 54). Interpreted from a utilitarian point of view, the formation 

 of gemmules is a life-saving expedient. As Professor Sollas says, " the 

 gemmules serve primarily a protective purpose, ensuring the persistence 

 of the race, while as a secondary function they serve for dispersal." 



All Sponges produce sex cells, which seem to arise from 

 amoeboid mesoglcea cells retaining an embryonic character. 

 In the case of the ovum, the amoeboid cell increases in size, 

 and passes into a resting stage; in the case of the male 

 elements, the amoeboid cell divides into a spherical cluster 

 of numerous minute spermatozoa. The similar origin of 

 the ova and spermatozoa is of interest. Most sponges are 

 unisexual, but many are hermaphrodite. In the latter case, 

 however, either the pioduction of ova or the production of 

 spermatozoa usually preponderates, probably in dependence 

 on nutritive conditions. 



Development. — It is not surprising to find that there is 

 great variety of development in the lowest class of Metazoa ; 

 it seems almost as if numerous experiments had been made, 

 none attended with progressive success. 



The minute ovum, without any protective membrane, 

 usually lies near one of the canals, and is fertilised by a 

 spermatozoon borne to it by the water. It exhibits a certain 

 power of migration, as in some Hydroids. Previous to 

 fertilisation, the usual extrusion of polar bodies has been 

 observed in a few cases, and is doubtless general. Seg- 

 mentation is total and usually equal, and results in a 

 spherical or oval embryo more or less flagellate. This 

 leaves the parent sponge, swims about for a time, then 

 settles down, and undergoes a larval metamorphosis often 

 difficult to understand. It is peculiarly difficult to bring 

 the history of the germinal layers in Sponges into line with 

 that in other Metazoa. 



(a) In the small calcareous sponge Sycandra raphanus (Fig. 51), as 

 described by F. E. Schulze, the segmentation results in a hollow ball of 



