CALCAREOUS SPONGES. 799 



-of descent Avere cleai-ly indicated twenty years ago, as represented 

 by the Leucascidfe and SycettidsB respectively. 



It must, of course, be remembered that the real difference 

 between the two types of collared cells concerns, as Minchin has 

 shewn [1909], the relation of the flagellum, with its basal granule, 

 to the nucleus. This relation has, of course, only been deter- 

 mined in a very few cases. In Leucosolenia coriacea, for example, 

 the basal granule is situated at the apex of the cell and the 

 nucleus at the base, while in L. comjylicata the flagellum appears 

 to spring from the nucleus itself, which is apically situated. 

 There can be no doubt that the actual position of the nucleus 

 itself in the collared cell may vai-y temporarily under certain con- 

 ditions, but in good spirit-preserved mateiial it appears always 

 to settle down into a. characteristic position, which is either 

 basal or apical, and which may be determined without resort to 

 special methods of cytological investigation. We do not wish 

 to lay undue stress upon this character at present, and we 

 should not venture to use it were it not associated with other 

 distinctive features, but we have been surprised, in view of our 

 former opinion as to the systematic value of such a character, to 

 find how constant the position of the nucleus is in the two lines 

 of descent indicated. This will be sufficiently evident fromi 

 reference to the table given in the Introduction. 



We have in vain attempted to split up the unwieldy genus 

 Leucosolenia into smaller groups. The utmost we have been 

 able to do has been to isolate from the main body of species three 

 well-marked types, Ascyssa, Asaute and Denclya. We do not 

 consider that Minchin's proposal to divide the homoco?! sponges 

 into two families, Leucosoleniidse and Clathrinidte, is at all 

 practicable in the present state of our knowledge, and if it be 

 true, as he himself has pointed out [1909], following Goldschmidt, 

 that the two types of relation of flagellum to collared cell may 

 occur in the same genus of Protozoa (^Mastigina), we see no reason 

 for supposing that both may not occur in the genus Leucosolenia. 

 According to our view, this is a large and heterogeneous group 

 of primitive forms all closely related to one another and merging 

 into one another to a large extent, from which the two lines of 

 descent referred to have led tlie way to the evolution of the 

 higher Calcarea. 



We will take the Denclya, or Leucascid-Leucaltid, line first, 

 in which the nucleus of the collared cells is basal. The starting 

 point of this line seems to have been from some form closely 

 related to Denclya. The radiate arrangement of the colony in 

 this genus formerly misled Dendy [1893 A] into regarding it as 

 on the line of evolution of the Sycettidee, but there are several 

 strong arguments against this view. The radiate arrangement 

 appears to be but a modification of a reticulate " Clathriuid " 

 character, and actual open anastomoses may occur between the 

 radial tubes, which, in spite of what has been said by more than 



