736 PROF. A. DENDY AND ME. R. W. H. ROW ON 



Family 3. LEUCALTID^ nov. 



Diagnosis. Sponge colony tubular and ramified, or even anas- 

 tomosing, with many oscula, or individualised with large 

 central cavity and single osculum. Wall of colony composed 

 of at least two distinct layers, namely, a dermal cortex with 

 strongly developed skeleton of tangential radiates, and a 

 chamber layer with a skeleton greatly reduced or even 

 absent. A thin gastral cortex or membrane may or may 

 not be present. Skeleton composed, mainly at any rate, of 

 equiangular radiates. ISTo sub-gasti"al sagittal radiates. 

 Nuclei of collared cells probably always basal. 



The members of this family appear to have been derived from 

 a Dendya-\\\ie ancestor by the development of a thick dermal 

 cortex with a strongly developed coi-tical skeleton, and the con- 

 sequent more or less complete reduction of the no longer necessary 

 skeleton of the chamber layer. As in other families of Calcarea, 

 the flagellate chambers range from greatly elongated and even 

 branched, and more or less radially ari-anged, to small, sub- 

 spherical and scattered. 



Bidder [1898] has already pointed out that the nuclei of the 

 collared cells in Leucaltis clathria Haeckel {Heteropegma nodus- 

 gordii Polejaeff) ai-e basal in position, and w^e are able to confirm 

 this observation and to add that they are basal also in Lettcettiosa 

 dictyogaster Row [1913 MS.]. 



The reduction of the skeleton of the chamber layer, correlated 

 with the development of a thick dermal cortex with a special 

 cortical skeleton, finds its parallel in the genus Gra7itiopsis 

 amongst the Grantiidfe, but in that case the syconoid ancestry 

 is very clearlj'- indicated in the remains of an articulate tubar 

 skeleton, while in Leucaltis the vestigial skeleton of the chamber 

 layer shows no indications whatever of an articulate origin. 



The characters above mentioned, in our opinion justify the 

 close associatioia of Leucaltis with Leucettusa, and the wide 

 separation of these genera from both Leucandra and Leucilla 

 in our scheme of classification. 



We prefer to derive the Leucaltidse directly from a Dendya- 

 like ancestor, rather than indirectly through Leucascus, because 

 Leucaltis still preserves the more primitive type of radial colony 

 formation with what we presume to be a true central gastral 

 cavity, while Leucascus has adopted a massive type of colony 

 formation in which the exhalant canals are possibly to be regarded 

 as pseudogastral in nature. 



