86 THE VOYAGE OF H.M.S. CHALLENGER. 



which have lost the siliceous spicules. In my opinion, the common ancestral group of 

 all sponges (provided that the whole class is monophyletic) has been skeletonless, 

 and the various main groups (subclasses or orders) descending from it have acquired the 

 different skeletal forms in different ways polyphyletically. This does not exclude the 

 possibility that in some skeletonless sponges the want of a proper skeleton is secondary, 

 produced by reduction. 



For the sake of brevity and clearness I will here call the hypothetical common 

 ancestral group, in which originally no skeleton was formed, Archispongiae. To this 

 primordial group may perhaps belong some Myxospongise (Halisarcidse, Chondrosidae) 

 and some Psammospongias (Arurnoconidae, Psamrninidae). From the same common 

 ancestral group may have arisen, as independent main branches, on one side the Calci- 

 spongiae, on the other side a part of the true Keratosa (not descended from Silicosa), and 

 further the Demospongiae (Monaxonida and Tetractinellida) and the Hyalospongiae 

 (Hexactinellida). It is quite possible that a horny skeleton, produced by the formation of 

 spongin-fibres, has arisen polyphyletically, independently in different groups of sponges. 

 The now prevailing opinion of their monophyletic origin seems to me not very probable. 



The nature and origin of the horny skeleton is an important point in these phylo- 

 genetical problems. In my opinion, the spongin-skeleton must not be regarded as a 

 formation of the same order and value as the calcareous skeleton of the Calcispongise, or 

 the siliceous skeleton of the Silicosa. Regarded from a general histological point of view, 

 the horny tissue of the sponges seems to present many analogies in form and develop- 

 ment to the elastic tissue in the higher Metazoa. The different forms of thin fibrillar and 

 strong fibres, simple and branched fibrillar, isolated and reticulated fibres, bundles and 

 networks of fibrillar, which are found among the numerous modifications of the elastic 

 tissue, and which arise in the maltha or the ground-mass of the connective tissue, occur 

 also in the horny fibrous tissue of the sponges. The chemical nature is little different, 

 and even the origin may often be similar. The strong fibres of many Keratosa are 

 produced by series of associated spongoblasts (F. E. Schulze), but the fine fibrillas of the 

 Stannomidae and the spongin-capsules and lamellae of Cerelasma are certainly formed in 

 another way. Perhaps each fine fibrilla of the Stannomidae is the filiform product of a 

 wandering amcebocyte, in a similar way as a dentin-fibrilla is secreted from an odonto- 

 blast. But it may also be that these and similar spongin-fibrillae are produced by a 

 chemical and physical alteration of the ground-mass, without the direct action of a cell, 

 in a similar way as is the case in the fibrous cartilage. 



Comparing the horny skeletons in the new Keratosa here described (especially the 

 StannoniidEe), and in the various groups of the so-called Cornacuspongiae, it seems to me 

 very probable that horny fibres, as strengthenings of the maltha, have arisen in different 

 groups of Keratosa and of Silicosa, independently one from another ; it is even very 

 probable that the fossil Pharetrones (Zittel), that remarkable group of Calcispongiae which 



