DENDY— HOMOSCLEROPHORA AND ASTROTETRAXONIDA 261 



of their mutual relationships is no less difficult than that of the name of the genus. 

 There is no doubt that members of the genus exhibit very great variability in all those 

 characters that might be used for the purpose of specific distinction ; in the character 

 of the surface, the form and size of the microscleres, the arrangement of the microscleres, 

 the thickness of the cortex, &c. 



Lindgren [1898] proposed to arrange the described forms in three groups, each 

 of which might be regarded as a species. The names of the three species v/ould then 

 be lyncurium, ingalli and japonica respectively. This arrangement is based on the 

 characters of the smaller asters. In Donatia lyncurium the smaller asters are supposed 

 to be oxyasters (0"012 — 0"02 mm. in diameter) only, variable in form and size but never 

 with tylote rays. In D. ingalli the smaller asters are tylote chiasters (0"006 — 0*016 mm.) 

 and oxyasters (0'02 — 0"052 mm.). In D. japonica the smaller asters are all tylote 

 chiasters (O'OOS — 0'016 mm.). D. lyncurium would appear to be characteristic of the 

 North Atlantic and Mediterranean ; D. ingalli of the Red Sea, Indian Ocean, Malay 

 Archipelago, Australia and the Pacific ; D. japonica of the Philippines, Java and 

 Australia. I do not think that this arrangement can be accepted, the "lumping" of 

 species appears to be too comprehensive. 



In my report on the sponges of Ceylon [1905] I proposed to regard all the specimens 

 of Donatia in Professor Herd man's collection as belonging to three varieties of Donatia 

 {Tethya) lyncurium, which I distinguished as a, h and c respectively. The undoubted 

 existence of transitional forms of the small asters at first sight seems to make it impossible 

 to recognise such forms as D. lyncurium, D. ingalli, D. seycliellensis and D. japonica 

 as distinct species. There is no reason, however-, why either specific or varietal 

 distinctions should be based exclusively upon the characters of the chiasters and 

 oxyasters. The foum, size and arrangement of the large spherasters, and the arrange- 

 ment of the megascleres may also prove useful in this respect, and by taking these 

 into account I now believe it possible to arrive at fairly satisfactory specific distinctions. 



It is impossible to discuss in this place all the species and varieties that have 

 been proposed. Nothing but a thorough and tedious revision of the whole genus, 

 based, if possible, upon the examination of type specimens, can lead to any really 

 satisfactory general conclusions. 



There are thirty-one specimens of Donatia in the " Sealark " collection, and after 

 careful examination, including measurement of the microscleres of all of them, I have 

 decided to arrange them in five species. None of the specimens are much more than 

 an inch in diameter and many of them are much less ; they attain nothing like the 

 size often seen in Australian Donatias. 



It will be seen from the locality lists that the difierent species occur very much 

 mixed together. I have found three species (lyncurium, japonica and seychellensis) in 

 the same jar from Salomon, presumably all obtained in the same haul of the dredge, 

 and a different three (japonica, seychellensis and ingalli) in the same jar from Praslin 

 E,eef. These facts show how necessary it is to examine every specimen microscopically 

 before determining to which species it belongs. 



