1886.] Zoological Observations made in North Celebes. 325 



has been brought by Moseley, Zittel, and others, to prove that this 

 latter genus is Alcyonarian, there are still some authors who main- 

 tain that it is Zoantharian. The peculiar structure of the present 

 form goes far to prove that the former opinion is right, and the latter 

 wrong. 



In the second place the resemblance of the young colonies of this 

 form to the genus Cornularia, and the resemblance of the adult 

 •colonies and polypes to those of Tubipora, justify the conclusion I 

 arrived at in a former paper, that Tubipora should be united with 

 the Cornularidse into a group, the Stolonifera ; this genus is, in fact, 

 the connecting link between these genera which was formerly missing, 

 unless we assumed that Syringopora was undoubtedly Alcyonarian. 



I hope in a future paper to be able to give some further particulars 

 of the anatomy of this form, perhaps also some account of the early 

 stages of its development, and some account of my researches upon 

 the other Stoloniferous Alcyonaria, which are present here in abun- 

 dance. 



Note on Tubipora and on Millepora. 



I have got the early stages of the development of Tubipora. It is 

 regularly holoblastic, and I think the gastrula is formed by invagina- 

 tion. Finding, however, that it is very difficult to keep the embryos 

 alive in this hot and dusty weather, I must wait until it becomes a 

 little cooler in December before' I can get any very satisfactory results 

 on this latter point. 



The generative products of Millepora are formed in little capsules 

 in the walls of the canals, and I have found both male and female 

 capsules in the same canals. The embryos, I believe from the evidence 

 of one preparation only, reach a certain stage of development in 

 chitinous capsules in the canals, and they are then discharged into 

 the water by the mouths of the gastrozooids. 



Fig. 2. 



s Jrr/f ran set fa 



Generative capsules of Millepora. 



