1898.] Funafuti, Rotuma and Fiji. 481 



channel. This fauna differed little from that inside the lagoon 

 at Funafuti ; but Radiolaria were more abundant, and often the 

 pelagic eggs of different species of fishes were extremely numerous. 

 Bits of green weed and pelagic alga? were very common, but 

 there was little sand, or dirt of any sort. When the tide was 

 ebbing the nets gave few animals ; there was generally, however, 

 a large accumulation of sand with bits of weed in the bottle. 



In the " auftrieb " collected in the daytime at Rotuma, animals 

 were rarely to be found, while the fauna at night was very rich. 

 In agreement with this it was noticeable that only at night time 

 were the corals of the family AstreeidaB fully expanded ; these 

 corals expand somewhat on the tide rising, but I never saw any 

 with the exception of Euphyllia, Syniphyllia and Mussa fully 

 out in the daytime, while at night parts of the reef looked quite 

 different by reason of the expanded polyps. 



The examination of the food, found in the ccelenteron of 

 different species of Pi'ionastrcaa, Favia and other genera, killed 

 in the early morning, was unsatisfactory. In most I could only 

 find pieces of green weed with sometimes a few small, oval, uni- 

 cellular algse ; animal remains were rare, and could seldom be 

 identified. I next tried to feed different species of corals, placing 

 them in bottles, into which a certain amount of the " auftrieb " 

 from the passages in the reef was poured, and then sinking them 

 in a pool off Solkopi ; I could get no results with any genera, 

 except Euphyllia and Mussa. Other genera with large calices, 

 such as Prionastrwa and Favia, could not be obtained in small 

 enough pieces without damaging some of their polyps, and, after 

 removal from the mass, were rarely found expanded either by 

 night or day ; these two genera further were never seen ex- 

 tended in the daytime, while most of the pelagic animals in the 

 bottles, although they were sunk in 4 feet of water under an over- 

 hanging ledge of rock, never lived more than a few hours. The 

 different polyps of the species Euphyllia and Mussa, on which I 

 experimented, were not in physiological continuity ; increasing by 

 fissiparity, their calcinal centres soon become very distinct, the 

 corallites separating, and never remaining in long series. As the 

 living tissues only cover the top ^ — f of an inch of the coral I um, 

 the branches, where they are broken off, are dead and these corals, 

 when placed in the bottles, easily expanded. The animals in the 

 " auftrieb " gradually became less active, until they moved but 

 slowly over the bottom, so that the conditions should have been 

 eminently favourable for the polyps. The tide would not allow 

 the polyps to be kept in the position I chose for more than four 

 hours, so that in about this time the bottles were lifted and the 

 corals killed by being dropped into a saturated solution of corrosive 

 sublimate. 



