20 PROFESSOR W. J. SOLLAS. 



or series of deposits ; the material of the pinnacles being, on this hypothesis, the 

 youngest now surviving. All that we can conclude from the facts we have just con- 

 sidered is that no remnants of the ancient lagoon mound are open to observation.* 

 They may have been washed away simultaneously with the destruction of the ancient 

 outer ridge, or they may be concealed under the modern lagoon mounds now existing. 

 Instead of pursuing these speculations further we may proceed in search of additional 

 evidence, and we turn first to the Mangrove Swamp. This interesting feature is 

 situated at the " occiput " of the atoll and may be regarded as a widened extension 

 of the central flat of Funafuti islet, lying within the Ijend, where the islet passes 

 from a N.E. — S.W. to a N. W. — S.E. direction. A similar flexure separates the islets 

 of Funafara and Telele further to the south. The eastern side of the swamp is 

 sharply bounded by the inner slope of the ocean ridge, here rather steeper tlian usual ; 

 on the north and for a good part of its western margin it is bounded by low clifls of 

 consolidated coral breccia, and these are crowned in places by a beach of coral 

 fragments and slabs of breccia, which lie within the arcaded roots of the surrounding 

 mangrove trees. Some of the fragments of breccia seem to be of far too large a 

 size to liave ])een thrown into their present position by the waves of the water in 

 the swamp, and thus suggest that the breccia platform may have undergone some 

 erosion before the swamp was cut off" from the ocean. At its southern end the 

 Mangrove Swamp passes into the sandy flat of the Taro Swamp. Its northern arm 

 is divided into two parts by a longitudinal ridge of hard coral breccia, which rises 

 about 2 feet above the flat floor adjacent on either side ; the western side of the arm 

 is bounded by a clifl of coral breccia about 3 feet in height, the eastern half ends 

 against the ocean-ridge and is covered by water for a great part of the day, aff'ording 

 a liome to patches of close-growing madder-tinted algse and delicate Renierid 

 sponges of a bright rose tint. 



The feature of chief interest in the swamp is afforded by its southern main portion, 

 where a great part of the floor is formed by a dead coral reef, consisting almost 

 wholly of two species, one a massive Pontes, and the other Heliopora cceridea. The 

 Porites projects as rounded flat-topped slabs or blocks, from 6 inches to a foot above 

 the general level ; these are surrounded by radiating growths of Heliopora, which 

 may extend, in branching lamellse and finger-shaped processes (figs. 13 and 14), 

 for a distance of three or four yards from their origin. The clumps of Porites 

 often attain a diameter of 4 feet, are frequently depressed in the centre, and traversed 

 by more or less radiating chaimels, which owe their origin to included lamellae of 

 Heliopora, since removed by erosion. In many cases the Heliopora was the first 

 of these corals to commence its growth ; the Porites following, formed around the 

 centre of the Heliopora system, and in rare cases outgrew the latter, so as to 

 completely surround it ; usually, however, the Heliopora increased more rapidly 



*' The position, however, of the ancient rock in the cliffs of the lagoon is scarcely consistent with the 

 assumption of depression. 



