TJIE GEOLOGY OF THE TONGA ISLANDS. 609 



their descent to the sea. Opposite it, on the right of the passage is 

 the ishmd of Falevai (called " Kopa" in the Admiralty chart), which 

 consists in part of a two-storied mass, in part of a one-storied tongue 

 of land which extends out towards Mo'unga Lata, while a terrace 

 at the same level interru])ts the slope on the side of the higher part 

 of the island. Small islands are dotted over the sheltered expanse 

 of water, two of which, A'a (called " Koto " in the chart) and 

 Langito'o, present features of interest which will be alluded to below. 

 As we advance, the hill Talau comes into view, exactly repeating 

 the shape of Mo'unga Lafa, in the level outline of the summit and 

 the two lines of terraces on its sides. 



The rough aneroid measurements that I obtained give about 140 

 feet as the upper limit of the first "story,"' 260 to 350 feet as that of 

 the second, and 420 and 520 for the summits of Talau and Mo'unga 

 Lafa, the two highest points of land in this part of the group. 

 Although these numbers show that the levels of the terraces at 

 different points do not agree accurately, the conclusion was strongly 

 impressed upon me on the spot, that the terraces and islands at each 

 level do in fact correspond, that is, that at successive stages in 

 the elevation of the group, the summit of each " story " has stood 

 at sea-level. The differences in their level may be in part accounted 

 for by the unequal action of the elevatory forces at different points 

 of the area acted on. Thus, on the hills Talau and Mo'unga Lafa, 

 which are between three and four miles apart, while the summit of 

 the lowest story is at about the same level in each (140-150 feet), 

 the two upper ones stand at about 270 and 420 feet in Talau as 

 against 360 and 520 feet in Mo'unga Lafa. Prom this it appears 

 that the second elevation was of greater extent in the neighbourhood 

 of Mo'unga Lafa than to the eastward near Talau. 



One remarkable feature in the contour of the land is the fact that 

 in some cases the summits of the islands, though level in profile, are 

 excavated by well-marked depressions. 



The most perfect example of this structure is presented by the 

 little island of A'a (the Koto of the chart) above mentioned. The 

 island is of an oval shape, being about three quarters of a mile long 

 and half a mile broad. It rises out of deep water, and is not sur- 

 rounded by a reef. In the chart a sounding of over 40 fathoms is 

 shown close to the southern shore, and except for a small shallow 

 area at the eastern end the other soundings in the neighbourhood 

 indicate nearly as great or greater depths. 



The sides slope steeply up to a height of from 85 to 95 feet (an 

 elevation considerably less than that of the top of the first " story " 

 in the surrounding islands), but the interior of the island is de- 

 pressed 50 or 60 feet below the level of the margin, being occupied 

 by a flat which stands at a height of some 36 feet above the sea. 

 The margin is of a nearly uniform height and forms a complete 

 unbroken ring. On its summit rough masses of limestone rock 

 stand up, but the flat interior is formed of brown earth. It seems 

 clear that when the rock forming the edge of the island stood at sea- 

 level a small lagoon some 10 fathoms deep occupied the centre. 



