3JE. J. MARTIN" ON THE TERRACES OE ROT OM AH ANA. 



173 



places it was broken through, by clumps of trees or small shrubs, 

 and many rush-covered patches appeared on the softer parts. 



Towards the lake the margin became uncertain and treacherous, 

 and the boundary between terrace-deposit and lake-mud was unde- 

 fined. 



Changes in Appearance. 



The want of a series of systematic observations of the rate and 

 form of deposit can never be supplied. 



Comparing photographs and notes taken in 1883 with experiences 

 two years later, it was found that marked changes were taking place 

 upon the upper part of the Terrace. 



The surface previously covered by an acicular deposit had now a 

 more granular character, and instead of crunching under foot like 

 hoar-frost, it had the yielding nature of a layer of snow. 



The overflow had also worn a slightly depressed channel, leading 

 from the cauldron across the platform, between the two shallow 

 upper basins already described. The flood rising to the lip of the 

 cauldron flowed through this channel and poured down the centre 

 of the Terrace. 



The erosion of this stream was evidently deepening its own 

 channel and smoothing the surface of those masses over which it 

 poured, while it threatened in time to alter the whole appearance 

 of the Upper Terrace. 



As the water rose more rapidly than it could escape by this new 

 channel, it next filled the circular outer basin (jSTo. 2), and then the 

 other large area (No. 3) ; a further rise spread over the western 

 end of the platform and down the trees on that side, the receptacles 

 on the east side being supplied by a cooler stream overflowing from 

 the basin (No. 3). 



This change of direction from that of the previously charac- 

 teristic uniform overflow was also apparent in the orange coloration 

 which was extending over the eastern tiers, probably due to con- 

 fervas. The temperature of the water in these depressions was only 

 100°, while in those at the same level on the western side it was 30° 

 warmer. 



Before attempting an explanation of the peculiar architecture 

 of the Terrace, it will be necessary to take into account the enor- 

 mous amount of material removed from the hill in excavating the 

 crater and cauldron ; from careful measurements this cannot be 

 estimated at less than 2| million cubic feet, an amount equal to a 

 deposit of eight feet in thickness over the entire area of the 

 Terrace. This detritus must either have been carried away with 

 the overflow into the lake or deposited upon the hill-slopes. 



Observing that the erosive action of the overflow was cutting a 

 channel through the hard siliceous pavement of the upper platform, 

 the torrent would certainly have opened a gorge through the soft 

 rock, if in its initial form the Terata Geyser had exhibited similar 

 characteristics to those with which we have been familiar. 



Or had its earlier activity been more feeble and intermittent, the 



