PHYSICS: C. BARUS 
475 
represents mere ledges of coral rock on the Admiralty chart: the 100- 
fathom line around each of three such reef rocks near the Exploring 
Isles is less than a mile in diameter, and the rocks are mere points; 
hence these minute reefs, beginning on small volcanic cones in the stage 
of sector G, must have been extinguished in sector J, resurgent in sector 
K, a Httle enlarged by outward growth in sector L, and almost extin- 
guished again in sector M. 
The upheaval and the sub-recent subsidence mentioned in the fore- 
going paragraph were not uniform, as has thus far been implied. The 
recent subsidence is believed to have increased to the east or northeast; 
first, as Agassiz pointed out, because the floor of the great lagoon of the 
Exploring Isles increases in depth from 20 fathoms or less on its western 
side to 80 or 100 fathoms on its eastern side; second, because several 
^drowned atolls' or submerged banks lie to the northeast; indeed the 
northeastern-most of the seven outlying atolls is mostly submerged; 
third, because Mango, 10 miles to the southwest, has elevated reefs 
moderately dissected at an altitude of 500 feet, and 30 miles to the 
west, Yathata and Vatu Vara are uplifted, undissected atolls at alti- 
tudes of 840 and 1030 feet. The intermediate island of Kanathea, 
nearer to the Exploring Isles barrier reef than Mango is, also seemed 
to me to bear small uplifted reefs at a height of about 600 feet, but I 
was too far from this island to make sure of it. 
The upheaval that preceded the subrecent subsidence must also 
have been unequal and greater to the east than to the west, because 
while the Exploring Isles were thus uplifted long enough ago to have 
been afterwards well dissected, Yathata and Vatu Vara were at that 
time presumably sinking and growing, preparatory to being uplifted 
recently as above stated. These unequal changes of altitude cannot 
be explained by changes in the level of the ocean, which are every- 
where alike; they can be explained only by unequal subsidence and 
upheaval of the islands concerned. 
INTERFEROMETER METHODS BASED ON THE CLEAVAGE 
OF A DIFFRACTED RAY 
By C. Barus 
DEPARTMENT OF PHYSICS, BROWN UNIVERSITY 
Received by the Academy, June 20, 1916 
The prismatic method of cleaving the incident beam of white hght is 
available for the superposition of non-reversed spectra, under conditions 
where the paths of the component rays may have any length whatever. 
