220 W. M. DAVIS 
The glacial-control theory was framed chiefly to account for 
atolls; barrier reefs receive less attention, but their formation 
appears to be conceived as follows: Let a volcano, VU, Fig. 3a or 4, 
of less ancient origin than those which were reduced to low relief 
in preglacial time, be submaturely or maturely dissected at the 
beginning of the glacial period, so that its spurs, W PU, are separated 
by radial valleys, WGU, while an alluvial delta-belt, UD, lying on 
a bank of wave-washed detritus, B’, with more or less coral in it, 
surrounds the non-embayed shores. Fig. 3a represents conditions 
which will result in producing a narrow-lagoon barrier reef in post- 
glacial time; Fig. 4 those which will result in producing a wide- 
lagoon barrier reef. [According to my own view this statement 
Fics. 3a (above) and 36 (below).—Structure of a barrier reef, showing conditions 
which will result in producing a narrow-lagoon barrier reef in postglacial time. 
should be modified by excluding preglacial reef growth around the 
stationary island, for reasons stated in connection with the struc- 
tural features of reefs formed during subsidence, and by adding a 
less or greater amount of preglacial abrasion, producing a rock 
platform, UL, Fig. 3b, back of which the island spurs would be 
truncated in cliffs, LF, and outside of which a detrital bank, UB’, 
would be accumulated. | 
As a result of the lowering and cooling of the glacial ocean, the 
valleys, WGU, are deepened to GC, Fig. 3a or 4, and widened as 
much as the duration of lowered sea-level allows; at the same time a 
marginal platform, B’’C, is cut down and built out about 40 fathoms 
below normal sea-level; but the retrogressive abrasion of the plat- 
form is supposed to go only so far as to cut small cliffs, C, in the 
resistant lavas of the volcanic island. Later, as a result of the 
