Terraces and River Valleys — Spencer. 159 
Spanish shelf is reduced to six to twenty miles in width, 
but it is indented by several short ravines or deep bays. 
Other amphitheatres indenting the Biscay shelf are noticed. 
The canyons are repeated off Spain and Portugal. The 
Caneia (opposite to cape Penas) in a distance of twenty 
miles descends to 8,280 feet and bends in the middle of its 
course. The Lima dessects the platform to a depth of 
5,622 feet, where it is already submerged another 1,200 feet. 
The channel of the Douro incises the platform to within 14 
miles of Oporto. Ofif JNIondego and ofiF cape Carvoeiro are 
canyons 5,000 feet or more in depth cutting back into the 
continental platform. The Grand canyon of the Tagus 
is especially well developed, with its apex five to six miles 
from cape Razon. At 35 miles from the apex is a branch 
round an isolated rock or sea stack which rises to within 
396 feet of the surface. Here and adjacent portions of the 
declivity suggest a lower platform corresponding to the 
submergence of the Blake plateau in a general way (re- 
viewer). The sea stack is only an isolated fragment of 
the platforms. Of several of the canyons Prof. Hull gives 
sections sustaining his conclusions, which may be cited in 
his own language. 
"Were there no other suboceanic channels than that of the 
Adour, it would of itself be sufficient to demonstrate its own fluvi- 
atile origin and that of all others here described. For what are the 
characteristics of a river-valley draining a plateau and adjoining 
regions? They are first, a continuous deepening of the bed of the 
channel in the direction of the outlet; second, continuous widening 
of the channel in the same direction; third, a widening course; 
fourth, lateral tributaries. All these characterize the suboceanic 
channel of the Adour. On the other hand they are not characteris- 
tic of seismic fissures, or of fissures formed by faults or any other 
processes with which we are acquainted on the land surface. We 
are familiar with valleys with similar characteristics, but waterless, 
entering the great valley of the Nile or the shores of the Red sea, 
or traversing the region of Arabia Petraea and southern Palestine, 
where rainfall is either absent or only intermittent; but we do not 
hesitate to recognize in them the channels of former streams and 
rivers, though they are now dry. No other theory than that here 
advanced will, I venture to hold, serve to explain their origin and 
presence under the waters of the ocean." 
In this paper Hull reviews the question of the date of 
