Foerste.J 
410 
[April 6, 
streams from opposite points of the same fold should begin cutting, 
the operation would be still more effective. The deepening of 
such a cross valley might furnish an outlet to the drainage behind 
the fold, by “tapping.” The inconspicuous part played by the 
lateral streams on the sides of Jura folds makes such a theory very 
improbable. Numerous dry notches or “wind gaps” ought to 
exist all over the Jura Mountain system, representing this back- 
ward erosion, in various stages of progress, to bear witness to the 
method of operation employed in the production of such cirques. 
While backward erosion has done much at various points along 
the Bernese Jura, it nevertheless is hardly of sufficient magnitude 
and frequency to account for the cirques. This theory is particu- 
larly at fault when the strange grouping of cirques along lines 
transverse to the folds is observed. Thus the Boujean and Pery 
cirques , the Sornetan and Undervelier cirques , the Court, Moutier, 
Roche, and Choindez cirques , lie along certain lines, and it is un- 
reasonable to suppose that backward erosion should have selected 
points so intelligently as to bring about these results. 
9. Consideration of additional Theories. It has thus been seen 
that almost all known theories of the origin of cross valleys have 
been applied at some time or other in the explanation of those of 
the Jura Mountains ; yet none of them seems entirely satisfactory. 
But there are two theories that, as far as I am aware, have never 
been considered in the case of the Jura. Both are intended to 
explain the origin of cross valleys under certain conditions. 
The first theory requires that a fold or even only a hard stratum 
dipping at more or less of an angle should be covered unconform- 
ably by layers of more recent formations. Streams adapted to 
the topography offered by the later formations might cut down 
their channels until they reached the concealed fold or hard 
stratum beneath. If the erosive power of the stream were suffi- 
cient, the stream would cut through its barrier before its head 
branches could be captured by other streams. If the unconform - 
ably superimposed beds were more or less removed, but the fold 
or hard stratum beneath resisted erosion well, then, in the course 
of time, the surrounding and overlying formations would be re- 
moved from this area, but the river, having already fixed its course 
in the harder rocks formerly concealed, would retain this channel. 
The direction of such a superimposed stream would not be given, 
however, by the beds in which the stream is now running, but by 
