< i °9 ) 
Derlyjbire, the Hilly Parts of fVaJes, or other Coun- 
tries, muft be fatisfied that they abound with Caverns 
of many forts. Some of them are dry, others ferve 
only for Paflages, or Channels to Streams, which run 
through them ; and a third Sort colleft and hold Wa- 
ter, ’till they are full. They muft likewife have ob- 
ferved, that there are fometimes narrow Palfages run- 
ning between the Rocks which compofe the Sides, and 
going from one Cavern to another. Such a Paflage, of 
whatfoever Shape or Dimenfions, how crooked and 
winding foever in its Courfe, if it be but tight, and runs 
from the lower Part of the Cavern, firft upwards to a 
lefs Height than that of the Cavern, and then down* 
wards below the Mouth of the faid Paflage, will be a 
natural Syphon. 
A natural Refervoir then. Tab. II, Fig. i. A BCD,' 
with fuch a natural Syphon, MNP may be fup- 
pofed. Let a Stream, which I Ihall call the Feeding- 
Stream, enter it, near the Top at O. The faid Cavern 
muft contain all the Water which comes in at O, ’till 
it is filled to the Top of the Syphon at N. Then the 
Syphon beginning to play, and being fuppofed always 
to difcharge more Water than comes in by the Feeding- 
Stream at O, will empty the Cavern, ’till the Water is 
funk in it below the Mouth of the Syphon at M; when 
it muft flop, ’tiil the Cavern is filled, and the Syphon 
runs again as before. If the Water difcharged by fuch 
a Syphon, M P be brought out of the Earth by a Chan- 
nel P Q., the Water will flow out of the Earth, and 
flop alternately, making an intermitting Fountain at Q_. 
By this plain and eafy Contrivance, feverai of 
the flowing and ebbing Springs obferved by the 
Naturalifts, may probably be explained ; and even 
a much greater Variety of them than is hither- 
S s z to 
