rs; he was also the first who
proposed to drive ships with paddles worked by steam.

We now come to Thomas Savory, who got a patent in 1698 for a method of
condensing steam to form a vacuum. Savory describes his discovery in
this way:--Having drank a flask of wine at a tavern, he flung the empty
flask on the fire, and then called for a basin of water to wash his
hands. A little wine remained in the flask, which of course soon
boiled, and it occurred to him to try what effect would be produced by
putting the mouth of the flask into the cold water. He did this, and in
a moment the cold water rushed up and filled the flask, this being
caused by the steam being condensed and leaving a vacuum, which Nature
abhors, and rather than permit this the water rushed up and took the
place formerly occupied by the now condensed steam. We see by this in
how simple a way great ends are produced, and in the age in which this
happened, the result may be indeed be said to have produced a great end.
The engine of Savory was used for some years as a machine to raise
water. The principle of his engine was just as I have stated, and
consisted of two cases and other various parts, and this engine
possessed advantages over that of the Marquis of Worcester in sucking up
the water as well as forcing.

Savory's engine consisted of two steam vessels connected to a boiler by
tubes; a suction pipe, or that pipe which leads from a pump of the
present day to the well, and communicating with each of the steam
vessels by valves opening upwards; a pipe going from these steam vessels
to any required height to which the water is to be raised. The steam
vessels were connected to this pipe by other valves, also opening
upwards, and by pipes. Over the steam vessels was placed a cistern,
which was kept filled with _cold_ water. From this proceeded a pipe with
a stopcock. This cistern was termed the condensing cistern, and the pipe
could be brought over each steam vessel alternately from the boiler.
Now, suppose the tubes to be filled with common air, and the regulator
placed so that one tube and the boiler are made to communicate, and the
other tube and the boiler closed, steam will fill one of the steam
vessels through one tube; at first it will condense quickly, but erelong
the heat of the steam will impart its heat to the metal of the vessel,
and it will cease to condense. Mixed with the heated air, it will
acquire a greater force than the air outside the va