[ 2y6 ]

Instead of a fall of five feet, a stream was directed intothe funnel from only about half that height. The gagestill rose considerably, though not so high as before.

It is observable, that in the circumstances of these ex-periments, a twirling motion communicated to the waterin the funnel impeded the carrying down of air, the gagealways sinking on the water receiving such a motion ;whereas, in those of the preceding article, it seemed tobe by the twirling of the water that the air was pusheddown.

It appears therefore that there are two ways of makingair pass down with water through a funnel, one by di-recting the stream against the side of the funnel, the otherby letting it fall from a great height: that in the one cafethe air enters between the spiral circumvolutions whichthe water forms in the funnel, and in the other betweenthe drops into which a considerable part of it is reducedby the fall; that we cannot avail ourselves of both waysat once, the one impeding the effect of the other; andthat in either case the air holes under the throat, so neces-sary in other machines, can have no place, as they give avent to the air brought down from above.

Water falling from a funnel through a pipe with air holes.

The six-foot pipe, used in the foregoing experiment,continuing fitted into the air vessel, its upper orifice waswidened, that the small end of a funnel-shaped copperpipe, of the same bore with the preceding funnel, mighthang freely in it, without touching the sides. The funnel-pipe reached up to the reservoir, and was kept always full,that the water might receive little or no air but at thevacuity between the nose of the funnel and the leadenpipe.

In