[ 292 ]

Judging that the motion of the water must be re-tarded in this last circumstance, I measured by a pen-dulum the times in which equal quantities of water runthrough the crane in both cafes ; and found, in manytrials, that the quantity which took the time of a hundredswings of the pendulum to run in when the sucking pipewas open, run in ninety-three, and sometimes ninety-two, when it was stopt.

As these differences seemed to proceed from air intro-duced into the water through the lateral pipe ; I tried tomake this air sensible, by raising the vessel which re-ceived the water from the crane, and keeping the noseof the crane immersed in it. As often as the suckingpipe was opened, air bubbles arose in the water of thereceiver, and fresh bubbles succeeded while it continuedopen ; but so long as it was kept stopt, no air bubbleswere seen.

To collect the air, a cask without a bottom was funknine or ten inches in a tub of water, and the nose ofthe crane inserted into a hole made in the top of thecask : into another hole in the top was fitted a small pipefor giving vent to the air ; and within the cask was fixedan inverted mortar for the stream to fall on. So long aswater was kept running though the crane with thesucking pipe open, a sensible blast issued from the blow-ing pipe of the calk, and a burning coal exposed to itwas excited in the same manner as by a common bellows:the sucking pipe being stopt, no blast was perceived, norwas any motion produced in the flame of a candle ap-plied to the orifice.

It appears therefore that water, running down throughan upright pipe, and filling its bore, admits air to enterthrough a lateral pipe : that after this admission, thewidth of the column of water contracts, the introduced

5

air