[ 294 ]

out from them, the air vessel in this cafe yielding none ;so that here the air must have been introduced at the topand passed down the funnel, and afterwards escaped whereit first found a vent. To be further satisfied in this point,I repeated the experiment with a somewhat different ap-paratus, in the following manner.

Water falling through a funnel.

The glass receiver of an air pump, about two feethigh, open at both ends, had its lower end immersedabout seven inches in a vessel of water, and supported ata proper distance above the bottom for the free passage ofthe water under the edges. A brass plate being pressedclose on the top, with leather between, a glass funnel,about twelve inches deep, and above half an inch diame-ter in the throat, was fixed into a hole in the plate ; andinto another hole was fitted a small blowing pipe.

A stopper being introduced into the funnel, till thewater it was filled with had become perfectly quiet, andthen cautiously removed, the water run in a stream, whichfalling into that in the receiver, produced air bubbles :but no blast issued from the pipe; and when the pipe wasstopt, the water in the receiver did not sink lower thanthe level of that in the outer vessel, whereas, if any airhad entered with the water, and been compressed in thereceiver, it must have forced a proportional quantity ofthe water out below.

The funnel was then supplied from a pipe, by whichthe water was made to dash against one side of it. By thismeans the fluid received a spiral motion, and twirlinground the funnel, left a large vacuity in the middle,reaching down sometimes to the funnel’s throat. Thestream, as it run through, was also twisted ; a sensible

blast