OF ASPECT, ROTATION , LIBRATJON, ETC. 171

monly supposed. We shall not, therefore, presentthem here,* but proceed at once to mention twoexperimental proofs of the fact in question. The firstexperiment is very simple. Let a tolerably heavyball be suspended by a long fine cord. Let it be lefthanging until all signs of twisting have passed away;then, having placed a mark upon the ball anywhereexcept near the top or bottom, cause it to swing ina circle, communicating this motion by means of thestring held at a point high above the ball, so thatno rotational movement can be imparted. It will befound that the mark continues always to be directedtowards the same point of the compass, not turning soas always to bear in the same direction with respect tothe centre of motion. The second was suggested byGalileo, who pointed out that if a body be set to floatin a basin of water, and this basin be held out at arm’slength while the holder turns round, it will be foundthat the floating body does not partake in the turningmotion; so that the side turned towards the holderof the basin at the beginning is turned directly awayfrom him when he has made half a turn. It is, how-ever, by no means easy to carry out this experiment ina satisfactory manner, the most striking phenomenon

* It may suffice to remark that if a body circuits round E in themanner shown in fig. 55, the total quantity of work done accordsexactly with that due to the imparted velocity ; but if it moves inthe manner shown in fig. 56, the amount of work done exceeds thatdue to the imparted velocity by the amount corresponding to onecomplete rotation of the body.