The same result can be brought about without the aid of a stream of water in the following manner. Take once more a pedestal perfectly air-tight, A B C D (fig. 79), having a partition across the middle, E F. Let the tube from the mouth of the animal, G H K, lead into the pedestal, and another tube, L M N, pass through the surface A D and the partition E F. In the tube, L M N, perforate a hole, X, just above the partition E F, and let another tube, O P, fit into it closely, having a hole, R, corresponding with the hole X. To the tube O P attach a figure of Pan, or some other figure with a fierce look, and, when the figure is turned towards the animal, it shall not drink, as though frightened; when the figure turns away, it shall drink. Now, if we pour water into the compartment A D E F through a hole, G´, which must afterwards be carefully closed with wax or some other substance, it will be found that, if the holes R and X are made to coincide, the water which was poured in will pass into the compartment E B C F. As A D E F becomes empty, it will attract the air through the mouth of the figure, which will then drink when a cup is presented to it. [Illustration: Fig. 79.] [Illustration: Decorative floral border] APPENDIX. There are four MSS. of Hero contained in the British Museum. Two belong to the Harleian Collection: they are numbered 5605 and 5589, and are assigned, the first to the fifteenth, the second to the sixteenth century; they are distinguished in the table below by the letters _a_ and _b_. The others are among the Burney MSS., (108 and 81), and are both assigned to the sixteenth century: they are cited as _c_ and _d_. Of these MSS. the first and third, (_a_ and _c_), are by far the most trustworthy: they generally supply the clauses missing in the printed text, and often furnish emendations of passages otherwise corrupt. The other two agree, for the most part, with each other, and with the printed text, both in the omission of numerous clauses and in the grossest errors in individual words. It is thus rendered probable that the defects of the Paris edition are in great measure due, not to the Editors, but to the MS. followed. In the following table of emendations made in the text of Hero as it appears in the _Veterum Mathematicorum Opera_, those only have been noticed which appeared curious or important. Many more have been passed over in which the correction was obvious even without the aid of any MS. The references are to the page and line of the Paris edition.