WITHDRAWN FROM THE ACTION OF GRAVITY. 429 



much less persistence tlian the large one; they frequently break before the 

 'measurements can be accomplished. Lastly, although the manometer of my 

 instrument has an interior diameter of one centimetre, equilibrium is established 

 therein very slowly, and great errors would result if uo regard were paid to this 

 circumstance. 



We will further remark that, when a bubble is just formed, there is, in gen- 

 eral, a drop suspended at its base, the weight of which somewhat elongates the 

 bubble in a vertical direction. To get rid of this small additional mass without 

 causing the film to break, we lightly touch it with one of the corners of a piece 

 of filtering paper; the drop is thus partially absorbed, and the same operation is 

 repeated with the other corners of the paper, wntil the excess of liquid has en- 

 tirely or apparently disappeared. A last remark should be added : if, when a 

 bubble is to be formed, we plunged the whole funnel of the adjutage into the 

 liquid, this would ascend, by capillary action, into the interior of the narrow tube 

 which surmounts this funnel, and would be only, in part, expelled by the breath, 

 so that, after the inflation of the bubble, it might collect in a small mass at the 

 lower part of the tube in question, and thus interrupt the communication between 

 the bubble and the manometer. To avoid this inconvenience, the whole exterior 

 surface of this little funnel should, at the commencement, be moistened with 

 glyceric liquid, and then the extreme border only of the funnel be immersed. 



Let us see now by what steps we have proceeded. For the largest diameters 

 of the table to be given in the next paragraph, we first inflated the bubble to 

 about 6 centimetres, applied the ball of wax, and then waited five minutes; after 

 which we again opened the tube of insufilation, allowed the bubble to diminish 

 until it seemed to have the desired diameter, and then applied the wax anew. 

 For all the smaller diameters we began by inflating the bubble to about 4 cen- 

 timetres, and after having applied the wax, waited ten minutes before allowing 

 the diminution to take place. Previous trials had shown that, with these pre- 

 cautions, the diameter afterwards remained invariable. The bubble on which 

 it was proposed to operate having thus attained the desired dimension, we re- 

 moved the drop suspended at its lower part, inclined the instrument to right and 

 left with a view to moisten well the two branches of the manometer a little above 

 the two levels, and we measured the diameter. We then allowed the instrument 

 to remain at rest for ten minutes, in order to give time for the establishment of 

 the equilibrium of the manometer, when we measured the pressure, and five 

 minutes afterwards measured it again. If the results of these two measurements 

 were not exactly the same, we made, after the lapse of another five minutes, a 

 third measurement, and so on, until two identical results were obtained in suc- 

 cession, or the difference was in a contrary direction from those preceding; in 

 the former case, the two last results were considered as giving the value of the 

 pressure ; in the latter, as their difference would be attributable to a slight error 

 of observation, we assumed, for the value of the pressure, the mean of these two 

 results. 



§ 28. The following table contains the results of these experiments ; I have 

 arranged them, not in the order in which they were obtained, but in the ascend- 

 ing order of the diameters, and I have distributed them into groups of analogous 



