74 



7. The equations for u- and y are not so complex that they can not be used. 

 In Table II are given the values of 2' deduced by formula (8). It will be noted 

 that the last frame is about sixteen times as thick as the first, yet the greatest 

 difference in these values is but a little more than one part in two hundred. Of 

 the results for the first four frames, the greatest difference is one part in seven 

 thousand. The thicker frames can not be expected to give such consistent results, 

 as the water tends to creep in between the thin layers of which the mica sheet is 

 made up. 



The Temperature Coefficient. 



Previous determinations of the temperature coefficient of surface tension 

 give results not more consistent than the values obtained for the tension itself. 

 Brunner gives the coefficient as .14 dynes per degree, and Merian as .253 dynes. 

 The latter result is almost double the former. Other observers give intermediate 

 values. In view of these differences, I concluded to make a determination of the 

 temperature coefficient by the mica frame maximum weight method. This 

 investigation is not yet completed, so I shall not go into detail. 



I am using a Troemner balance, No. 5, easily sensitive to one one-hundredth 

 milligram. The arrangement of the balance and box or closet is very much the 

 same as in Hall's experiment. Inside the wooden box I have a double-walled tin 

 box, open on the side next the glass door. The space between the walls of the 

 tin vessel (the walls being about two inches apart) may be filled with a bath to 

 regulate the temperature of the enclosure. This temperature is obtained by read- 

 ing three thermometers, placed in different positions. A rotary fan is used to 

 equalize the temperature throughout the enclosure. It is arranged so that the 

 water whose coefficient is to be determined is siphoned in and out of the vessel 

 inside, without opening the door or disturbing the balance. 



I have tried four methods of regulating the temperature of the enclosure. A 

 current of air from a blower giving a very constant pressure was passed through 

 an iron pipe heated by from one to a dozen Bunsen burners, and then through 

 the tin box. By varying the air supply and the number of burners, a fairly con- 

 stant temperature could be maintained. But I was not able to raise it above 50°. 

 I next tried a water bath, the water being heated in a tube outside, but con- 

 nected with the box — somewhat upon the principle of an incubator. I could 

 easily maintain any desired temperature between 0° and 70°. But for higher 

 temperatures I found that the convective circulation of the water was too slow to 

 prevent the water in the tube from boiling. I substituted oil for water, but I was 

 not able to extend my observations above 80°. 



