120 Mr. J. J. Waterston on the Expansion 



small flame, the tube after a pull will contract of itself with- 

 out being drawn; the glass is thus kept thick and strong; 

 and the flame being slowly withdrawn while the tube is kept 

 revolving, a strong conical point may be obtained. This is 

 filed down until a small needle-point opening appears. Before 

 filling the tube, a scale of twenty divisions to the inch is 

 finely and carefully marked upon the glass, as shown in 

 fig. 3. The mode of doing so is shown in fig. 6. The 

 feather edge of the ivory scale resting on the glass, the lines 

 were drawn with a fine-pointed lithographic steel pen, using a 

 watchmaker's lens. The ink was of gamboge, rubbed down in 

 water and darkened with black. When dry, the marked portion 

 of the tube was passed through a small flame, gradually heating 

 until, when near red heat, the lines are observed suddenly to turn 

 black; it is then to be quickly withdrawn. The lines are thus 

 ingrained upon the glass, and cannot be easily rubbed off. The 

 tube is then filled with distilled water to about the proportion 

 required. The small hole is then sealed by holding the extreme 

 point only in the flame for a few minutes. 



§ 7. To find total contents of the tube and the proportion occupied 

 by the liquid. — At a small Bunsen-flame the upper part of the tube 

 is dried, beginning at the top and working downwards ; then, im- 

 mersing in a water-bath of known atmospheric temperature, the 

 scale-reading of the lower part of the capillary cup is noted. The 

 liquid is then shaken to the other end of the tube and a similar 

 reading noted, taken by applying the ivory scale. Halfway 

 between these is the middle point. Add to this the distance 

 between this middle point and the zero of scale, the sum is the 

 value of zero of scale, to which we have to add the scale-reading 

 of an observation, and half the depth of the capillary cup is taken 

 by estimation as the volume to be added to reading of bottom of 

 cup, which is always a very distinct line by transmitted light. 



Next shake a small part of the liquid into the conical top; 

 then dry, and take readings. Thus, and by allowing for capil- 

 lary cup, we obtain the contents of half the tube, and thence 

 that of the liquid with great exactness. By further operations 

 of the same kind we may discover any tapering of the tube. 

 This was slight, and omitted in the reductions. 



§ 8. Graduating the tube thermometers. — The liquid em- 

 ployed for these thermometers was burnt linseed oil. After 

 the tubes were marked with scale divisions and sealed, they 

 were subjected for an hour to a temperature of 350° before gra- 

 duation. Three were constructed (the least of them is shown 

 in fig. 3, lashed with copper wire to the water-tube). In the 

 top of the funnel F, fig. 1, a large test-tube was fixed and filled 

 with lard. In this bath the standard mercurial thermometer, 



