THERMO-ELECTRIC DIAGRAM FROM - 200° C. TO 100° C. 



739 



determined with great accuracy — ( 1 ) the slope of the tangent at any point of the curve ; 

 (2) the value of the temperature at that point. Neither of these is a simple matter of 

 attainment. The following method was adopted to obtain the slope of the tangent : — 

 A fine dark-blue silk thread, about 60 cm. long, was threaded through two pieces of 

 sheet lead about the size of a florin, and the ends knotted to prevent the thread slipping 

 out. The thread was then stretched across the paper and carefully adjusted so that one 

 edge of it. was tangential to the outer edge of the curve at some point. A weight was 

 laid on each piece of lead to maintain the straightness and position of the thread during 

 the observations to be made. To prevent any error from parallax (through the thread 

 not actually touching the paper), a heavy strip of plate glass was laid along the thread 

 in the neighbourhood of the point of tangency, and by means of a strong light in a 

 vertical plane containing the thread, and with the help of a magnifying glass, it was 

 possible to adjust the coincidence of the edges of the curved line and thread to a 

 thousandth of an inch. The value of the tangent of the angle of slope, in the position 

 attained, was then easily found by reading from the millimetre paper the co-ordinates 

 of two points, one near each end of the thread, these two points being from 400 to 600 

 mm. apart. To find the temperature for the point of contact of the thread with the 

 curve, another geometrical property of the parabola was employed, as it is impossible to 

 determine by mere inspection the exact point at which two curves touch. If equal 

 distances be measured in opposite directions along any tangent to a parabola from its 

 point of contact with the parabola, the points thus found will be equidistant from the 

 parabola, when measured along lines parallel to the axis of the curve. To a first 

 approximation at least, for such points at distances from the curve not exceeding 6 mm. , 

 it was allowable to take the lines engraved on the millimetre paper parallel to the axis 

 of E.M.F. as being parallel to the axis of the curve. (In some curves, where the axis 

 of the parabola made a great angle with the axis of E.M.F. , theory showed that for 

 distances up to 4 or 5 mm. no error exceeding the limits of observational errors would 

 be made in adopting the same course.) Thus it was relatively easy to read a series of 

 pairs of temperatures at which two points on the tangent were 2 mm. distant from the 

 curve, 4 mm., 6 mm., or 8 mm. distant from the curve. If everything were correct, 

 each of these four pairs of temperatures would give the same mean temperature, and 

 this would be the temperature of the point of contact of the thread with the curve. 

 Usually a single pair of readings was satisfactory, and in this case the deviation taken 

 was 4 mm. Two specimens of such observations will elucidate the process. 



Deviation. 



Upper 

 Temperature. 



Lower 

 Temperature. 



Readings for 

 the Tangent. 



Temperature 

 of Contact. 



100 x tangent. 



4 mm. 



301° 



-15° 



397|+157| 

 250 + 100 



* 



158-6 



