DETERMINATION OF CRITICAL POINT OF DEATH l6l 



i° C. the drum was rotated, say, through a distance of 2 mm., 

 and the position of the spot recorded. In this way, by con- 

 necting the recorded points, a curve was obtained, in which 

 the length corresponding to each temperature was known. 

 In this curve an abrupt inversion, due to sudden death- 

 contraction, was found to occur at about 59 C. The curve 

 thus obtained, however, though the successive points recorded 

 were very near each other, is the result of intermittent 

 observations. Again, two observers were required, one to 

 read the temperature, and the other to take the record. It 

 was therefore subject to error of thermometric reading. 



Means of obtaining automatic record.— For this reason 

 I was desirous of obtaining a curve which should be con- 

 tinuous and practically automatic, the plant itself being made 

 to record its own variations of length, and its own death-point. 

 The problem resolves itself into that of making the reflected 

 spot of light partake of two motions simultaneously, namely, a 

 horizontal movement proportional to the change of tempera- 

 ture, and a vertical movement proportional to change of length. 

 The horizontal, or thermometric, component of the movement 

 I secured as follows : I constructed a thermo-electric element 

 of iron and nickel, one junction of which was placed in melting 

 ice, and the other junction in the vessel of water containing 

 the specimen whose temperature was being subjected to change. 

 This element was placed in circuit with a resistance box and 

 a sensitive reflecting galvanometer. The amount of the 

 movement of the galvanometer spot of light could now, by 

 interposing suitable resistance, be brought to any appropriate 

 value. In my experiments, with a particular galvanometer, 

 the movement of this spot of light, for each degree of rise of 

 temperature, was 2*5 mm. — i.e. one-tenth of an inch — when 

 the recording surface was at a distance of 125 cm. from the 

 galvanometer. This extent of movement was quite sufficient 

 for the purposes of the experiment, as it enabled estimates to 

 be made with ease, correct to one-fifth of a degree. By inter- 

 posing smaller resistances, however, one-twentieth of a degree 

 could easily be discriminated. The excursion of the spot of 



M 



