Professor Draper's Description of the Tithonometer. 413 



peppercorn be left, the operator will be abundantly satisfied 

 with the sensitiveness of his instrument. Commonly, at first, 

 a large residue of hydrogen gas, occupying perhaps an inch or 

 more, will be left. It is to be understood that even this large 

 surplus will disappear in a few hours by absorbing chlorine. 

 But this is not to be waited for ; as soon as no further rise 

 takes place in a minute or two, the siphon is to be inclined on 

 one side, and the residue turned out into the open leg. 



Now, recurring to what has been said on the equilibrium, 

 it is plain that this excess of hydrogen arises from a want of 

 chlorine in the tithonometric liquid. A proper quantity must 

 therefore be furnished by proceeding as follows. 



The sentient tube being filled with the liquid by inclination, 

 connect the polar wires with p q, as before. These may be 

 called generating wires. Allow the liquid to rise in b c, until 

 the third platina wire z, which may be called the adjusting 

 wire, is covered an eighth of an inch deep. Then remove the 

 negative wire from the cup p into the cup r, and now the con- 

 ditions for saturating the liquid are complete; hydrogen esca- 

 ping away from the surface of the liquid at z, and chlorine 

 continually accumulating and dissolving between x and d. 

 This having been carried on for a short time, the gas in a d 

 is to be turned out by inclination and the instrument re- 

 charged. That a proper quantity is evolved is easily ascer- 

 tained by allowing total condensation to take place, and ob- 

 serving that only a small bubble is left at a. 



It will occasionally happen in this preliminary adjustment, 

 that an excess of chlorine may arise from continuing the pro- 

 cess too long. This is easily discovered by its greenish-yel- 

 low tint, and is to be removed by inclining the instrument 

 and turning it out. 



Thus adjusted, everything is ready to obtain measures of 

 any effect, there being two different methods by which this 

 can be done, — 1st, by continuous observation; 2nd, by inter- 

 rupted observation. 



Of the Method of continuous observation. — This is best de- 

 scribed by resorting to an example. Suppose, therefore, it is 

 required to verify Table I., or, in other words, to prove that 

 the effect on the tithonometer is proportional to its time of 

 exposure. 



Put on the cap of the sentient tube a d, connect the polar 

 wires with p q, and raise the liquid to zero. 



Place the tithonometer so that its sentient tube will receive 

 the rays properly. 



At a given instant, marked by a seconds watch, remove the 

 cap A D, and the liquid at once begins to descend. At the 

 end of the first minute read off the division over which it is 



