328 
BOTANICAL GAZETTE [APRIL 
the mercury will hang on the wire, and regulation thereby becomes 
Pic. t 
impossible. The seal must be gas-tight. (3) The 
constriction of the capillary at C must be narrow 
enough so that a short section of the mercury column 
above will not slide down past the constriction even 
with slight jarring. It must be possible, however, to 
shake mercury down past the constriction. (4) The 
band holding the upper contact wire should not cover 
the contact point itself, otherwise adjustment is 
ifficult. 
To adjust the instrument for a given temperature 
an excess of mercury is shaken down below the con- 
striction to unite with the column in the capillary. 
The remainder of the mercury above the constriction 
is allowed to flow, with gentle tapping or shaking, if 
necessary, into the upper bulb, the thermometer 
being inclined during the process. The constriction 
should be narrow enough to prevent the mercury 
below from flowing past it. The bulb of the regulator 
is now plunged into a large beaker of water kept 
exactly at the temperature for which the regulator is 
to be set. The mercury column will stand a little 
above the upper contact, since an excess of mercury 
was first shaken down past the constriction. The 
length of the mercury column above the contact is 
noted, and a section of nearly the same length is 
forced above the constriction by carefully lowering 
the bulb into another vessel of water kept at a suffi- 
ciently higher temperature. The bulb is then replaced 
in the first vessel and the process repeated. Finally, 
only single globules are forced past the constriction 
until the end of the column stands exactly at the upper 
contact wire, when the bulb is kept in water at the 
desired temperature. If the constriction has been 
properly made, the short piece of mercury column 
above it may be left in place. 
This regulator is designed to actuate a telegraphic relay which 
interrupts the heating current. The installation shown in fig. 2 
