33 
Heat by Wounded Plants . 
even a fifth of one division could be read with ease. There 
are many sources of error in the use of this instrument which 
require great precautions to exclude ; all due care was taken 
to do so in so far as they might influence the work in hand. 
The galvanometer was naturally exceedingly sensitive to 
masses of iron ; hence, so far as possible, all iron objects, 
particularly movable ones, were removed from the room. 
Another necessity is that the room in which the experiments 
are being carried on should be of constant or nearly constant 
temperature. This, together with a no less important condi- 
tion, that of a perfectly solid foundation for the instruments, 
was realized in the room in which these experiments were 
performed ; where the temperature remained for the most 
part exceedingly equable, within a few tenths of a degree, 
and the cement floor being directly on the ground, was quite 
free from vibration. The thermometer on the level of the 
table where all the plants were kept, registered a little over 
24 0 C. For rendering the galvanometer astatic, a bar-magnet 
was used and so arranged that the observer at the telescope 
could alter its position and thus control the zero-point. The 
change of the declination of the magnet of the galvanometer, 
due to the earth’s magnetism, could not be disregarded ; hence 
before each observation a zero-point was established by cutting 
out the thermo-element from the circuit with an ordinary 
mercury commutator. 
In the use of the thermo-electric element Rodewald 1 gives 
a number of precautions against errors, which were followed 
in so far as necessary in this work. The most important is 
the complete insulation of the element by means of an even, 
but not too thick, coat of hard shellac, which at the same 
time serves to protect the metal from the chemical action of 
the juices of the plant. A careful watch was kept to see that 
the shellac coating always remained intact, and very fre- 
quently the element was washed clean and re-lacked. The 
binding-screws which connect the iron wires of the element 
with the copper wires coming from the galvanometer were 
1 1. c. xviii, p. 281. 
D 
