34 Richards . — The Evolution of 
another source of trouble, since any disturbance of their 
constancy of temperature was followed by uncertain deflections 
of the galvanometer. Finally, however, they were encased in 
cork, which was found to thoroughly protect them from 
outside influences. As has already been said, a large portion 
of the element itself was also protected with cork and sealing- 
wax. The above are the most weighty sources of error as 
regards the galvanometer and the thermo-electric element ; 
in the actual experiments on living plants there are other 
precautions to be taken, of which more will be said later. 
Frequent experiments established the fact that, with the 
scale used and at the distance at which it was placed from 
the mirror, one division had the value of 0-07° C. It was 
found for the purpose desired that water of different tempera- 
tures, measured by standard normal thermometers, was as 
accurate as a more complicated apparatus with oil. Alcohol 
was also used, and the results obtained thereby agreed with 
the determinations made with water. In the experiments 
which will be found given at the end of this paper, one 
division of the galvanometer scale is taken then as the 
equivalent of 0-07°. 
In the manipulation of the plants under experimentation, 
there are also many precautions to be taken against possible 
errors. In the first place the plants must be kept in an 
atmosphere saturated with moisture, both to prevent their 
withering and also to avoid changes of temperature due to 
evaporation from the cut surfaces of the injured plant. It is 
also necessary that they should have acquired the temperature 
of the warm room. A day or two was always allowed before 
the plants brought into the warm room were used. Another 
point which is exceedingly important is the placing of the 
thermo-needles in the objects. When the needle is thrust too 
much into that portion of the wound which gapes open, 
naturally little or no effect is observed ; but when it is properly 
arranged near the base of the cut where the sides come 
together, the needle-point receives the full influence of both 
the cut surfaces. The latter method was always practised, 
