ents--those of Joule and
Rowland--give values exceeding by several units that found by Hanssen.
A committee of the British Association, appointed for this purpose,
reported in 1876 that sixty of the most reliable of Joule's
experiments gave the mean value 774.1. The experiments were made with
water at a temperature of about 60 deg. F., according to the mercurial
thermometer, and reduced to its value at the temperature of melting
ice, according to the formula given by Regnault for the variation of
the specific heat of water at varying temperature under the constant
pressure of one atmosphere. According to this formula the specific
heat of water increases with the temperature above the melting point
of ice, so that the equivalent would be somewhat less at 32 deg. F. than
at 60 deg. F. It will be found in Regnault's _Relation des Experiences_
that he experimented on water at high temperatures, but more recently
Professor Rowland has found that the specific heat of water is
_greater_ at 40 deg. F. than at 60 deg. F., thus reversing between these
limits the law given by Regnault; the increase, as given by the most
probable values, being, roughly, about 1/250 of its value at 60 deg. F.
The proper correction due to this cause would make the equivalent over
777 foot pounds, instead of 774.1. Professor Rowland's experiments,
when reduced to the same thermometer, same temperature, and same
latitude as Joule's, agreed very nearly with those of the latter,
being about 1/1000 part larger; so that the chief difference in the
ultimate values consists in the reductions for temperature and
latitude. The force of gravity being less for the lower latitudes, the
number representing the mechanical equivalent will be greater for the
latter, since the unit pound mass must fall through a greater number
of feet to equal the same work; so that the equivalent will be greater
at Paris than at Manchester. Professor Rowland also found that the
degrees on the air thermometer from 40 deg. F. upward to above 60 deg. F.
exceeded those on the mercurial thermometer throughout the
corresponding range, and that from 40 deg. to 41 deg. the degree was between
1/150 and 1/200 of a degree larger on the air thermometer than on the
mercurial. Although this fraction is too small to be observed by
ordinary means, yet, if it exists, it cannot be ignored if absolute
values are sought. Regnault employed the air thermometer in his
experiments, while Joule used the mercu