156 THE ATLANTIC. [oHaP. 11. 
A.M. for each day, and the three intermediate lighter lines the 
hours of 10 a.m., 4 p.m., and 10 p.m. respectively. The spaces 
between the transverse lines have the value of th of an inch on 
the barometer scale referred to a column on the left side of the 
diagram, and to sth of a degree centigrade referred to a cor- 
responding scale on the right. The black curved line indicates 
the variations in atmospheric pressure reduced from bi-hourly 
observations taken with a mercurial barometer. Later in the 
cruise, we had out from Messrs. Elliott, of London, a self-regis- 
tering aneroid barometer, a beautiful instrument, but somewhat 
too delicate in its construction for use on board a ship. Owing 
probably to a trace of friction, the aneroid was usually very 
slightly behind the mercurial barometer in its indications, but 
otherwise the close correspondence between the two curves was 
most satisfactory as an evidence of the care and accuracy with 
which the barometrical observations were taken. 
The red line gives the variations of the temperature of the 
air projected from observations of a standard mercurial ther- 
mometer housed in the usual way. An instrument devised by 
Mr. Thomas Stevenson, of Edinburgh, with a view of elimina- 
ting any transient and accidental changes of temperature, and 
thus arriving at a more natural mean for the day, was observed 
twice daily, at 6 a.m. and at 6 p.m., by Mr. Tizard. A maxi- 
mum and a minimum registering thermometer have their bulbs 
immersed in a small flask filled with brine, and suspended on 
gimbals. The whole of the contents of the flask must thus be 
affected by a change of temperature before that change can be 
registered by the thermometers. The size of the flask is so ad- 
justed that there is ample time for this to occur with the grad- 
ual and normal cycle of daily temperature, while spasmodic os- 
cillations are neutralized before they are recorded. 
The general result of this arrangement is, that the maximum 
stands somewhat higher in Stevenson’s than in the ordinary 
maximum thermometer, while in the minimum there is very 
