404 
PACIFIC EXPLORATION: L. J. BRIGGS 
therefore does not need to be known with an accuracy greater than 
0.5 mm. Determinationsof the constant agreeing to 0.1 mm. can be 
obtained with the auxiliary manometer to which reference has aheady 
been made. 
Thermal hysteresis of the glass envelop. — This phenomenon is generally 
recognized in precision thermometry, and results in the so-called de- 
pression of the zero. This effect would tend to introduce an uncertainty 
in the determination of the difference in the acceleration of gravity at 
two stations where it is necessary to remove the apparatus from the 
ice in proceeding from one station to the other. A slight change of 
this kind (2 parts in 100,000) was observed at Balboa, following the 
dismantling of the apparatus and its transportation across the isthmus 
by rail, owing to the slides in the canal. These instruments were made 
of German glass. Waidner and Dickinson^ state that thermometers 
made of the best boro-silicate glass show a depression of the zero of 
0?03. This would correspond to a change of 1 part in 200,000 in the 
volume of the bulb. Consequently if boro-silicate glass were used in 
the construction of the glass parts of the apparatus, the error due to 
thermal hysteresis would fall within the limit set in this discussion. 
Disturbances arising from the motion of the ship. — Three classes of dis- 
turbances are encountered on board ship: 
1. Tremors, due to the engines or auxiliary machinery or to the impact 
of waves. The effect of such disturbances can be greatly reduced by suspend- 
ing the ice tank from spiral springs. This form of support does not eliminate 
all vibration. Slight tremors however appear to be advantageous rather 
than otherwise, as they help to bring the mercury surface to its true position 
in the bulb. 
2. Horizontal translation, due chiefly to rolling, which tends to swing the 
apparatus from its vertical position. Such translations produce two effects 
which can be made to counteract each other to some extent, (a) The vertical 
component of the column is shortened by an amount proportional to the sine 
of the deflection from the vertical. Such deflections therefore tend to increase 
the length of the column, (b) The centrifugal force resulting from the de- 
flection tends to depress the column when the latter is mainly below the point 
of support. This can be eliminated by mounting the column so that it is 
bisected by the gimbal plane. By suitably adjusting the position of the 
column, the two effects can thus be made to compensate in part.^° 
3. Vertical motion, due to rolling and pitching or to the rise and fall of the 
ship as a whole in a heavy sea. - This is the most serious of all the disturbances 
to contend with, for the motion is accelerated, and is superimposed on the 
gravitational acceleration. Extensive damping through the use of the 
capillary column has been employed in the apparatus here described to reduce 
