The indications depend upon the hydrostatic principle, that floating bodies displace a quantity of the fluid which sustains them equal to their own weight. According, therefore, as the specific gravities of fluids differ from each other, so will vary the quantities of the fluids displaced by the same body when floated successively in each. The specific gravity of distilled water, at the temperature of 62 deg. _F_, being taken as unity, the depth to which the instrument sinks when gently immersed in such water is the zero of the scale. The graduations extend from 0 to 40; the latter being the mark which will be level with the surface when the instrument is placed in water, the specific gravity of which is 1.040. In recording observations, the last two figures only--being the figures on the scale--are written down. Sea-water usually ranges from 1.020 to 1.036. A small tin, copper, or glass cylinder is useful for containing the water to be tested. It should be wider than the hydrometer, and always filled to the brim. If fitted to a stand, which is supported by gimbals, it will be very convenient. Water in a bucket, basin, or other wide vessel, acquires motion at sea, and the eye cannot be brought low enough (on account of the edges) to read off the scale accurately. Errors of observation may occur with the hydrometer, if it be put into water without being clean, or without being carefully wiped. The instrument is extremely accurate if correctly used. It should be kept free from contact with the sides of the vessel; and all dust, smears, or greasiness, should be scrupulously avoided, by carefully wiping it with a clean cloth before and after use. Whenever the temperature of the water tested differs from 62 deg., a correction to the reading is necessary, for the expansion or contraction of the glass, as well as the water itself, in order to reduce all observations to one generally adopted standard. Negretti and Zambra's hydrometer, with thermometer in the stem, shows the density and temperature in one instrument. For the following Tables we are indebted to the kindness of Admiral FitzRoy:-- TABLE for reducing observations made with a BRASS HYDROMETER, assuming the linear expansion of brass to be 0.000009555 for 1 deg. F. The correction is additive for all temperatures above 62 deg., and subtractive for temperatures below 62 deg.