which Mossman, at my suggestion, adopted. Another important consideration in the placing of thermometers on board ship is to see that they are placed in a thoroughly exposed position in good louvred screens, which can get a free breeze across them: not up against a bulkhead or under a bridge or any other such place. On the Scotia they were fitted up against stanchions on each side of the quarter on the poop deck, about five feet above the level of the deck, and projecting as much as 18 inches clear of the ship's side, where they were in an absolutely open position.
When the temperature observations were being made the dry and wet bulbs on both sides of the ship were read, and the readings of those on the weather side were recorded as the correct ones. It is interesting to note that errors of several degrees were observed on the leeside thermometers on many occasions, especially when the Scotia was in high latitudes and low temperatures prevailed. Furnaces, galley and cabin stoves, and the general higher temperature of the ship itself all affected the readings. On rare occasions when the wind was absolutely ahead, and it was thought that both the port and starboard thermometers might be affected, Mossman used sling thermometers on the foc'sle head, but these, as a rule, did not vary a tenth of a degree Fahrenheit from
