The Table, showing the Thermal Windrose, accompanying this paper, was computed from the observations made at the Ben Nevis Observatory during the three years ending May 1887. It shows the mean temperatures, on the mean of the three years, of the different winds for each month, for the year, and for the seasons. The direction of the wind is observed to the thirty-two points of the compass, but in this table the temperatures are only shown for eight points, the intermediate points having all been added on to these eight points in the same way as that described by Mr Omond inh is paper on " Winds and Rainfall" in the Journal of the Scottish Meteorological Society, namely, the by winds were added to their adjacent octants, and the points half-way between the octants were on the odd day of each month added to the octant to their right, looking out from the centre of the compass card, i.e., they were veered. two points, and on the even days to that on their left, i.e., they were backed two points. The mean temperature of each direction of wind for each month was found by tabulating the hourly observations of temperature under the direction of wind observed at the same time, or under its octant as above, and then taking the arithmetical mean. When the wind was variable, or its direction doubtful, the temperature was entered in the column for calms and variables. These variable winds belong chiefly to the northern half of the compass, their existence being principally due to the abrupt and precipitous character of the north side of the Ben. The temperatures given in the table are those indicated by thermometers which in summer and autumn are protected in the regulation Stevenson screen, and in winter and spring in a smaller pattern of the same, which can be shifted up or down a ladder-like stand, so as to be always at or near the standard height of 4 feet above the surface of the snow.