During the week of September 5 to 11, 1897, kite-flights were made daily on Blue Hill. Twice the kites were maintained in the air, and continuous records were obtained during most of twenty-four hours. These records furnish an example of the small diurnal changes of temperature in the free air at short distances above the ground, which were deduced from the average changes at different hours and at different heights. From 2 p.m. of the fifth to 2 p.m. of the sixth, the altitude of the self-recording instruments varied between 500 and 1000 metres above sea-level, averaging about 700 metres and varying little from this height during much of the night. The times when the kite-meteorograph crossed the 700-metre level in ascending and descending were determined from its barograph trace, and the synchronous temperatures and humidities were read from the records of its thermograph and hygrograph. The results have been plotted in Plate XI., Figs. 1 and 2, together with the temperatures recorded simultaneously at the summit and valley stations of the Observatory and the humidities at the summit. Fig. 1 shows that the diurnal variation of temperature, well marked at the lower levels, is very slight or has entirely disappeared at 700 metres. Fig. 2 shows that the course of the relative humidity at 700 metres is exactly opposite in phase to that recorded at lower levels, for at 700 metres the minimum humidity was recorded at night and the maximum during the day, while the opposite conditions prevailed on the hill. Repeated kite-flights indicate that these are the normal conditions at the two levels. In Plate XI., Fig. 3, is plotted a curve from the hourly readings of the thermograph at the Blue Hill valley station (fifteen metres) during the week, and also a curve connecting temperatures recorded by the kite-meteorograph once or twice each day during the same week at a level of 500 metres, obtained in the way described or computed from the adiabatic change. All the night records show that it was decidedly warmer at the height of 500 metres during the night than it was at the ground, except during the cool wave on the seventh and eighth. Furthermore, the curves in Fig. 3 indicate a control of the surface temperatures during the day by those above. For instance, on the seventh there was a distinct flattening of the day curve, evidently because, as the temperature on the ground rose 10∞ above that at 500 metres, the air was in unstable equilibrium, and colder air descended to take the place of the surface air so that its temperature could rise no higher. On the tenth, the temperature at 500 metres was considerably greater than the mean of the day at the ground, and the air at the ground did not acquire the unstable condition in any volume until the warmest part of the day, so that the diurnal curve at the lower station forms a sharp peak. [Illustration: PLATE XI.--KITE OBSERVATIONS AT BLUE HILL, SEPT. 5-11, 1897.]