This historic book may have numerous typos and missing text. Purchasers can download a free scanned copy of the original book (without typos) from the publisher. Not indexed. Not illustrated. 1918 Excerpt: ... were superposed. During the winter there was, as stated above, no regular diurnal variation, the temperature difference sometimes remaining almost constant for days, when the ground was covered with snow, and at other times, in the absence of snow, undergoing large but irregular fluctuations. Because of their opposed seasonal variations and because of their similarity in all other respects, the autumn and spring values only were used in the preparation of the mean diurnal-variation curves shown in Fig. 2. Results were included only from those days on which the records were complete for all three elements or could be made complete by a small amount of justifiable interpolation. On this basis of selection the number of days available for computation of the mean hourly values represented in Fig. 2 was limited to 87. Curves A, B, and C, of Fig. 2 represent, respectively, temperature at upper thermometer minus the temperature at lower thermometer, P. D. between the terminals, and E2 of the insulated circuit,, being positive, and P. D. between the terminals Pi and P. of the non-insulated comparison circuit, P, being positive. During the last two months of the observation period, unfortunately some errors were introduced into the temperaturedifference records, due to a short piece of iron pipe which had been driven into the soil not very far from the upper thermometer for the purpose of making some auxiliary measurements with a mercury thermometer. Comparison of the curves obtained with this pipe in position with similar curves obtained in absence of the pipe showed that its effect was negligible except during the hours of intense sunshine. Further, the curve representing the mean hourly values of the air temperature in Observing House A, obtained f...