798 MR J. Y. BUCHANAN ON THE 



normal difference clue to solar and terrestrial radiations alone would be about 

 17° R 



Nocturnal heating is observed in foggy weather in most of the months, and it takes 

 the form of a sharp rise of temperature between midnight and 1 a.m., which then falls 

 gradually to a minimum at or near the hour of sunrise ; but there are exceptions to this, 

 as in November, and particularly in January, when the temperature rises very uniformly 

 from a minimum at midnight to a maximum at noon, and then falls again as uniformly 

 to the midnight minimum again. 



In dealing with the winter months at Ben Nevis, it must not be forgotten that the 

 solar influence is very small. Lying in lat. 56°48' N. the sun's meridian altitude at the 

 winter solstice is only 9°35', and the length of the day is under 6|- hours. It is, there- 

 fore, chiefly at this season that we might expect terrestrial or geographical influences to 

 produce their most apparent effect. With the march of the season the influence of the sun 

 increases very rapidly, and it has a tendency to obliterate the effects of other agents, 

 especially in the hours of the day when its heating power is increasing or diminishing 

 most rapidly. In the summer months, when the diurnal range of temperature is con- 

 siderable, it is only in the neighbourhood of the epochs of maximum and of minimum 

 temperature that other influences can make themselves felt. At these times, and 

 especially at the time of maximum temperature, the heating and cooling influences are 

 for a time in a condition approaching equilibrium, during which the temperature remains 

 nearly constant, and its curve runs sensibly parallel to the line of time-abscissae. Here 

 we might expect other influences to show themselves ; and, in fact, if we inspect the 

 curves, especially those relating to clear weather, we see that most of them show great 

 irregularities in the neighbourhood of the extremes. Nearly all the clear weather curves 

 have strongly-marked irregularities near the date of minimum temperature, and most of 

 them, as February, March, August, and October, show similar irregularities near the date 

 of maximum. In foggy weather the hour to hour irregularities of mean temperature are, 

 as might be expected, much less striking, with, however, the exception of the month 

 October, when we have a very remarkable oscillation of temperature from hour to hour 

 during the whole morning. If we were to calculate the mean temperature of the morn- 

 ing from the odd hours, 1, 3, 5 . . . . 11, we should find it quite half a degree lower than 

 if we used the even hours, 2, 4 .... 12. One reason for the greater uniformity of the 

 temperature curves in foggy weather than in clear weather is that there are no effects of 

 alternating cloud and sunshine. In clear weather we have these effects, and to some 

 extent they must be held responsible for the irregularities apparent. This, however, 

 applies only to the daylight hours, and we see that the irregularities in clear weather 

 are by no means confined to these hours. 



In dealing with the barometric pressure we have found something very similar to 

 what we have just noticed in regard to temperature, namely, that the curves for foggy 

 weather are much smoother and more uniform than those for clear weather, and the 

 clear weather barometric curves present nearly as irregular and serrated an outline as 



