34 MR JOHN AITKEN UN THE NUMBER OF DUST PARTICLES IN THE 



particles actually present, the estimates of clearness were always made before the 

 dust, temperature, and humidity readings were taken, and were entered in the field note- 

 book along with the direction and force of the wind. The clearness was measured by 

 estimating the amount to which a mountain was hazed. Then, knowing the distance of 

 the mountain selected, it is easy to calculate the extreme limit of visibility of that particular 

 sample of air. For instance, if the hill selected were 20 miles distant, and half hazed, 

 then 40 miles would be the extreme limit of visibility of the air at the time. In this way 

 all the observations can be reduced to one scale for comparison. It should be stated that 

 at Kingairloch all observations of this kind bad to be done in a south-easterly direction, 

 as only in this direction could distant hills be seen. In the other directions the view 

 was closed in by mountains only a few miles distant. 



In working out these observations on the haze they were divided into sets, and 

 arranged in tables according to the humidity of the air at the time, for the reason already 

 given in Part I., — all the observations taken when the wet-bulb depression was 4° being 

 entered in one table, all those taken while it was 5° in another, and so on. (When the 

 wet-bulb depression is less than 4° the real humidity is frequently uncertain, because 

 when the depression is slight it is generally after rain, and the wet-bulb depression is 

 thus very much influenced by local conditions, such as wet ground, trees, &c.) The 

 observations in the different tables were then rearranged, — the observation which has 

 the greatest number of particles being put at the top of the column, then the next 

 greatest, and so on, down to the observation which has the fewest particles. Then 

 as the humidity of all the observations in each table is about the same, the limit of 

 visibility of the observations at the top of the table ought to be the least, as at this end 

 of the column there is most dust, and the limit of visibility should increase towards the 

 foot of the column, and be greatest in the last entered observation. 



When this is done and all doubtful dust observations are rejected, as well as all 

 observations taken in or immediately after rain, on account of the uncertainty of the 

 value of the humidity readings taken under the conditions, it is found that in the tables 

 for depressions of 4°, 5°, and 6°, the order of the limit of visibility is, in a general way, 

 inversely as the number of particles. In all three tables the lowest limit is associated 

 with the greatest amount of dust, and the widest limit with the least. There is, how- 

 ever, some mixing of the relative positions of the intermediate numbers, as might be 

 expected. 



The results obtained by this method of working are not so satisfactory when we 

 come to test the effect of the dust in very dry air, such as that giving depressions of 7° 

 and more. In the tables for high depressions the figures for the limit of visibility are 

 very much mixed, large numbers appearing near the top of the column as well as near the 

 foot. There are many reasons for this. One is the conditions under which the estimates 

 of haze have to be made. When the air is very dry, it is clear with even much dust in 

 it. All estimates have, therefore, to be made on a thin haze, as seen on distant and high 

 mountains, and it is difficult to estimate a thin haze. And, further, as these estimates 



