ADDENDUM TO PAPER ON WIND PRESSURE. 
743 
The plane surface was 10 inches long and 6 inches wide. 
'With this numerous experiments were made at angles of incidence 
varying from 90° to zero with very consistent and satisfactory results, 
which are shown in Fig. 2, in comparison with the results of other 
experimenters, and are further stated in the subjoined table : — 
Angle of Incidence. 
r~ 
Hutton. 
Normal pressure according to 
Duchemin. Crosby. 
a 
Kernot. 
10° 
24 
34 
10 
28 
20° 
45 
61 
54 
30° 
66 
80 
37 
72 
40 3 
83 
92 
83 
50° 
95 
99 
68 
90 
60° 
100 
100 
95 
70° 
100 
100 
91 
98 
80° 
101 
100 
99 
90° 
100 
100 
100 
100 
The above figures are in the case of Hutton and Duchemin taken 
from Professor Warren’s Engineering Construction , p. 2S6, and in the 
case of Crosby from Engineering of 6th June, 1890, p. 690. 
An inspection of Fig. 2 will show that the results I have now 
obtained when plotted give a much more regular curve than those 
of the other authorities mentioned, a fact which should, I think, be 
taken as greatly in favour of their probable accuracy. In the most 
useful part of the range, that corresponding to ordinary roofs, they 
are very nearly a mean between Hutton’s and Ducheiuin’s. 
In the previous paper, p. 579, it was pointed out that in ordinary 
buildings in which roofs are combined with walls, and occasionally 
parapets, the wind is deflected upward and away from the roof, and the 
normal pressure largely reduced. A considerable number of further 
experiments have been made to illustrate this point, and all have 
emphatically vindicated the conclusion that a roof supported on walls 
experiences very much less pressure than one supported on columns only, 
and under which the wind can pass freely. A. number of experiments 
with small flags were made to determine the curve taken by the air 
current passing the top of the wall, with the result given in Fig. 3, 
where, if the height of the wall be unity, the distances AB and BC are 
•5, while BD is *5, CE is *75. Above the curve A DE the full current is 
felt, but below it are only eddies or nearly still air. The pressure 
upon any roof not rising above ADE will therefore be but an 
insignificant fraction of what has been hitherto supposed. Thus the 
limit of what has been called the “wind shadow” is accurately 
defined. It was notable how very definite was the curve ADE, 
within and without which the behaviour of the little flags was entirely 
different. 
With the aid of the apparatus used to obtain the results shown in 
Fig. 2, numbers of experiments were made as to the wind pressure 
on roofs with and without walls, and the general results of the paper 
of 1893 were confirmed. In the case shown by Fig. 4, the following 
table gives the conclusions : — 
Angle of roof 
/ 
Without any 
Normal pressure 
With wall. 
-> 
With wall. 
to horizontal. 
wall. 
‘ AB = U EC. 
AB BC. 
60° 
95 
55 
75 
45° 
87 
25 
55 
30° 
72 
10 
14 
20° 
54 
0 
0 
