939 
§ 4. Before proceeding we will communicate the results of an 
investigation about the observations of the roar of cannons made 
by a number of observers of the Royal Dutch Meteorological Insti- 
tute, temporarily reinforced by other correspondents. These observa- 
tions have been continued until now by some observers, but have 
been rather numerous during the months October—February. A 
preliminary report about the results.for October Sh appeared in the 
periodical “Hemel en Dampkring” for October 1914, p. 81, and was partly 
reprinted by Prof. Mrinarpus in “Meteorologische Zeitschrift” for 
May 1915, p. 199. The latter paper gave us the opportunity of 
extending the observational area for October Sth east and southeast- 
ward, so that almost half a circle round the source of sound near 
Antwerp was formed. 
On account of the extraordinarily large number of separate sounds 
there was no possibility of attributing the observations of definite 
sounds to well-defined causes. The only thing possible was to classify 
the reports according to the degree of clearness with which the 
sound was heard, the duration of the observations and the degree 
of certainty with which a direction was indicated. In figures 9—16 
this is shown by indicating a report without particularities by —, 
a report of moderate intensity and number of the sounds by + and 
the reports of very heavy sounds, accompanied by vibrations of 
window-panes, ete. by =. A circle, which sometimes occurs, means 
that there is a definite statement that #o roar of cannons was 
observed. 
In all figures circles have been drawn indicating the distance 
from the source of sound. 
For the bombardment of Antwerp, especially heavy on October 
Zh Gh, the place from where the heaviest gunfire was directed 
was known exactly enough, a little north of Malines. For the other 
cases of more numerous reports we had first to find out from 
what part of the battlefield the sounds might have come. Very 
useful in this inquiry has been a complete number of the mail- 
„edition of the “Nieuwe Rotterdamsche Courant” and a collection of 
the copies of that paper for periods of 10 days from the date chosen, 
which we obtained through Mr. van Manen, one of the editors of 
the paper, from Miss Dr. Cu. van Manen. We readily grasp this 
opportunity to repeat here our cordial thanks for this co-operation. 
In the second place we used the monthly war-reviews, published 
by the “Times”. 
By combining the information from different sources we succeeded 
in every case in obtaining a high degree of certainty about the 
to} o o 
