438 PIRB-CONSEEVANCY. 



motion directed towards the trace, thereby not only smo"thering up 



the fire, but also brushing away on to the trace burning fragments 



that must otherwise only cause it to flare up anew before the broom 



could be brought to bear on them again. If the broom were brought 



down vertically, these burning fragments would be scattered about 



in every direction and some might fall amongst and set on fire the 



standing grass or accumulation of dead leaves just outside the 



trace. The force -with which the broom should be brought down 



will of course depend on the violence of the advancing fire, and 



the more violent it is, the closer together must the broom-men be 



ranged along the edge of the trace. Indeed, when the conflagratioa 



is very violent at any point, several men must beat it out, bringing 



down their brooms rapidly -one after another so as to give the fire 



no chance of reviving. The fire in the cushion of dry grass may 



be so fierce that the men, while waiting to attack it, may be 



obhged to stand in -the midst of the grass and bushes beyond th« 



guide line. To protect themselves -then, and also when they 



advance and attack the fire, they should screen themselves behind 



a bunch of leaves or twigs held in the left Jiand. Without such 



protection the men would -be exposed to extreme discomfort, and 



would often be unable to approach the fire near enough to attack 



■it. Indeed they should never be without such fire-screens. In 



■beating out a violent fire care must be taken that when the broom 



is lifted back to strike again, fragments of burning grass or leaves 



are not thrown backwards from it into the protected area, a by no 



means unusual occurrence. 



If, while a trace is being fired, -a dangerous wind should spring 

 up, the work should be stopped at once. It is false economy to 

 go on with the operation on the plea that the men being all out on 

 the spot, their services jnust on that account be utilised. Should 

 the fire spread into the forest, as it is bound to do in nine cases 

 out of ten, the diminished area actually protected would mean 

 an appreciable increase in the cost of protection per acre. If 

 the grass in the forest is very dry, the £rst occurrence of any 

 symptom of a whirlwind should be the signal for stopping work 

 at once. 



In fire-tracing the effect of a large mass of fire to produce a 

 strong draught of colder air towards itself should never be forgotten 

 and should be made the most of. A few remarks on this subject 

 will not be out of place. The air being still, if a single line of fire 

 is burning, the flames will all ascend vertically, the draught from 

 opposite sides behig equal ; but if a second line of fire parallel to, 



