312 Ml' Fowler and Mr Lock, The origin of 



these causes, and it is the object of this paper to show by analysis 

 of the available experimental data that it is barrel vibrations that 

 mt *^® dominant cause, at least in these particular experiments. 

 Whether the barrel vibrations are the cause in general it is for 

 future experiments to decide. 



(2) Consequences of the random ^pressure variation theory. If 

 random pressure variations and random blast disturbances are the 

 dominant cause of the initial angular velocity of the axis of the 

 shell, certain well-marked characteristics of the initial motion of 

 the axis can be deduced. For shortness it is convenient to define 

 the position of the axis OA relative to the direction of motion of 

 the centre of gravity OP by the angles 8, cj) of Fig. 1 

 V 



Fig. 1. 



The plane POV is vertical, and the angle S is usuallv called the 

 yaw. Then the initial circumstances of the axis of the shell are 

 defined with sufficient precision by the initial values of </> and 

 dhjdt^ or 8', that is to say, by <f>^ and 8,,'. The initial value of 8 itself, 

 obtained by extrapolation, is in practice very little different from 

 zero, as is to be expected, and may be ignored in what follows. 



Now random variations of gas pressure across the base of the 

 shell and random asymmetry of the blast will result in a disturbing 

 couple (practically impulsive) acting on the shell, whose plane of 

 action must be expected to vary from round to round in an entirely 

 arbitrary manner. That is to say, this theory demands that all 

 values of (j>^ should be equally probable. Again, the impulsive 

 disturbing couple which acts on the shell can hardly depend in 

 any way on the axial spin of the shell or the twist of the rifling. 

 The mean value of the disturbing impulsive couple, and therefore 

 of So , should thus be the same for similar shells fired at similar 

 velocities from similar guns with different twists of rifling. These 

 are two deductions from the theorv which can be tested by the 

 expenmental results. We may say at once that neither deduction 

 IS sufficiently fulfilled. We may be certain that, though no doubt 

 such random pressure efl^ects occur and are appreciable, the main 

 cause of the disturbance Kes elsewhere. 



