370 MESSRS. R. H. FOWLER, E. G. GALLOP, C. N. H. LOCK AND H. W. RICHMOND : 
small a calibre as is consistent with obtaining accurate measurements of the jump 
cards. 
With regard to the effect of variation of shape we have very little evidence. 
If we compare the moment coefficients f M for shells of 2 and 6 calibres radius of head, 
as shown in figs. 4 and 5, it is obvious that the difference is much less marked than 
the difference between the two curves of f u , and that the two curves of f M are very 
nearly of the same shape. No great errors would be introduced by assuming that the 
values of f M for the two shells were in a constant ratio. Thus it seems reasonable, 
until the appearance of evidence to the contrary, to consider that the value of the 
moment coefficient for any shell can be obtained by multiplying the value, obtained 
in this experiment, by a constant independent of the velocity. It will then be 
sufficient to determine the value of this constant at a single velocity, which may even 
be a low velocity attainable in a wind channel. The value of the cross-wind force 
factor for any shell may be obtained in a similar manner but the results will be more 
uncertain. For rough purposes it may even be sufficient to assume f h and f m 
independent of the velocity except when dealing with velocities very near the velocity 
of sound. It thus appears to be possible to treat f M in the classical way in which 
was treated, in which it was assumed that the values of f n for two different shells are 
in a constant ratio at all velocities. This treatment is inadequate in the case of f K , 
but on present evidence is far more valid in the case off^. 
By applying the results of the present trial in this way, we may even hope to get 
reasonably accurate estimates of the drift and stability for any type of shell, on the 
basis of wind channel experiments only on the particular shape of shell required. 
The method would be especially valuable in connection with the design of new 
shapes of shell. It is known that, in general, the longer and more pointed a shell is, 
the less is its drag coefficient; by a series of wind channel tests on a series of shell 
shapes it would be possible to determine the greatest length of shell that would be 
sufficiently stable in a gun of given rifling, or the sharpness of rifling required to make 
a given shell stable. Useful information was obtained on this point from wind 
channel experiments before the jump card trial provided certain data for the extra¬ 
polation to high velocities. It must be emphasised, however, that this one experiment 
needs extension and confirmation before the structure sketched above can confidently 
be reared upon it. 
We have now discussed in general terms the applicability of our theory and 
experiments to the calculation of drift, stability, the effect of wind, and the design of 
improved forms of shell. Though the details of the calculations on these various 
points are not given here, enough has been said to show that the results form some 
advance in the subject of the application of aerodynamics to the flight of shells. 
