230 
MESSRS.. R. H. FOWLER AND C. N. H. LOCK ON 
they show remarkably good agreement between the new results and the old. This 
confirms the substantial truth of the whole theory ; in particular, the agreement of 
the results for the two twists of rifling verifies that the couple M is unaffected by a 
change in axial spin in the ratio 3 to 4. 
Fig. 1. Values of tlie couple coefficient/ M (v/a, 0) and the normal force coefficient / N (v/a, 0) plotted 
against vja, corrected for the effect of the cards. The full curves for/ M for types I.-IV. are reproduced 
without alteration from (A). The plotted points show the means of observations here analysed for the 
first and second half-periods, and the numbers against them show the number of rounds contributing to 
each mean. The group marked * was analysed in (A) by a different method with identical results. The 
origin of/j, for type IV. is displaced downwards 4 units. 
The full curve (V) for f s (shells of form A) represents the complete results of the experiment. The 
dotted curve reproduces the partial results of (A). 
Curve I.—for 3-inch shells of form A with centre of gravity 4-20 inches from the base. (Type III.) 
Curve II.—The same with centre of gravity 4-73 inches from the base. (Type I.) 
Curve III.—The same with centre of gravity 5-08 inches from the base. (Type II.) 
Curve IV.—The same for form B with centre of gravity 4-965 inches from the base. (Type IV.) 
Curves of / M sin S, the complete moment coefficient, considered as a function of 
botli variables vja and S corrected for the effect of the cards, are plotted against 
S in fig. 2. The information is somewhat fragmentary : in addition to wind channel 
results, values of / M , when d>10 degrees, are available for shells of type I.—III. for 
two velocities near v — a. For type IV. (pointed shells) values of / M for large $ are 
wanting in this region, but exist for two high velocities, and one less than a. 
