October 14, 1920] 



NATURE 



215 



account for it, such, for example, as that the shell 

 "yawed" from its path to a degree varying from 

 gun to gun, the " yaw " being supposed to affect 

 the pressure at the escape holes, and therewith 

 the time of burning. The true explanation, how- 

 ever, proved to be the hitherto unsuspected effect 

 of "spin" — i.e. the angular velocity of the shell 

 about its axis, and this factor has since proved 

 practically the most important one in the behaviour 

 of a powder train fuze. The shell, in order to 

 secure stability in its flight, is given a high 

 angular velocity by the rifling of the gun working 

 on the copper driving-band. In all the guns em- 

 ployed (varying in velocity from 900 f.s. to 

 2500 f.s.) the twist of the rifling was i turn in 

 30 calibres — i.e. in 30 times 3 in. or 7J ft. This 

 gave angular velocities varying from 7200 to 

 20,000 revolutions per minute in the five guns. 

 The angular velocity of a shell falls off compara- 

 tively slowly in flight, so that it could be regarded 

 as approximately constant along the trajectory of 

 each gun. The peculiar differences observed in 

 the gun trials could be explained only as an effect 

 of spin, and it was clearly necessary to carry out 

 spinning trials on fuzes "at rest" — i.e. without 

 forward velocity — to see if the effect of spin could 

 be isolated. Such trials were carried out at speeds 

 up to 30,000 r.p. m., and an enormous effect of 

 spin was established. It was possible to double 

 the time of burning of a fuze, or even to make 

 it cease burning altogether, merely by spinning it. 

 The effects of a fall of pressure also were exag- 

 gerated by spin, as was shown in the laboratory 

 at University College, by spinning a fuze under 

 reduced pressure. 



The explanation of this effect of spin is inter- 

 esting. It could not be due to any "dynamic pres- 

 sure " effect at the escape holes, or to a centri- 

 fugal effect on the gases in the groove ; these 

 were investigated and found to be far too small. 

 The real explanation is the centrifugal effect on 

 the slag produced by the gunpowder in its com- 

 bustion. When the spin is high the gunpowder, 

 warmed, softened, and just ignited by the com- 

 bustion of the previous layer, is "spun" outwards 

 to the outer edge of the groove before it has had 

 time properly to burn and to ignite the next layer; 

 consequently, combustion is slower, and may fall 

 altogether. The absence of any effect of spin in 

 the case of a special powder giving no slag, as 

 well as the fact that "blind" fuzes arc found to 

 have failed first on the inside edge of the ring, 

 make it clear that the centrifugal effect on the 

 slag is the prime cause of the trouble. At 



30,000 r.p.ni., a spin reached in fuzes fired from 

 small guns, it is almost impossible to attain any 

 accuracy at all. The rapid increase of fuze-trouble 

 with spin is due to the fact that the centrifugal 

 effect varies as the square of the spin. 



One obvious means of avoiding the excessive 

 effect of spin was to reduce the rifling of the gun 

 and therewith the rotation of the shell. The possi- 

 bility of doing this is strictly limited, as with too 

 low a spin the shell becomes unstable. Two 

 similar guns were rifled respectively i turn in 

 30 and I turn in 40 calibres, and in all respects 

 the fuzes fired from the latter were found to be- 

 have more satisfactorily, thus confirming the 

 results and predictions of laboratory trials. All 

 similar guns were provided thereafter with the 

 smaller rifling, with good effects. 



Another factor affecting the behaviour of fuzes 

 is their temperature. This effect, also previously 

 unknown, is a smaller one, but by no means 

 negligible. .\ fuze burns more quickly at a higher 

 temperature, and allowance must be made for this 

 in accurate firing. A curious phenomenon arises in 

 connection with this. It was usual to test a fuze 

 at rest as well as in the gun, and in the case of 

 some long-burning fuzes at rest the fuze heats 

 itself by its own combustion to such an extent 

 that its time of burning is seriously decreased. 

 This " self-heating " effect does not occur in a 

 fuze fired from a gun, which is cooled by its pas- 

 sage through the air. Consequently, for accurate 

 comparison with gun trials the fuze fired at rest 

 must be cooled while it burns — e.g. by subjecting 

 it to a rapid spray of water. This was actually 

 done in later trials, the fuze being rotated in a 

 closed box at any required spin and pressure, and 

 subjected the while to a rapid jet of water to 

 ensure the constancy of its temperature. 



We may summarise as follows : The rate of 

 burning of a fuze is a function of the total pres- 

 sure at its escape holes, which is made up of the 

 atmospheric pressure .\ and some function f(v/V) 

 of V its velocity and V the velocity of sound. It 

 is a function also of the spin S and of the tem- 

 perature T. Expressed mathematically, the rate 

 of burning is equal to F[{A-f /)/(v/V)), S, T], 

 where F is some complicated function of the three 

 variables. It is easy to see that fuzes are likely 

 to cause trouble when subjected to conditions, as 

 they were in the late war, far exceeding in severity 

 any under which they had previously been used ; 

 and to foretell that in the next war — if there be 

 one — reliance will be placed mainly on clockwork 

 fuzes unaffected by these various factors. 



The Iridescent Colours of Insects.* 



By H. Onslow. 



III. — Selective Metallic Reflection. 

 T N the two preceding articles various insects 

 *• have been described and illustrated, which 

 owe their principal iridescent effects to the colours 

 of " thin plates " and to the diffraction of ribbed 



1 Confiaii«<) frn*ii p. 1S3. 



NO. 2659, VOL. 106] 



structures or "gratings." However, more than 

 one physicist of repute has stated that most insect 

 colours are due to selective metallic reflection. 

 The arguments against this theory, as applied to 

 scales, were considered in the first article; briefly, 

 they are due to the facts that both reflected 



