286 BELL SYSTEM TECHNICAL JOURNAL 



conveniently be made. Paraffin may surround ^* the target and the 

 source, to slow down the primary neutrons; or the water itself may 

 perform this office. Again, the target may be diffused throughout the 

 entire water-mass, in the form of a soluble salt of uranium. The 

 detector is a substance becoming radioactive when exposed to slow 

 neutrons. It may itself be spread throughout the water in the form 

 of a soluble salt, or it may be in the form of a thin foil which can be 

 moved from place to place in the water. In the former case, the water 

 is thoroughly stirred after the exposure is over, and then a sample is 

 taken, the activity of which is a measure of the average density — and 

 therefore of the total quantity — of thermal neutrons in the entire 

 tank during the exposure. In the latter case, the foil is used for 

 mapping out the density of thermal neutrons in the water as function 

 of the distance r from the target in the middle, and what is usually 

 plotted is the "/r^ curve," / standing for the strength of the activity 

 of the foil. 



The total quantity of thermal neutrons, existing at any moment 

 dispersed throughout the water, is greater in the presence of the 

 uranium than in the absence thereof ^^ (Anderson, Fermi and Han- 

 stein) ; this is the simplest proof of the fundamental result. When 

 the Ir^ curves are compared, it is found that the presence of the ura- 

 nium lowers the curve in the close neighborhood (within 13 or 14 cm) 

 of the neutron-source, but raises it further out. Presumably this is 

 because the uranium swallows up the slow primary neutrons, and those 

 which it gives out in exchange are themselves not slow until they 

 have gone a long way onward in the water. In tanks of sufficient 

 size, the increase farther out more than balances the diminution nearer 

 in, and the total quantity of thermal neutrons is augmented by the 

 presence of uranium (Halban, Joliot and Kowarski); this agrees with 

 the other result. It is therefore established what when the primaries 

 are slow, the fission-process delivers more neutrons than it consumes. 

 The same holds true when the primaries are fast, for when a beam of 

 RnBe neutrons is sent through a plate of uranium oxide the detector 

 beyond reveals a greater quantity of rapid neutrons than when the 

 plate is absent (Haenny and Rosenberg, experimenting with a plate 

 8 cm thick). 



How many neutrons then emerge, for every one which is spent in 

 producing a fission? This is a remarkably difficult question to put to 



18 It is not necessary that the " slowlng-down " substance be actually between the 

 target and the source, since slowed-down neutrons come out of it in all directions. 



19 To make the situations strictly comparable, the uranium is replaced in the 

 control experiment by some substance possessing an equal absorbing-power for 

 neutrons, but not liable to fission. 



