DISCOVERY OF THE NEUTRINO — COWAN 427 



day the films would be removed and developed for reading. At that 

 time tests would be made of the detector and electronic system to catch 

 any changes that might have occurred. 



THE FIVE ELEMENTS OF PROOF 



Having the equipment operate as planned near the reactor and 

 observing the correct patterns of pulses now and then was most satis- 

 fying. But now the work remained to test these signals to ascertain 

 whether or not they were in fact produced by antineutrinos from the 

 reactor. Five experiments were performed using these pulses, with 

 objectives as listed below : 



1. The rate at which they were recorded must be correct, knowing 

 the reactor power and detector efficiency. This rate must drop to 

 zero (or to a relatively low and well- understood background) when 

 the reactor is shut down. 



2. The first pair of pulses must be shown to be due to the annihila- 

 tion of a positron by an electron. 



3. The second pair of pulses must be shown to be due to the capture 

 of a neutron by cadmium, and the neutron must have appeared in 

 the detector at the same instant as did the positron. 



4. The signal rate must be proportional to the nmnber of protons 

 in the water target tanks. If the amount of hydrogen is changed, 

 the signal rate must change accordingly. 



5. The signal, when shown to be associated with the reactor being 

 run, must be shown to be independent of gamma rays and neutrons 

 leaking from the reactor shield. 



The following months saw these tests undertaken. In each test, 

 the two water tanks operated as independent targets, and the data 

 obtained from each were analyzed and required to check one another. 

 The checks were made in various, sometimes redundant, ways, in 

 order to apply every test we could devise. The details of these checks 

 and the resulting data are reported in the relevant papers listed in 

 the bibliography, and will be described only in general terms here. 



Dependence of the signal rate on reactor power. — Tliis is the 

 easiest to describe. The equipment was operated for 893.5 hours 

 (in two separate runs) with the reactor on, and for 263.4 hours (again, 

 in two separate runs) with the reactor off. With the reactor on, the 

 signal rate was about 1.8 per hour, and with the reactor off, it was 

 about one-fifth of this. This background rate was understood in 

 terms of cosmic ray interferences, similar to the ones which had forced 

 us to stop work at Hanford. But there, the cosmic ray backgi'ounds 

 were some 10 times higher than the signal rate produced by the reac- 

 tor. We could also work our data "in reverse," calculating a cross 

 section for the reaction from them, then comparing it with the 

 theoretical one. The two — experimental and theoretical — agreed 



