178 BELL SYSTEM TECHNICAL JOURNAL 



particular experiment which I will quote, the apparatus of Fig. 12 

 was set up under a skylight on the top floor of a building not far above 

 sea-level (the Reichsanstalt in Charlottenburg) without the upper part 

 of the metal armor, so that rays coming vertically from above should 

 reach the upper counter after passing through the atmosphere and a 

 little glass. Between the upper and the lower counter, a block of gold 

 4.1 cm. thick was alternately inserted and removed. Over a period of 

 twelve half -hours with the gold-block absent there were 19897 dis- 

 charges in the upper counter, 12209 in the lower, 986 coincidences; over 

 a period of twelve intercalated half-hours with the gold-block pres- 

 ent there were 19814 discharges in the upper counter, 10562 in the 

 lower, 743 coincidences. The presence of the gold had thus reduced 

 the coincidences by 24.6 per cent; this the authors qualify by a 

 considerable statistical uncertainty, stating it as (24.6 ± 4.2) per 

 cent. 



Rossi did the like experiment, with lead and roofing over the upper 

 counter, equivalent in thickness to 77 cm. of water; in place of the 

 gold he used a slab of lead 9.7 cm. thick; it reduced the number of 

 coincidences by sixteen per cent, which he expresses as (16 ± 3)%. 

 Only a week or so before I add these words, he reported yet another. ^^ 

 There were three counters in a vertical line, 7 cm. of lead and 6 cm. 

 of roofing above them, 25 cm. of lead permanently between them; 

 and an additional thickness of no less than seventy-six cm. of lead, 

 which he could interpose between the counting-tubes (half of it be- 

 tween the upper two, half between the two lower). This enormous 

 absorbing layer cut down the number of (triple) coincidences by only 

 some forty per cent (Rossi writes 38.5 ± 5.1). 



Now, the coincidences are ascribed to ionizing particles such as those 

 of which the tracks are made visible by the cloud-chamber method: 

 particles which make densely-crowded trains of ion-pairs along their 

 paths. Let us forget for the moment about high-frequency photons, 

 and think only of such particles, descending almost vertically from on 

 high. The numbers 24 and 16 and 38, quoted above, are the percent- 

 ages of these which are stopped by plates of gold and lead, of the 

 stated thicknesses, in the situations stated. Now the remarkable and 

 perplexing point about these numbers is, that they are about the same 

 as the percentage declines of ionization which are found, by Millikan 

 and Regener, on lowering an electroscope through the equivalent thick- 

 nesses of water! ^^ 



'" I have just received (February 1) a preprint of the report from Dr. Rossi; it is 

 marked NatHnvissenschaften 20, 65 (1932). 



^^ Bothe and Koihcirster, and Rossi, elect the following way of making the com- 



