1895. ] NEW YORK ACADEMY OF SCIENCES. 181 
50° De 60° 65° 
Left. Right. Left. Right. Left. Right. Left. Right. 
40.6 28.5 38.0 27.0 3 Ee pc | SPU Pala) 
40.0 29.0 Bioo) © wy OES 35:0 24.0 By TRA Las: 
40.0 29.8 ade te ded: 34.0 24.1 32.6 21.0 
40.5 29.1 37.5 = 26.3 SD.4) - 20-5 32.5 20.5 
40.0 28.8 38.0 26.5 Sa Bay Biles: 
34.5 24.0 
40.2 29.04 37.8 26.8 34.9 24.0 32°45 21-3 
— ee _— — = —LL 
2 w = 69.24° 2 w = 64.6° 2 w = 58.9° 2p —— on lis 
p = .304 fi Se pi—aaoled jp = 8 
70° “oe 80° 
Left. Right. Left. Right. Left. Right. 
30.0 20.0 27.0 16.5 24.1 13.5 
30.0 19.5 27.0 16.5 OF ela. D 
29:6 719:9 27.0 16.5 24.2 14.0 
30.0 20.4 PAG oye Allee al 23.9) 13a 
30.5). 19.5 7.0 16.4 24.0° 13.9 
26.5 15.5 
30.0 19.9 26.8 16.25 pF a laa 
_——— — —=— 
2 w = 49.9° 2 w = 43.05 ee 
p= .644 p—.731 (p= cred 
All of the previous observations had been subject to errors of 
unknown magnitude, aside from the errors of observation; and 
the results, while agreeing very closely with the calculated 
values for some angles, differed from them by considerable 
amounts at others. For example, the agreement for platinum 
at 80° was very close, while at 70° the difference was as large 
as .045. Similarly for uranium glass, the difference at 50° and 
65° was quite large. Hence I did not consider the results 
given by the experiments upon uranium glass and platinum 
altogether trustworthy as accurate quantitative measurements. 
The experiments upon silver, however, were free from all pos- 
sible error so far as I was able to discover, except the obdserva- 
tional error. Mention has already been made of the fact that 
Violle had previously made a number of determinations of 
the same general nature upon silver. His results do not agree 
very closely with those given above, being uniformly larger. 
I am altogether unable to account for the uniformity in the ex- 
cess of his values over those given by these experiments. His 
results are here inserted for the sake of comparison. 
