568 JOHN F. HAYFORD 
in the table are the computed deflections on the different assump- 
tions indicated. They correspond to the totals at the foot of the 
preceding table. 
Longitude Azimuth Latitude Latitude Azimuth 
Sta. No. 1, | Sta. No. 59, | Sta. No. 54, | Sta. No. 164 | Sta. No. 115, 
Point Arena.| Mt. Ouray, | Uncompah- Calais, Knott Island, 
Calt Colo. gre, Colo. Me. Va. 
Hayford 
M= 1.0, hy=113.7 +15767 —10703 +5755 —o'51 = 1.00 
Lewis 
M=0 2-1 —TOn20) aerate eee ==TOaO2a ls savenereal! “sack esters tll fen eeeeennee 
M=o0.5 h=19.29...| +53.73 — 8.22 —3.093 —16.00 — 27.19 
M=o0.9 h=19.29 ...| +13.06 — 5.86 +0.58 — 2.97 — 5.52 
Mir Ate TO\20) sols eee an Louie ee VA cone eras eh |ilera cela oie 
Mi Tc Al er TOMO Merl ne se cveree el ai cere EOL cate eee allie neem 
FE). AOE 2X0) “5 Bell — ao aude. > eb ao os ate Oe OF ill = ya oe eee 
If Mr. Lewis had constructed this table, which would have 
required less than a day of computation, he would have avoided 
the gross error into which he has fallen. Instead of believing, 
as indicated in his article, that a large reduction in assumed com- 
pleteness of compensation of about the same amount for every 
station would neutralize the large change of assumed depth from 
113.7 to 19.29 km. he would have known that for some stations 
such a neutralization requires a very large reduction of assumed 
completeness (to M=.2 at Mt. Ouray, for example), for others 
requires a very small reduction (at Point Arena, Calais, and Knott 
Island, for example), and for others requires an increase of assumed 
completeness, that is, over-compensation (at Uncompahgre, for 
example). He would have noticed that for these five stations 
the average value of M necessary to secure an agreement with 
the Hayford values is about .9, not far from unity. 
If beginning to be skeptical of his own proposition that a 
reduction in completeness of compensation at all stations would 
counteract a reduction in assumed depth of compensation, and 
beginning to suspect that such a counteraction is impossible in 
the computations as actually made, he had then proceeded to 
examine into the matter more carefully he might have noted also 
the following things: 
1. That the factors as printed on p. 617 differ very largely 
for separate rings from the Hayford factors. The maximum 
