DE. GLADSTONE ON THE EEFEACTION-EQTTIVALENTS OE THE ELEMENTS. 19 
Table VII. — Supplementary. 
Substance. 
Alcoholate. 
Hydrate. 
Silicate. 
Hypopbospliite. 
Nitrite. 
Univalent. 
Potassium 
Sodium 
A+a+9-27 
B + a + 9'21 
A+a-6-22 
B+a-6-20 
A+«+ 12-20 
B+b+11-88 
A+a+8'45 
A+«+0-48 
B+«+0T0 
G+*-6-9l 
II+«- 8-294 
Hydrogen 
Bivalent. 
H+a+6 637 
L+2(<*+7-30) 
T+20+ 7-26) 
Substance. 
Tartrate. 
Chromate. 
Bichromate. 
Oxalate. 
Carbonate. 
Univalent. 
Potassium 
Sodium 
Lithium 
Ammonium 
Hydrogen 
Bivalent. 
2(A+*)+20-21 
2(B+«)+20-05 
2(H+«) + 16-85 
2(A+«)+13-84 
2(B+«)+15-28 
2(C+«)+13-90 
2(G +<*)+ 13-46 
2(A+*)+44-89 
2(B+*)+44-07 
2(0 +«)+42-88 
2(G+«)+42-40 
2(A+«)+005 
2(G-+«) — 2-44 
2(H+<4— 5-00 
2(A+«) — 8-89 
2(B+«.) — 8-45 
Table VII. — Supplementary. 
Substance. 
Eluoride. 
Arsenite. 
Sulphocyanide. 
Sulphite. 
Potassium 
A+« — 9-28 
A+a+14-64 
2(A+«) — 2-56 
Sodium 
B+a+10-70 
Substance. 
Biborate. 
Permanganate. 
Eerricyanide. 
Eerrocyanide. 
Potassium 
2(A+*)+54T 
3(A+a)+46-56 
4(A+«)+39-40 
Sodium 
2(B+«)+15T 
The differential numbers along a line in Table VI., or down a column in Table VII., 
are sufficiently near to show that we are dealing with a reality ; but they are sufficiently 
wide apart to show that we must rely upon the average of the numbers and not on any 
single experiment, if we wish to get a refraction-equivalent true to the first place of 
decimals. Unfortunately all experimental errors fall upon this residuary number. 
The only exception to this regularity which is worth notice, is in the case of hydrogen, 
which is A — 4'5 or thereabouts in the hydracids, but drops to somewhere about A — 6*7 
in the organic acids. This seems to indicate that in the first group hydrogen has a 
refraction-equivalent somewhere about 2-2 higher than in the other. 
Though these Tables alone do not afford us the means of determining a single refrac- 
tion-equivalent of a metal or of any other element, it is evident that the refraction-equi- 
valents of the whole would be a simple matter of calculation if we could determine with 
certainty the value of any letter, Roman or Greek, that is, the refraction-equivalent 
of any one of the constituents. The means of arriving at this will be explained in the 
second part of this paper under the head of Potassium. 
