Report of Committee 
OIN 
Atomic Weights, Published 
During 1893. 
F. W. Clarke. [Reprinted from the Journal of the American Chemical Society. 
Vol. XVI, No. 3. March, 1894.] 
REPORT OF COnHITTEE ON DETERHINATIONS OF 
ATO1TIC WEIGHT, PUBLISHED DURING 1893.' 
By F. W. Clarke. 
To the Members of the American Chemical Society: 
YOUR committee, appointed at the last annual meeting, 
respectfully submits the following report, showing the 
advances made in our knowledge of atomic weights during 1893. 
The year has been one of reasonable activity in this branch of 
investigation, and it is thought that all the work actually pub- 
lished within its limits is here summarized and recorded. One 
partial exception possibly may be made to this statement. Mor- 
ley's work on oxygen, reported orally at the American Associa- 
tion meeting, and at the Chemical Congress in Chicago, is 
omitted, for only unofficial abstracts of it have yet appeared in 
print. That work is of such fundamental importance that it 
seems best to await its completion, rather than to do it possible 
injustice by fragmentary notices which might be inexact. 
In the "Stas Memorial Lecture," by Professor Mallet,2 will 
be found a remarkably full discussion of the theory or philosophy 
of atomic weight determinations, which should be carefully read 
by all students in this domain. Taking the life and work of 
Stas as text and example, Mallet considers the conditions neces- 
sary to the fruitful continuation of that work, and throws out 
many suggestions of great practical value. He particularly ad- 
vocates the multiplication of interdependent data, in order that 
errors may be eliminated; and he also recommends the estab- 
lishment of a special, endowed laboratory, in which a group of 
trained observers may co-operate towards the attainment of the 
best results. 
The individual atomic weight determinations of 1893 are sub- 
joined. 
Boron.-W. Ramsay and Emily Aston3 redetermine the 
atomic weight of boron by two distinct processes, one of them 
being entirely new. First, with many precautions, they estimate 
1 Read at the Baltimore meeting Dec. 28, 1893. 
2 J. Chem. Soc., August, 1893. 
3 J. Chem. Soc... 63, 207, February, 1893. 2 
F. W. CEARKE. DETERMINATIONS 
the percentage of water in crystallized borax, Na2B4O7, 10 H20; 
all weights being reduced to vacuum standards. The same 
reduction is also made in their other scries, but as they carry out 
the weights to seven decimal places, the corrections applied go 
farther than is necessary. In the subjoined tables I give the 
nearest tenth milligramme. 
Wt. Na2B4O„ ioH2O. 
IO.3582 
Wt. Na2B4O7. 
5-4784 
Per cent. H2O. 
47-1099 
At. Wt. B. 
11.04 
5-3440 
2.8247 
47-U33 
10.97 
4-9963 
2-6379 
47.2026 
IO.85 
5-7ooo 
3.OIOI 
47-I9I2 
IO.87 
5-3143 
2.8066 
47.1882 
10.88 
4.9972 
2.6392 
47-1865 
10.885 
5-2367 
2-7675 
47-I524 
10-955 
Mean atomic weight, 6=10.921, ± 0.010. 
In the second series of experiments fused borax was distilled 
with hydrochloric acid and methyl alcohol, and the residual 
sodium chloride was weighed. Results as follows: 
Wt. Na2B4OT. 
Wt. NaCl. 
At. Wt. BA 
4.7684 
2-7598 
II.OI5 
5-2740 
3-O578 
IO.925 
3-2344 
I.8727 
IO.992 
4.0862 
2.3713 
IO.879 
3-497° 
2.0266 
IO.949 
Mean atomic weight, 6=10.952, 
± O.OIO. 
The distillations were conducted in soft glass flasks of about 
100 cc. capacity, in which the sodium chloride was dried at 
about 3500, and finally weighed. It was found, however, that 
the flasks were somewhat attacked during the process, with lib- 
eration of silica, consequently some oxygen in the alkali of the 
glass had been replaced by chlorine, and the weight of the ves- 
sel increased. Another set of distillations was, therefore, made 
with flasks of hard combustion tubing, and these were slightly 
attacked also. 
Wt. Na2B4O7. 
Wt. NaCl. 
At. Wt. B. 
5-3118 
3-0761 
IO.983 
4.7806 
2.7700 
10-955 
4-9907 
2.8930 
10.936 
4-723I 
2-7360 
IO.968 
3-3I38 
1.9187 
IO.992 
1 Atomic weights used in Ramsay and Aston's calculations: O=i6, H=i.oo8, 
Ag=iO7.92, Na=23.05, 01=35.45, and Br=79.95. 
Mean atomic weight, 11=10.966, ± 0.005. OF ATOMIC WEIGHT. 
3 
As a check upon these results the sodium chloride was dis- 
solved in water, and after filtration precipitated with silver 
nitrate. The resulting silver chloride was collected on a Gooch 
filter, dried at 2000, and weighed. 
Wt. AgCl. 
Percent. Cl in NaCl. 
At. Wt. B. from Cl. 
7-5259 
60.493 
II.071 
6-7794 
60.515 
II.024 
7.0801 
60.516 
II.003 
6.6960 
60.514 
11.039 
4.693I 
60.479 
II.091 
Mean atomic weight, 6=11.052, ± 0.010. 
From all these results, and after discussing the supposable 
sources of error in them, the authors conclude that the atomic 
weight of boron is very nearly eleven. 
Still another determination of the atomic weight of boron, by 
an entirely independent method, is due to Rimbach.1 Taking 
advantage of the fact that methyl orange, as an indicator, is un- 
affected by free boric acid, he titrated solutions of borax with 
standardized hydrochloric acid, thus measuring the amount of 
sodium in the salt and thence calculating the value to be ascer- 
tained. The borax was prepared in platinum vessels from care- 
fully purified boric acid and sodium carbonate; and the hydro- 
chloric acid used contained 1.84983 per cent, of absolute HC1, 
determined gravimetrically as silver chloride. In the following 
table of results the titrated hydrochloric acid is given by weight 
in grammes: 
Wt. borax. 
Wt. HC1. sol. 
Per cent. Na.2 
At. Wt. B.2 
IO.OO2I4 
103.1951 
12.07081 
IO.9646 
I5-32772 
158.1503 
12.07138 
10.9598 
15.08870 
I55-727I 
12.07530 
IO.9273 
IO.1293O 
104.5448 
12.07517 
IO.9298 
5-25732 
54-257I 
12.07435 
IO.9361 
I5-O4324 
I55-23O7 
12.07283 
IO.9486 
15.04761 
I55-2959 
12.07448 
IO.9356 
IO.434O9 
107.6602 
12.07176 
IO.957I 
5-O47I3 
52-0897 
12.07480 
10.9330 
Calculating from the sum of all the weights the atomic weight of boron 
becomes 10.945, ± 0.003. 
1 Ber. d chem. Ges., 26, 164, February, 13, 1893. 
2 Calculations made with O=i6, 11=1.0032 (Keiser), Na=23.O575, 0=35.4529, and 
Ag=io7.9376. 4 
F. W. CLARKE. DETERMINATIONS 
Carbon.-Leduc' has determined the density of carbon mon- 
oxide, deducing thence the atomic weight of carbon. The 
globe used held 2.9440 of air. Filled with carbon monoxide the 
latter weighed 
2.8470 
2.8468 
2.8469 
Mean, 2.8469 
Hence the density of CO is 0.96702, when air =1, and with 
0=15.88, 
0=11.913, 
If O=i6, then 0=12.003. 
In a very brief note2 Wanklyn announces that he has studied 
a series of hydrocarbons in which the successive steps rise-not 
by CH.,= 14-but by | (CHJ = 7. If these observations are 
sustained, then the atomic weight of carbon becomes-6. 
Cobalt and Nickel.-Atomic weights determined by Winkler,3 
who starts with weighed quantities of metal, electrolytically de- 
posited. This is then dissolved in a platinum dish in pure hydro- 
chloric acid, and the chlorides produced are finally dried for sev- 
eral hours at 1500. In one set of experiments the chlorine in 
the chlorides is weighed gravimetrically, as silver chloride; in 
another set it is titrated with a standard solution of silver. In 
the first case we have the ratio R:2AgCl, in the other, R:Ag.2. 
Results as follows, for nickel: 
Wt. Ni. 
Wt. AgCl. (grav.) 
At. Wt. Ni. 
0.3011 
I.462I 
58.9102 
O.2242 
I.0881 
58-94l8 
O.5166 
2.5108 
58.8571 
O.4879 
2.3679 
58.9419 
O.3827 
I.8577 
58.9304 
0.3603 
I-75I7 
58.8334 
Mean result, Ni=58.9O33. 
Wt. Ni. 
Wt. Ag. (vol.) 
At. Wt. Ni. 
O.l8l2 
0.6621260 
58.9253 
0.1662 
0.6079206 
58.8665 
0.2129 
O.7775252 
58.9584 
O.2232 
0.8162108 
58.8811 
0.5082 
1.8556645 
58-9684 
0.1453 
0.5315040 
58.8631 
Mean result, Ni=58.9ic>4. 
1 Compt. rend., 115, 1072. 
2 Phil. Mag., (5) 36, 552, December, 289?. 
8 Ztschr. anorg. Chem., 4, 10, 2897. OF . ATOMIC WEIGHT. 
5 
The following data were obtained for cobalt: 
Wt. Co. 
Wt. AgCl. (grav.) 
At. Wt. Co. 
0.3458 
I.6596 
59.6O44 
0.3776 
1.8105 
59.66O9 
0.4493 
2.1521 
59-7215 
0.4488 
2.1520 
59-6577 
0.2856 
1.3683 
59-7O8I 
0.2648 
I.2678 
59-748O 
Wt. Co. 
Wt. Ag. (vol.) 
t. Wt. Co. 
0.177804 
O.6418284 
59.6495 
O.263538 
0.9514642 
59.6396 
0.245124 
0.8855780 
59-5996 
0.I90476 
O.6866321 
59-73H 
0.266706 
O.9629146 
59-6388 
O.263538 
0.9503558 
59-7O92 
Mean result, 00=59.6834. 
Mean result, 00=59.6613? 
In the case of the cobalt determinations it is of course con- 
ceivable that the chloride formed might be contaminated with 
traces of basic compounds. This question is considered by 
Winkler in a supplementary communication.2 The weighed, 
electrolytic cobalt, deposited upon the surface of a platinum 
dish, was treated with a solution of neutral silver sulphate. Sil- 
ver was thrown down, and was washed, dried, ignited, and 
weighed. Thus a direct ratio was measured between silver and 
cobalt, involving no intervention of cobalt chloride, and conse- 
quently no error due to basic salts. Results as follows: 
Wt. Co. 
Wt. As. 
At. Wt. Co. 
O.2549 
O.9187 
59-7421 
0.4069 
I.4691 
59.6377 
The precipitated silver, as a check upon its purity, was dis- 
solved in nitric acid, reprecipitated with hydrochloric acid as 
chloride, and filtered off. The filtrate, upon evaporation to dry- 
ness, gave traces of residue, containing cobalt. The amount of 
the latter in the first experiment was at most 0.5 milligramme, 
and in the second not over 0.2 milligramme. Correcting for 
these amounts the values found for the atomic weight of cobalt 
become 59.6356 and 59.6164, respectively. These figures fall 
within the limits of variation of those found in the first paper, 
1 All calculations with Ag=io7.66 and 01=35.37 
2 Ztschr. anorg. Chem., 4, 462, 1893. 6 
F. W. CEARKE. DETERMINATIONS 
and show that the supposed error, if it existed, could not be large. 
The author admits, however, that the degree of concordance 
among his experiments is not so great as could be desired. The 
main purpose of his work was to show the essential constancy of 
the values, as opposed to the views of Kriiss concerning the sup- 
posed composite nature of nickel and cobalt. 
Molybdenum.-Atomic weight redetermined by Smith and 
Maas.1 Sodium molybdate was converted into chloride by heat- 
ing in a current of pure, dry, gaseous hydrochloric acid. Results 
as follows with weights reduced to vacuum standards: 
Wt. Na2MoO4. 
Wt. NaCl. 
At. Wt. Mo. 
I.14726 
0.65087 
96.130 
0.89920 
0.51023 
96.094 
0.70534 
0.40020 
96.108 
0.70793 
0.40182 
96-031 
I.26347 
O.71695 
96.087 
I.I52I7 
0.65367 
96.126 
0.90199 
0.51188 
96.067 
0.81692 
O.46358 
96.077 
0.65098 
0.36942 
96.O73 
0.80563 
O.45717 
96.078 
Palladium.-Atomic weight determined by Joly and Reidie' 
by means of potassium palladio-chloride, K,PdCl4. In the first 
series of experiments the salt was dried in vacuo at ordinary 
temperature. It was then electrolyzed in a solution acidulated 
with hydrochloric acid, both the palladium and the potassium 
chloride being weighed. The palladium was dried, ignited in 
a stream of hydrogen, and cooled in an atmosphere of carbon 
dioxide. Results as follows: 
In mean, ^10=96.087. 
At. 
Wt. Pd. 
Wt. salt. 
Wt. Pd. 
Wt. KC1. 
From salt : Pd. Pd : 2KCI. 
1.0255 
O.3919 
0.5520 
IO5.749 
IO5.643 
1.2178 
0-3937 
0.5551 
104.881 
105.535 
1-2518 
0.4048 
O.5687 
104.923 
IO5.916 
This series was rejected by the authors upon finding that the 
salt retained water. In one experiment 0.23 per cent, of water 
was found. Accordingly a second set of determinations was 
1 This Journal, 15, 397, July, 1893. Results calculated with O=r6, ^3=23.05, and 
Cl=35-45- 
2 Compt. rend., 116, 147 i8q?. Results calculated with 0=15.96 0=35.37, KC1=74-4o. 
and H=i. OF ATOMIC WEIGHT. 
7 
made in which the palladio-chloride was dried in vacuo at ioo°. 
Wt. salt. 
Wt. Pd. 
Wt. KC1. 
At. Wt. Pd. 
Salt : Pd. 
Pd : 2KCI. 
I-3635 
O.4422 
O.6186 
105-374 
106.368 
3.0628 
0-9944 
I-3929 
105-545 
106.229 
I-4845 
O.4816 
O.6782 
105.405 
105.694 
r-7995 
O.5838 
0.8206 
105.427 
105.861 
The mean of the values found from the ratio K2PdCl4 : Pd is 
105.438. If O=i6, this becomes Pd= 105.702. With 0 = 
15.88 we get 104.910. The KC1 values were rejected. 
Still another set of determinations was based upon the reduc- 
tion of the double chloride in a steam of hydrogen, the data 
being in other respects like the foregoing. Results as follows : 
At. Wt. Pd. 
Wt. salt. 
Wt. Pd. 
Wt. KC1. 
Salt: Pd. 
Pd : KC1 
2.4481 
0-7949 
I.Il68 
105.560 
IO5.9I I 
1.8250 
0.5930 
O.8360 
IO5.671 
105.656 
It is noticeable that these values, in the set K2PdCl4 : Pd, run 
higher than in the eloctrolytic series. Inasmuch as the latter 
experiments were made in presence of free hydrochloric acid, it 
would seem possible that chlorine may have been liberated, pre- 
venting complete precipitation of the palladium, and so slightly 
lowering the apparent atomic weight of the latter. At all events 
the resuts obtained are not sharply concordant, and are, there- 
fore, inconclusive. 
Barium.-An elaborate memoir upon the atomic weight of 
barium, by Richards,1 makes a notable change in the value of 
that constant. Barium bromide was the compound investigated, 
and the procedure consisted in titrating its solution with solu- 
tions containing known quantities of silver. In many of the ex- 
periments the silver bromide thrown down during titration was 
collected and weighed. From the weights of Ag and of AgBr 
found, the ratio Ag : Br may be computed as a check on the 
accuracy of the work. 
Full details are given concerning purity of materials, pro- 
cesses, weighings, etc., for which the original memoir must be 
consulted. The final results are stated in the following col- 
ums : First, BaBr2 corresponding to 100 parts Ag. Second, 
1 Proc. Amer. Acad., 28, 1. Presented Jan. n, 1893. 8 
F. W. CLARKE. DETERMINATIONS 
BaBr2 corresponding to 174.080 parts AgBr. Third, percent- 
age of Ag in AgBr. 
Exp. 
I. 
11. 
ill. 
I 
... 137.746 
I37-783 
57-46o 
2 
••• I37-736 
i37-76o 
57-455 
3 
I37-723 
4 
••• I37-732 
5 
••• 137-735 
137-739 
57-447 
6 
••• I37-748 
I37-748 
57-445 
7 
••• 137-747 
8 
... 137.740 
137-747 
57-448 
9 
••• 137.755 
I37-748 
57-442 
IO 
••• I37-738 
I37-752 
57-451 
ii 
... 137-747 
I37-772 
57455 
12 
I37-726 
14 
••• 137-75° 
137-745 
57-443 
15 
••• I37-756 
137-754 
57-445 
16. 
••• I37-73I 
17 
••• I37-748 
18 
137-745 
19 
••• 137-759 
I37-758 
57-445 
Mean, omitting i and 2... 
••• 137-745 
137-747 
57-448 
Stas found • • 
•• 57445 
The last column, by its agreement with Stas' measurements, serves to 
confirm the other two, from which the atomic weight of barium is easily 
computed. From the ratio Ag2:BaBr2, when Ag=io7.93 and Br=79-955, 
Ba=i37-426. 
From the ratio 2AgBr:BaBr.„ 
Ba=i37.43i. 
The value 137.43 then must replace the lower value commonly 
held hitherto. The new results appear to be thoroughly estab- 
lished and unimpeachable. 
Thallium.-Lepierre1 has redetermined the atomic weight of 
this metal by several distinct methods. First, by the electrolysis 
of thallous sulphate in presence of an excess of ammonium oxa- 
late, weighing first the sulphate and then the metal obtained. 
The latter was dried and weighed in an atmosphere of hydrogen. 
The corrected weights and results are as follows: 
Wt. T12SO4. 
Wt. Tl2. 
At. Wt. Tl.(0=15.96), 
1-8935 
1-5327 
203.52 
2-7243 
2.2055 
203.67 
2.8lI2 
2.2759 
203.69 
1 Bull. Soc. Chim.. (3) 9, 166, March 20, 1893. OF ATOMIC WEIGHT. 
9 
Secondly, a weighed quantity of crystallized thallic oxide was 
converted by means of sulphurous acid into thallous sulphate, 
which was then subjected to electrolysis as in the preceding 
series. The results are subjoined. 
Wt. T12O3. 
Wt. Tl2. 
At. Wt. Tl. 
3.2216 
2.8829 
203.76 
2.5417 
2.2742 
203.53 
In the third series of experiments a definite quantity of thal- 
lous salt was fused in a polished silver crucible with ten times 
its weight of absolutely pure caustic potash. Thallic oxide was 
thus formed, which was washed with water and alcohol under 
various precautions, dried at ioo°, and finally weighed in the 
original crucible. 
At. Wt. Tl. 
3.1012 T12SO4 gave 2.8056 T12OS 203.57 
2.3478 " " 2.1239 " 203.44 
2.7591 TINO., " 2.3649 " 203.73 
Finally, crystallized thallic oxide was reduced in a stream of 
hydrogen, and the water so formed was collected and weighed. 
Wt. T12O3. 
wt. h2o. 
At. Wt. Tl. 
2.7873 
0.3301 
203.54 
3.9871 
0.4716 
203.82 
4.0213 
0.4761 
203.00 
The mean of the eleven figures given for thallium is 203.57, 
but stated by the author as 203.62. If 0-16 this becomes 
Tl-204.08 ; and with 0=15.88, T1=2O2.55. It will be observed 
that the values found in the individual experiments vary among 
themselves through a range of over eight-tenths of a unit, and 
that, therefore, the final mean result is entitled to little consid- 
eration. M. Lepierre is evidently unfamiliar with the deter- 
minations of his predecessors; for, quite ignoring the admirable 
work of Crookes, he speaks of the currently accepted value for 
Tl as resting without proof upon the calculations of Meyer and 
Seubert. He also neglects to state whether his weighings were 
reduced to vacuum standards, and he gives no evidence bearing 
upon the absolute purity and definiteness of the several com- 
pounds weighed. The metallic thallium which served as a 
starting point for the investigations, seems to have been carefully 
purified. 
The Composition of water by Volume.-The volume ratio be- 10 
f. w. clarke, determinations 
tween H and O in water, which controls the atomic weight of 
oxygen as deduced from its density, has been redetermined by 
Scott.1 The report published is only in abstract, but the sub- 
joined values for the ratio, H2: O, are given. 
2.0020 
2.0024 
2.0026 
2.0030 
2.0022 
2.0025 
2.0029 
2.0023 
2.0028 
2.0017 
2.0022 
2.0028 
Mean, 2.00245, ± 0.00007. 
Combining this mean with the density ratio, O :H:: 15.882 :1 / 
gives for the atomic weight of oxygen the value 15.862. 
Scott gives mean results of several other series of experiments, 
but only this one in detail. The hydrogen was obtained from 
palladium hydride, and the oxygen from silver oxide. With 
only an abstract available, the work done cannot be critically 
compared with Morley's. 
The Density of the Principal Gases.-These data, as determined 
by Lord Rayleigh,3 having important bearing upon atomic 
weights, fall properly within the scope of this report. Full de- 
tails are given as to the apparatus used, and all corrections are 
applied ; such, for instance, as for the difference in volume of 
the glass globes when filled or exhausted, for the value of grav- 
ity at the laboratory in latitude 5i°47", etc., etc. The final 
results are as follows, with air as unity. 
H 0.06960 
N 0.97209 
0 1-10535 
The weights in grammes per liter are : 
1 Proc. Roy. Soc.t 53, 130. Read March 23,1893. 
2 Rayleigh. 
3 Proc. Roy. Soc., 53, 134. Received March 4, 1893. OF ATOMIC WEIGHT 
11 
H 0.09001 
N 1.25718 
O M2952 
Air 1.29327 
Two of these constants have also been remeasured by Leduc;1 
who gives the following weights for one liter each: 
Air 1.29315 
N 1.2570 
Miscellaneous Notes.-Richards and Rogers, in a paper upon 
the occlusion of gases by metallic oxides,2 show that when the 
oxides of copper, nickel, zinc, and magnesium are prepared by 
the ignition of the nitrates, they retain appreciable quantities of 
gaseous impurities, such as nitrogen and oxygen, especially the 
former. The oxides of cadmium, mercury, lead, and bismuth, 
similarly prepared, contained no gaseous occlusion, excepting a 
mere trace in the case of cadmium. This observation reveals a 
possible serious source of error in several determinations of 
atomic weight, particularly in the cases of zinc and magnesium. 
In consequence of this fact, the atomic weight of zinc is now 
under reinvestigation in the Harvard laboratory, the compounds 
selected for study being the chloride and bromide. 
In the study of the rare earths, rough determinations of atomic 
weight are of great importance in the identification of material. 
Commonly they are made by the sulphate method, which is 
familiar to all chemists. Gibbs,3 however, in a study of these 
substances, makes use of the oxalates. By ignition, the oxide 
is left and can be weighed, while the proportion of oxalic acid 
is easily determinable by titration with permanganate solution. 
In some cases the oxalates were ignited with weighed quantities 
of pure sodium tungstate, the increase in weight representing 
R2O3. From the ratio between R2O3 and C2O3 the molecular 
weight of the oxide is easily calculated, and agrees well with 
determinations made by the sulphate process. 
Hinrichs, in a series of communications published in the 
Coinptes Rendu s' attempts to show that Stas' determinations 
1 Compt. rend., 117, 1072. 
2 Proc. Amer. Acad., 28, 200. Presented May 10, 1893. 
8 Am. Chern. J., 15, 546, November, 1893. 
4 On the KC1O3 work, in 115, 1074. On the lead determinations, in 116, 431. 12 
F. W. CLARKE. DETERMINATIONS 
of atomic weights are vitiated by systematic errors, depending 
upon the quantities of material weighed. According to him the 
larger the weight of the substance employed the larger the error; 
and he constructs curves to represent his views graphically. 
His mathematical procedure, however, his "method of limits," 
is not clear, and it savors somewhat of assuming Proutian values 
in advance, and then regarding variations from them as errors 
for which corrections should be applied. At all events he fails 
to point out, on the basis of experimental evidence, how the sup- 
posed errors were produced, and he does not show that they 
could have been of sufficient magnitude to account for the 
divergencies to be explained. If Stas' work is to be set aside, 
it can be by experimental researches only. 
In a third1 paper the same author gives his mathematical pro- 
cess in some detail, and still later he discusses Dumas' experi- 
ments relative to the atomic weight of oxygen,2 seeking to show 
that it is exactly sixteen. Very recently he has also discussed 
the determinations of the same constant made by Keiser, Cooke 
and Richards, Morley, and Dittmar and Henderson, reaching 
the conclusion that the ratio H:O=i:i6 rigorously.3 He ob- 
jects especially to the use of averages as fallacious, but does not 
fully establish the legitimacy of his conclusions. 
In another journal4 Hinrichs criticizes Richards' researches 
upon the atomic weight of copper. After showing that several 
published determinations of the electrolytic ratio between silver 
and copper agree quite well with the ratio 108 : 63.5, he takes up 
in particular Richards' analyses of copper sulphate, and points 
out that a part of the data correspond with the even values 
O=i6, S-32, and Cu-63.5. The other work of Richards he 
dismisses as erroneous, but neglects to show wherein the errors 
lie. 
In still another paper6 Hinrichs proposes to abandon gases 
like O and H as bases for atomic weight determinations, and to 
1 Compt. rend., 116, 695. 
2 Conipt, rend., 116, 753. 
8 Compt. rend., 117, 663. 
•4 Chern. News, 68, 171, October, 6, 1893. 
5 Conipt, rend., 117, 1075. OF ATOMIC WEIGHT. 
13 
adopt carbon (diamond) = 12 as the fundamental standard, on 
the ground that it can be exactly weighed. Oxygen and cal- 
cium are combined with this as substandards, becoming sixteen 
and forty respectively. 
Replies to Hinrichs' criticism of Stas have been published by 
Spring' and by Van der Plaats.2 Both writers object to the pro- 
cesses of extrapolation of which Hinrichs makes use, and Van 
der Plaats shows by tabulation of data that the systematic errors 
alleged to exist in Stas' experiments are not there. Only for- 
tuitous errors occur, and these are of insignificant magnitude. 
Furthermore, the ratios determined by Stas are so numerous, so 
varied, and so concordant in their final results that the pre- 
sumption in their favor is overwhelming. 
In the same paper, indeed, as its main object, Van der Plaats 
gives the results of a recalculation of Stas' twenty-one ratios by 
the method of least squares. He does not, however, give his equa- 
tions nor explain the special form of the method which he used. 
The results are stated in two tables, one, (A), computed upon 
the supposition that each ratio is entitled to weight inversely pro- 
portional to the square of its mean error; the other, (B), regard- 
ing all as of equal weight. O=i6 is the base of the system. 
A 
B 
Ag 
. 107.9244 ± 0.0136. 
I07.9202. 
Cl 
• 35 4565 ± 0.0049. 
354516. 
Br 
• 79-9548 ± O.OIOI. 
79-9497- 
I 
■ 126.8494 ± 0.0166. 
126.8445. 
s 
■ 32.0590 ± 0.0085. 
32.0576. 
K 
- 39.1403 ± 0.0059. 
39-I4I4- 
Na 
■ 23.0443 ± 0.0043. 
23-0453- 
Li 
7.0235 ± 0.0051. 
7.0273. 
Pb 
206.9308 ± 0.0404. 
206.9089. 
N 
14.0519 ± 0.0078. 
14.0421. 
NHt 
18.0740 ± 0.0034. 
18.0760. 
Hence H= . 
■ • 1.0055 ± 0.0021. 
1.0085. 
Another contribution to the Hinrichsian literature is furnished 
by E. Vogel,3 who criticizes Stas at considerable length. His 
1 Chem. Ztg., 17, 242, February, 22, zip?. 
2 Compt. rend., 116, 1362. 
s Bull. Acad. Bruxelles, (3), 26, 469. 14 
F. W. CLARKE. DETERMINATIONS 
paper is preceded by adverse reports from W. Spring and I,- 
Henry. 
In conclusion I submit a table of atomic weights revised to 
January 1, 1894. Oxygen=i6 is taken as the base of the sys- 
tem, but for provisional reasons only. Before long, with im- 
proved determinations, it may be practicable to return to the 
more philosophical H=i, when the entire system can be trans- 
formed once for all into something like permanent shape. A 
premature transformation of this kind, however, would only work 
confusion, without corresponding benefit. 
Name. 
Atomic Weight. 
Aluminum.... 
27. 
Antimony .... 
120. 
Arsenic 
75- 
Barium 
137-43 
Bismuth 
208.9 
Boron 
11. 
Bromine 
79-95 
Cadmium 
IT2. 
Caesium 
132.9 
Calcium 
40. 
Carbon 
12. 
Cerium 
140.2 
Chlorine 
35-45 
Chromium 
52-1 
Cobalt 
59- 
Columbium - • • • 
94- 
Copper 
63.6 
Erbium 
166.3 
Fluorine 
19- 
Gadolinium. • • ■ 
156.1 
Gallium 
69. 
Germanium - • • ■ 
72-3 
Glucinum 
9- 
Gold 
197-3 
Hydrogen 
1.008 
Indium 
II3-7 
Iodine 
126.85 
Iridium 
I93-1 
Iron 
56. 
Lanthanum - - - • 
13^-2 
Lead 
206.95 
Lithium 
7.02 
Magnesium - - - - 
24.3 
Manganese - • • • 
55- 
Mercury 
200. 
Molybdenum - - 
96. 
Name. 
Atomic Weight. 
Neodymium. • .. 
.... 140.5 
Nickel 
58.7 
Nitrogen 
14.03 
Osmium 
.... 190.8 
Oxygen 
.... l6. 
Palladium 
.... I06.6 
Phosphorus• • • • 
.... 31. 
Platinum 
195- 
Potassium 
39.II 
Praseodymium • 
•••• 143-5 
Rhodium 
103. 
Rubidium 
85.5 
Ruthenium 
.... 101.6 
Samarium 
.... 150. 
Scandium 
44- 
Selenium 
.... 79. 
Silicon 
28.4 
Silver 
•••• 107.92 
Sodium 
.... 23.05 
Strontium 
.... 87.6 
Sulphur 
.... 32.06 
Tantalum 
.... 182.6 
Tellurium 
•••• 125. 
Terbium 
.... 163. 
Thallium 
.... 204.18 
Thorium 
.... 232.6 
Thulium 
.... 170.7 
Tin 
.... 119. 
Titanium 
.... 48. 
Tungsten 
.... 184. 
Uranium 
.... 239.6 
Vanadium 
.... 51.4 
Ytterbium 
•• •• 173- 
Yttrium 
.... 89.1 
Zinc 
•••• 65.3 
Zirconium OF ATOMIC WEIGHT. 
15 
Discussion.-Dr. Wiley. By our organization we are able to 
present to American chemists the latest and most accurate num- 
bers for atomic weights. It is a matter for congratulation that 
this has been accomplished by the American Chemical Society.