D' (hydrochloric, ferrocyanide in excess). Ratio, 1 : 2.02 : 0.955, confirming D and corresponding to K,Cd Fe(CN)..

F" (acetic, ferrocyanide in excess). Ratio, 1 : 2.03: 0.95, confirming F and corresponding to K,CdFe(CN)。.

The abnormal results in G and G', probably due to cadmium hydroxide being carried down in varying quantities, necessitated a repetition of this work, so in July, 1901, G"" was made under the same conditions but was treated repeatedly with very large. quantities of ammonia before washing with water and then analyzed at once.

G" (ammonia, cadmium in excess). The iron cadmium ratio. was determined in duplicate and found to be I: 1.496, corresponding to K,Cd, (Fe(CN),)2.

When this precipitate was being washed it was noticed that there was a portion which did not settle as well as the rest; this was decanted from the part which settled well, washed and analyzed, giving a ratio of exactly I Fe: ICd.

In the results already given the duplicates on C' and the ratios on E and E' did not check satisfactorily, so these precipitates were made again (July, 1901) and as they settled well (cadmium being in excess) they were washed and analyzed within a week. C"" (hydrochloric, cadmium in excess). White, no decomposition; gave Fe: Cd: I: 1.07 and 1: 1.065.

E" (acetic, cadmium in excess).

gave Fe: Cd :: 1: 1.07 and 1 : 1.06.

White, no decomposition;

We see in these two a total lack of agreement in composition with those which were allowed to stand many months before analysis, indicating a change in composition after precipitation. On the other hand these last results are in exact agreement with the titration ratios in slightly acid solution;' for if the theory for IK,Fe(CN), or 1Fe: 1Cd' gives for the strength of the ferrocyanide solution I cc. = 0.00671 gram cadmium, then, if the ratio in the precipitate is I: 1.07, the strength of the solution in terms of cadmium becomes 1.07 X 0.00671 or 0.00718, while the average result of titration was 0.00717.

For ease of comparison let us disregard for the present the potassium, which has been found to be present always in the proper amount to satisfy the remaining valency of the ferrocyanogen radical and consider only the iron cadmium ratios.

1 Miller and Fisher: This Journal, 21, 542 (1900).

2 Based on Hermann's formula K2Cd Fe(CN)6.

(1) The results when ferrocyanide is in excess are entirely different from those where cadmium is in excess.

(2) In either acid or ammoniacal solution, ferrocyanide being in excess, the ratio is 1 iron to I cadmium or the precipitate is K,CdFe(CN), while in a neutral solution the cadmium ratio is higher.

(3) With cadmium in excess, in either neutral or hydrochloric acid solution the final composition is the same, corresponding to K.Cd, (Fe(CN)),, while in an acetic acid solution the final composition corresponds to K,Cd, (Fe(CN)).

10

(4) With cadmium in excess, in an acid solution the composition alters on standing with an increase of cadmium in the ratio, but when freshly precipitated the ratio agrees exactly with the results by titration.

(5) With cadmium in excess, in an ammoniacal solution the ratio exceeds the normal, but after washing with ammonia corresponds to Cd,Fe(CN),; when freshly precipitated and washed quickly with ammonia, the part which settles badly being removed by decantation, the composition agrees exactly with K,Cd, (Fe(CN)), while the part decanted corresponds to K,CdFe(CN)。•

1 Analyzed within a week of precipitation.

2 After treatment with ammonia.

3 Decanted portion.

(6) The constancy of the final composition of these precipitates does not favor the theory that potassium ferrocyanide is merely dragged down.

The results obtained when cadmium was in excess in an ammoniacal solution were so extraordinary that they deserved further study. The precipitate G"", K ̧Cd ̧(Fe(CN),)2, was treated seven times with strong ammonia (sp. gr. 0.90) to see whether a change could be effected by a difference in solubility in ammonia as the K,CdFe(CN), seemed the more soluble. In this way, a creamy white residue was obtained which gave on analysis Fe : Cd :: 1 : 1.99, showing Cd,Fe(CN), which was also obtained in G after rewashing with ammonia.

This result, together with the fact that the portion first obtained by decantation when the precipitate was originally treated with ammonia, gave a ratio of Fe: Cd of exactly 1: 1 shows that the original precipitate can be resolved into the two simple ferrocyanides Cd,Fe(CN), and K,CdFe(CN),, and affords important confirmation of the theory that these complicated precipitates are mixtures, advanced several years ago in connection with the ferrocyanides of zinc and manganese.'

Now, if this is true for the precipitate in an ammoniacal solution, it ought also to be true of those more complicated ones formed in acid solutions; accordingly, C", which gave 1:1.07 when freshly precipitated, was treated in the same way with strong ammonia, and the residual cream-colored portion analyzed. This gave a ratio of 1:1.90, showing that the same change had taken place, though it was not entirely complete.

To test this further, precipitate A (neutral, cadmium in excess, made June, 1900) was treated eight times with strong ammonia and the residue analyzed. This gave the ratio of 1: 1.99, agreeing almost perfectly with Cd,Fe(CN),.

In order to ascertain whether the change which takes place on standing can be hastened by heating, the precipitate E"" (acetic, cadmium in excess), which when washed quickly and analyzed gave a ratio of 1: 1.07, was heated with water containing acetic acid on a water-bath for three days and then analyzed. The result was Fe: Cd::1:1.40 compared with 1:1.383(E') after standing six months in the cold and 1: 1.392 (E) after one year. This precipitate (E"" after heating) was next treated eight 1 This Journal, 19, 556 (1897).

times, with strong ammonia, and the residual portion analyzed. The result was Fe: Cd: :: 1.97, showing again the presence of the normal cadmium ferrocyanide.

These experiments show that the ferrocyanides undergo a change in composition after precipitation which progresses to the same point under the same conditions and which is hastened by heating. The action of ammonia on these precipitates (cadmium in excess) has been confirmatory of the idea that they are either mixtures or else very easily decomposable double salts made up of K,CdFe(CN), and Cd,Fe(CN)。.

Written in this way the results are expressed as follows :

Character of

solution.

Cadmium in excess.

Ferrocyanide in excess.

4K,CdFe(CN)6.3Cd2Fe(CN)。 9K2CdFe(CN) ̧.Cd2Fe(CN)。

K,CdFe(CN)6

....

It would be premature to advance these formulas as representing relations of these complicated and variable precipitates, but the facts seem to warrant this as a working hypothesis, to be used in a further study of the other insoluble ferrocyanides.

Although the results are not all that are desired, many of the precipitates are so easily oxidized, readily decomposed and impossible to filter, that it seemed very doubtful whether any better results could be obtained if the work were repeated.

On comparing the results obtained on cadmium with those on zinc previously published' and with those of Stone and Van Ingen3 we see some cases of remarkable agreement as well as some discrepancies, but a discussion of these results is postponed as it is proposed to continue this work by a similar investigation of the ferrocyanides of the other metals of the periodic group containing cadmium and zinc and also of the analytical group containing zinc, manganese, nickel and cobalt so that, taking zinc as a starting-point, we can see whether the variations in composition have any connection with the analytical grouping or the periodic law and also possibly obtain more knowledge of their constitution.

PROXIMATE ANALYSIS OF SPENT ALKALINE LIQUOR FROM THE REDUCTION OF POPLAR WOOD FOR PAPER STOCK, BY THE SODA PRO

CESS, WITH A DESCRIPTION

OF THE METHOD.

BY MARTIN L. GRIFFIN.

Received November 11. 1901.

O far as I know, no attempt at analysis of this complex liquid

So fa

has ever been made, and I know of no references in literature relating to its composition. I have, at different times, made determinations of some of the constituents of this liquor, particularly the total alkali and acetic acid, but this is my first attempt at an investigation of the proximate constituents. The reason why the subject has not been taken up by technical chemists, employed in the paper industry, is because of the complex nature of this waste and the fact that manufacturers have had no idea that it contained anything valuable which they could recover except the alkali, for which greatly improved machinery has been contrived during the past fifteen years.

My interest in this subject was much increased by the action of the Scottish Paper Makers Association, which nearly two years ago offered prizes for the investigation of various waste products resulting from their paper industry, among which was the waste alkaline esparto liquor.

The subject is brought before you now simply in introductory form, but still in a way which I trust will be of some interest and value.

THE PROCESS BY WHICH THE SPENT LIQUOR RESULTS.

We are told that M. Meliner, a Frenchman, in 1865, was the first to discover the process of reducing spruce and poplar woods to paper stock by treatment with alkalies under steam pressure. The process now consists in cutting the wood into chips, the fiber of which is about inch long, and charging same into digesters holding from 3 to 5 cords. Strong sodium hydroxide. liquor, containing from about 90 to 100 grams of soda-ash per liter, is then furnished to an amount of about 4 gallons to each cubic foot of capacity of the digester. Steam pressure of 100 to 120 pounds is then applied for from seven to eight hours, when the