338 THE BNALYBT. ON BONE EXPERIMENTS WITH THE SILICATED CARBON AND SPONGY IRON FILTERS. By G. W. WIGNRB, F.C.S. WHEN staying at Margate recently I was dissatisfied with the character of the water supply, which was not, in my opinion, fit for drinking, and as it was therefore necessary to provide a filter, I considered it a good opportunity t o test two of the rival filters which are among the most extensively advertised.The London water, although it is not by my means perfect, i R yet too good t o give any value whatever to the results of such an experiment. The Margate water, besides being evidently contaminated by infiltration from the sea, has a very unpleasant smell and taste, and the microscope shows a variety of objects which might be more readily passed in surface drainage than in drinking water.It was therefore fairly suitable for such an experiment. The two filtem chosen for the purpose were the ‘( silicated carbon ’’ and the ‘ I spongy iron,” both were perfectly liew and bought direct from the makers. As regards the construc- tion I may just note for the information of those who have not seen both, that the principal mechanical advantages of the silicated carbon are-that the filtering body is a porous slab, cemented into the stoneware-so that it is impomible for any portion of the liquid to escape without passing through the entire thickness of the filtering material. The advantage of this< is obvious, while the fact that this porous slab is effectively aGrated every time the filter is allowed t o run dry is no small advantage. On theother hand the speed of filtration cannot be regulated, and in practice it is somewhat irregular.In the spongy iron filter the regulation of the speed of filtration, though somewhat troublesome to the user, is very efficient; but owing t o the very coarse grain of the principal filtering material it is not only possible but) probable that a good deal of the water escapes without having been subjected to the influence of the Bpongy iron at all.The rate of filtration is much dower than in the silicated filter. I did not consider it of any use to test these filters while quite new, as it must beTHE ANALYST. 830 evident that if tliey were not capable of working satisfactorily for a few weeks on such water as that used the experiment was concluded.They were therefore put up side by side-filled regularly as nearly as possible with the same quantity of water and at the same time, the amount being about two gallons per day each and allowed to work thus for fire weeks. During this time I found that the silicated water was always preferred for table use. Even those who were entirely ignorant of the fact that filters were used would sometimes complain of a flat taste in the spongy iron water.Since it is obvious, from what I have said, that each filter possesses adrantages peculiar t o itself, I determined, in addition to testing the water which had simply passed through each filter, t o test samples which had passed through both, and as i t seemed possible that the order in which the filters were used might make some difference, it was decided to try with silicated filter first and with spongy iron first.I therefore had in all five samples, viz. :- 1. Water from main. 2. Ditto filtered through silicated carbon filter. 3. Ditto filtered through spongy iron filter. 4. Bitto filtered 1st through silicated carbon and 2nd through spongy iron. 5. Ditto filtered 1st through spongy iron and 2nd through silicated carbon.The samples were distilled for ammonia as soon as possible. The following are the results of the fire analyses :- No. 1 No. 2 Tap in Silicated Main. Carbon Filter. Colour ... ... Yellow Green Pale Blue Suspended Matter . . . Traces None Smell ... ... Offensive Very Slight Taste ... .. Slightly Saline Slightly Opaque Offensive Hardness before boiling 27.20 Hardness after boiling 9-40 Total Solids ...... 69.60 Loss on Ignition ... 5.80 Total Mineral Matter 63-80 Chlorine as Chloride of Sodium ... } 36'66 Nitrogen as Ammonia Xone Do. Albuminoid} ,oo33 Ammonia Do. Nitrates -7500 Do. Nitrites Traces these 4 forms ... ) '7533 from Permaneanate 1 Total Nitrogen in Oxygen absorbed 16.90 5 30 66.90 6.80 60.10 36.66 None None Traces -7300 07300 None No.3 Spongy Iron Filter. Pale Blue None Offensive Saline and offensively flat 11.60 9-30 58.35 6.35 63.00 36-46 *Oi28 -0161 -6600 Traces -7489 -0630 " Microscopical Zxamina- Very bad Organic Satisfactory tion of Residue ... Animal and debris minute It was very vegetable traces only. free from debris. Some Mineral matter suspended small spores.traces. matter. Quartz. No. 4 1st Silicated. 2nd Spongy, Opaque Blue Traces Decidedly Offensive Unpleasant and very flat 1 4*0° 10.7O 4.80 66.90 36.50 60.70 -0875 00140 -6600 -7616 -0314 Satisfactory Traces No. 5 1st Spongy. 2nd Silicated. Chalky Blue Sligb t Traces Decidedly Offensive Excessively flat and ohjectionable 17.10 7 *2" 69.75 9.40 60.35 36-60 -0149 -0051 07300 ,7600 Traces *0183 Satisfactory On looking a t these results, the fnst general feature which deserves attention is that tlic two filtere when used together do not appear to have effected any additional340 THE ANALYST.purification, but rather the reverse. It is very difflcult to account for this fact. I n both cases an ample quantity of the once filtered water was allowed to flow through the second filter ao as to rinse it out, and the rinsings were rejected. It is true the filters were not continuously worked under these conditions for any length of time, but this ought hardly to be an essential for this experiment. Leaving this I would next point out the special features shown by columns No.2 and No. 3, which are the two filtered waters, as compared with column No.1, which is the unfiltered. Both filters have equally removed the yellow green colour of the water, and the ettspended matter. The silicated f&er has greatly reduced the smell, but the spongy iron has merely changed its character, so that instead of resembling decaying vegetable matter it is a sort 6f indescribable smell, perhaps more resembling the smell of impure hydrogen prodaced by the action of acid on iron than anything else; but of course very faint, though offensive.The silicated removed the offensive taste, and left only the saline. The spongy iron left the saline, but added to it a taste a little worse than if the water had been boiled, that is, there wae eomething besides the flat taste. Bischaf has alluded t o a gas which he has observed within the bulk of spongy iron after it had been in use some time, which gas he says is sometimes explosive, sometimes not, and on one occasion he found this gas t o contain a hydro-carbon ; may not this hydro-carbon be the cause of this smell and taste? The effects on the hardness are singular, The silicated reduced the temporary hardness by 6*2p, and the permanent hardness by 4*1°, total reduction 10.3O.The Bpongy iron reduced the temporary hardness by 15*5O, but only reduced the permanent hardness ? ~ y a decimal. The spongy iron therefore reduces the total hardness the most, but bearing in mind the greater importance of permanent hardness, there is probably little to choose. It is remarkable that the reduction of total mineral matter does not more closely follow this reduction of hardness.I have not yet made full mineral analyses of the residues to determine this point. It is, however, in the ammonia and albuminoid ammonia, which are so often viewed as almost the sole proof of purity or otherwise, that the most important resulte appear. The silicated removes even the small proportion of albuminoid ammonia which waa present, but the spongy iron produces a very marked proportion of free ammonia and so much albuminoid ammonia that the water would probably be condemned by those who tivst mainly to that determination. If the albuminoid ammonia determination is of value, then pro tanto the filtered water is worse than the unfiltered.I f the filter is a good one, and certainly many other experiments have pointed out that it is, then any analyst who reported on such a water, in ignorance of the fact that it had been through spongy iron, might be deceived.But still further, if ppongy iron acts in this way, is it possible that metallic iron may in a less degree? I n considering this point all the special circumstances of the case must be borne in mind. For instance, the very saline Character of the water, and the fact that this salt is partly due, in all probability, to infiltration from the sea, but still the fact rebains, The taste has apparently been affected in a similar way. This result is a serious one, in whichever light we view it.THE ANALYST. 34 1 'It is unnecessary to go at length through the other columns.It is well, however, to point out that when two filters were used, in the case where the spongy iron was last the ammonia and albuminoid ammonia were almost as high as in No. 3, while, where the silicated was last, they were far less, though not so low as when the water passed through the silicated only. It woald be very desirable to have reports of similar trials on the water suppls of other places, but I hope it will be possible t o take the precaution of using new filters and keeping them at work for a month before &king the samples.