I04 



NATURE 



\yunc 6, 1872 



collected therein so soon evaporates, especially in hot 

 weather; vide " Symon's British Rainfall, 1868 ;" " Rain 

 Gauge Experiments at Strathfield Turgiss, Reading," by 

 the Rev. C. H. Griffith, F.M.S., &c., p. 23, which further 

 establishes my remarks. The absence of outlet for con- 

 fined air heie fpoken of, might be remedied, if indeed 

 needed, by drilling small air-holes in the bottom of the 

 receiving cylinder and upright flange, hut net facing each 

 other. I have two Glaisher's gauges fitted with the in- 

 verted flange arrangement, both of which answer re- 

 maikably well. 1 believe the improvement which I have 

 adopted is more effectual against loss by evaporation 

 (during all weathers) than the present (Glaiskei's) system. 



One of my gauges has been further improved at my 

 suggesticn by being fitted with a spiral or helical pipe in 

 the place of the J -shaped pipe, thereby presenting no direct 

 opening for evapoiation, at the same time offering little or 

 no hindrance to the speedy descent of the rain-water ; 

 but this is a matter I hope to enlarge upon in a paper (as 

 yet unpubUshid) which I hope to communicate to you 

 shortly, '■ On a Proposed New Form of Rain Gauge (the 

 Atmospileomtter)," m which a similar, but more extensive, 

 idea, is shown. 



The particulars mentioned in paragraphs i and 3 of 

 this letter were lorg since communicated to and approved 

 by the Secretary of the Rainfall Committee of the British 

 Association (G. J. Symons, Esci , F.M.S., &c.) one of the 

 highest authorities in matters relating to rainfall. 



John Ja.mes Hall 



WATER ANALYSIS 

 I. 



IT is now upwards of twenty years since the inhabitants 

 of this country, and especially of the metropolis, weic 

 awakened, by a succession of virulent attacks of epidemic 

 cholera, from the profound indifference with which thty 

 had regarded all matters connected with the public health 

 for the hundred and eighty years which had succeeded 

 the Great Plague. During that interval builders had been 

 allowed to cover land with hundreds of acres of dwellings 

 built without re.gard to ventilation, drainage, or water 

 supply. In all the towns and villages of the country the 

 ground was honeycombed with cesspools and wells, the 

 latter deriving their supply of water at least in part from 

 the former, anel in all riverside towns the liver either re- 

 ceived the town sewage at once or after it had passed 

 through the cesspools. Attention once being drawn to 

 the matter, it became the duty of the chemist to detect 

 the various polluting matters introduced by the sewage 

 into the different sources of water-supply, and to discover, 

 if he could, waters that were free from this pollution ; and 

 a still greater freld was opened up, for the two first in- 

 cjuiiies naturally led to the allied questions — How can 

 sewage be rendered harmless ? and Can slightly polluted 



water be rendered sa'ely drinkable by the removal from 

 it of the contaminating matter introduced by sewage? We 

 propose in this article to look only at the first of 

 these e|uestions, the one on the successful solution of 

 which dcpend'all the others — Can organic contamination 

 be detected and estimated with accuracy ? As soon as 

 the question was approached it was found to be one of 

 extraordinary difficulty, and in 1856 Hofmann and Blyth 

 drew attention to the inaccuracy of the then existing pro- 

 cesses, especially of the one known as " loss on ignition " 

 obtained by igniting the solid residue on the evaporation 

 of the water. This loss, then generally looked on as 

 affording a measure of the organic matter present in the 

 water, they proved to consist of a loss of carbonic anhy- 

 dride, nitric acid, ammonia, and moisture, &c., and they 

 proposed to render the determination more accurate by 

 the addition of a known weight of sodic carbonate, which, 

 while it drove off the ammonia (usually a very small frac- 

 tion of the loss), retained the acids and prevented the 

 aqueous magnesic chloride from losing hydrochloric acid. 

 The same chemists pointed out the necessity of determining 

 the amount of nitrogen present, but were unable to recom- 

 mend any process for its estimation. The methods for 

 estimating the ammonia were also very unsatisfactory, for 

 we find Dr. Dundas Thomson in 1855 distilling as much as 

 fifty gallons of the metropolitan water-supplies in order to 

 estimate the ammonia, which was done by titration with 

 standard acid ; and this when some of the metropolitan 

 supply was taken from the Thames at Vauxhall, and 

 " Fibrin from F:eces " could be distinctly recognised in 

 the Southwark Company's water. Another process, de- 

 vised by Forchhammer, was also in use for the deter- 

 mination of the organic inatter, which consisted in 

 adding a standard solution of po'assic permanganate 

 until no further loss of colour occuired. This process 

 had been improved from time to time, anel was and 

 is largely used. The only other test was that for hard- 

 ness, invented by Dr. Clark, and which is still in use, 

 with but slight modification from the original m.elhod.' 



If these processes are considered but shortly, the de- 

 fects they possess are at once apparent. Take, for in- 

 stance, Hofmann and Blyth's improved solid residue 

 process. On ignition there was great danger of decrepi- 

 tation and consequent loss, notwithstanding the high 

 temperature (120' to 130" C.) to which the residue had been 

 exposed. Frankland and Armstrong have shown that por- 

 tions of the nitrogenous matter were liable to remain 

 fixed in the ignited residue as cyanogen compounds. 

 Again, in the case of some artificial residues prepared by 

 treating dilute solutions of urea as in the above process, 

 from 44 to 59 per cent, of the urea used was found to have 

 been lost during the preliminary evaporation, the sodic 

 carbonate having expelled it as amnionic carbonate. And 

 on the treatment of similar residues by ignition, from 58 

 to 85 per cent, of the organic matter was left in the resi- 

 due. It was usual also to restore the lost carbonic anhy- 

 dride to the ignited residue by evaporating a solution of 

 that gas on it and weighing until a fresh treatment did 

 not increase the weight ; but to still further increase the 

 difficulty of this unhappy process, it was shown that some 

 residues seemed to have the power of taking up such 

 quantities of carbonic anhydride, that they weighed more 

 after this treatment than they did before ignition. The 

 est'mation of ammonia by titration with acid neeels no 

 argument against it ; the enormous quantities of water 

 necessary for the determination sufficiently condemn it ; 

 and it has been long superseded by the admirable quan- 

 titative form of the Nessler process invented by the late 

 Mr. Hadow, of King's College. The permanganate ])ro- 

 ccss, however, being an easy one to perform, still survives 

 in the laboratories of many analysts. Indeed, not content 

 with giving the results of this determination as " oxygen 

 required to oxidise the organic matter present," the lively 

 imagination of some led them to the rctnarkable conclu- 



