The most convenient mode of analysis was found to be as follows. About 1 gram of the salt was ignited over a burner, by which treatment the cobalt was converted into tricobalt tetroxide, Co304) whilst the potassium and sodium remained as nitrites. The residue was boiled in water and filtered, the cobalt being weighed as Co304, or as metal, after reduction by hydrogen in a Rose's crucible. The filtrate, after acidification with hydrochloric acid, was evaporated to dryness, gently ignited, and the residue weighed as a mixture of potassium and sodium chlorides. Mixed KCl and NaCl...... 45.3 45.8 45.5 45.5 per cent. The potassium was estimated in the mixed chlorides by the usual platinic chloride method. The alcoholic filtrate from the potassium platinichloride was evaporated, the platinum removed by reduction, and the sodium weighed as chloride, and again as sulphate. The nitroxyl (NO2) was estimated by boiling the salt with sodium hydroxide solution, filtering off the cobalt hydroxide, which was weighed as tricobalt tetroxide, Co3O4, and titrating the filtrate with potassium permanganate solution, after acidification with sulphuric acid, as described later. The water of crystallisation was not readily given off, even at 130°. It was estimated directly by igniting the salt in a boat in a combustion tube, passing the evolved gases over heated reduced copper gauze to destroy the oxides of nitrogen, and collecting the water in a weighed calcium chloride tube. So far as the authors have been able to ascertain, the dipotassium sodium salt, K,NaCo(NO2),H2O, has not been previously described. Solubility of the Salt.-The solubility of the salt was determined as follows. About 1 gram of the salt was suspended in 200 c.c. of 10 per cent. acetic acid and shaken repeatedly during 3 days; it was then allowed to settle for 24 hours. The slightly turbid liquid was removed by a pipette and filtered, but could not be made perfectly clear, even by filtration through a hardened filter. Fifty c.c. of the filtered solution were boiled with sodium hydroxide solution, filtered from the precipitated cobalt hydroxide, acidified with dilute sulphuric acid, and titrated with N/10 permanganate as before; 0.6 c.c. was required, which makes the solubility rather less than 1 part of the salt in 20,000. Gravimetric Estimation.-The definite composition and the insolubility of this salt make it suitable for the gravimetric estimation of potassium, as the following results show: 10 c.c. of a 1 per cent. solution of pure potassium chloride were precipitated by 10 c.c. of the sodium cobaltinitrite reagent, with the addition of 1 c.c. of strong acetic acid. The precipitate was allowed to stand overnight, filtered either through a weighed Gooch filter, or weighed filter paper, washed with 10 per cent. acetic acid until the washings were colourless, and, finally, once with water, dried at 125° until the weight was constant, and weighed. Wt. of KCl, calc. from wt. I. II. III. Mean. Taken. of K,NaCo(NO2),H2O... 0103 0-101 0.096 01 0.1 gram. These numbers suffice to show that, under the conditions of the experiment, the potassium is completely precipitated, and that the method gives accurate results with very little trouble. To secure reliable results, it is most important to precipitate from a solution containing between 0.5 and 1 per cent. of potash, KO. At this concentration, the precipitate settles readily, and in such a physical state that it is easily retained by a filter. At greater dilutions, the precipitate settles out very slowly, adheres very firmly to the sides and bottom of the beaker, and passes through the filter so persistently that at a concentration of 0.2 per cent., the weight of precipitate obtained was 6 per cent. below the calculated, whilst at 0·1 per cent. only two-thirds of the calculated amount could be collected. The deficiencies are far larger than the solubility of the salt could account for, and are apparently due to mechanical difficulties in collecting so fine a precipitate. Volumetric Estimation.-The precipitation and washing are carried out as above; the filtration must, however, be made through asbestos in a Gooch filter, as paper interferes with titration by potassium permanganate. When washed, the precipitate and the asbestos plug are blown out of the Gooch filter into a beaker, and boiled with dilute sodium hydroxide solution. The asbestos and the precipitated cobalt hydroxide are filtered off, and the filtrate containing all the nitrite of the precipitate is made up to 100 c.c. The authors find the most convenient method of titrating to be as follows. 20 c.c. of the solution are acidified and rapidly titrated by N/10 potassium permanganate solution. As some loss of oxides of nitrogen may occur on acidification, a second and more accurate titration is made by adding the volume of permanganate solution thus found to the nitrite solution before acidification, and then adding more permanganate solution until a permanent colour is produced. As an example, 20 c.c. took 16.3 c.c. of permanganate solution when acidified first, but when to another 20 c.c., 163 c.c. of the permanganate solution were added, followed by dilute sulphuric acid, it was found that a further addition of 0.1 c.c. of permanganate solution was required to produce a permanent coloration. The end point is determined quite sharply by 1 drop of permanganate solution producing a colour which lasts for 1 minute. Estimation of Potash in Manures and Soils.-The method is very suitable for the estimation of potash in such commercial fertilisers as muriate, or sulphate of potash, kainite, &c., or in soils. The method adopted by the authors for fertilisers is as follows: weigh out 10 grams of muriate or sulphate of potash, or 40 grams of kainite, and dissolve in water, making the solution up to 1 litre. Filter off a portion through a dry filter, take 10 c.c., add 10 c.c. of the sodium cobaltinitrite reagent and 1 c.c. of acetic acid, allow to stand until the precipitate has settled, pass through a Gooch filter, and proceed to the volumetric estimation as above. The following results have been obtained: By cobaltinitrite volumetric method ...... Commercial Sulphate of Potash : K20 per cent. 49.67 49.64 I. II. ...... I. 12.22 11.94 12.70 III. By cobaltinitrite volumetric method......... Kainite: By cobaltinitrite volumetric method By platinic chloride gravimetric method... 11.98 12.07 12.53 For soils, the authors have found the following method to give very satisfactory results. The soil is extracted by heating 10 grams in a loosely stoppered Jena glass flask with 20-30 c.c. of strong hydrochloric acid for 48 hours. The solution is then freed from all bases which might interfere with the method, by boiling with excess of sodium carbonate. The precipitated bases are filtered off, and the solution concentrated, after addition of acetic acid, to 10 c.c., or to a volume which gives the concentration of potash recommended above. Ten c.c. of the sodium cobaltinitrite reagent are now added, with more acetic acid if necessary, and the precipitate allowed to settle. The volumetric estimation is now carried out as before. The following results have been obtained by various extractions of the same peaty soil under different conditions. The solution from each extraction was divided into aliquot parts, and comparative estimations of the potash made by the cobaltinitrite volumetric and by the platinic chloride gravimetric methods: The results as a whole show that the cobaltinitrite volumetric method is nearly as accurate as the methods hitherto available for estimating potassium by means of platinum, whilst it is much easier to work and presents no special difficulties. CAMBRIDGE. XCVII.-Notes on the Chemistry of Chlorophyll. By LEON MARCHLEWSKI, Ph.D., and C. A. SCHUNCK. DESPITE the great amount of work performed and published in regard to the question of the real spectrum of chlorophyll, scientific opinion on this subject is by no means unanimous. Chlorophyll is notoriously one of the most changeable compounds known, and the disregard of this circumstance has led several authors to ascribe to substances, isolated by methods in which the action of chemical reagents has not been excluded, the name of chlorophyll which they cannot possibly claim. 1. THE SPECTRUM OF CHLOROPHYLL. In order to be able to ascertain whether a certain process leads to the isolation of unaltered chlorophyll, or whether the product under observation is in fact a chlorophyll derivative, two methods have been used, one a physical and the other a chemical one. The physical method is the well-known optical one. The spectroscopical properties of the product isolated are compared under the same conditions with the properties of a solution of chlorophyll which has not been under the influence of any reagents whatever. As is well known, crude alcoholic green leaf extracts produce an absorption spectrum in the visible region which consists of four bands. between the lines B and F, and are also characterised in the more refrangible region, as one of us has shown (Proc. Roy. Soc., 1898, 63, 389), by three bands situated between F and the potassium Ks line, the first of which by artificial light is visible to the eye.* * As regards the spectrum given by the living leaf, our observations, using direct sunlight as the source of light, show that in a thin leaf only the characteristic band in the red is visible accompanied by a distinct lighter shading on its more refrangible side, the rest of the visible spectrum being continuous. In a denser leaf, the band and shading become of the same intensity, forming a broad intense band, and at the same time there is an indication of a band in the vicinity of the F-line. A still denser leaf cuts off all the visible rays more refrangible than the b-lines, but shows an indication of a band adjacent to D situated on its less refrangible side, whilst the band in the red has become still broader. This band, as seen in a thin leaf, lies approximately between A 6975 and x 6700, the shading extending to a 6500. In The relative intensity of the three first bands between B and Fis almost the same in all freshly prepared solutions, but the fourth band in the vicinity of the line E varies, being even sometimes darker than the third band, and sometimes hardly visible, a fact which suggests the supposition made by E. Schunck (Annals of Botany, 1889, 3, 65), that this fourth band does not belong to chlorophyll, but to some derivative of it. The question now arises whether these four bands in the less refrangible region, or at least three of them, and the three in the more refrangible region characterise a chemical individual, or whether they are caused by several substances. The majority of observers contend that the bands exhibited by a crude leaf extract between the lines B and F, or at least three of them, are indeed caused by a chemical individual-chlorophyll-a view which we also share, and we believe that the three bands in the more refrangible region are also due to the same individual, for the following reasons. It is well known that crude alcoholic chlorophyll extracts contain at least one green colouring matter and several yellow substances, the xanthophylls. It is possible, however, using the following method, based on the very elaborate and important observations of Sorby (Proc. Roy. Soc., 1873, 21, 442), to get rid of the greater portion of these accompanying colouring matters and at the same time to cause no alteration in the chlorophyll proper. This method, which is explained fully in the second part of the paper, depends upon the difference of solubility of the coloured constituents of crude chlorophyll extracts in alcohol or carbon disulphide, when the alcoholic extract is agitated with successive quantities of the latter, and by which means the chlorophyll can be obtained comparatively free from the other coloured constituents of the crude extract. On now examining the spectrum, it will be found to be exactly similar to that of the crude extract, with the exception that more of the ultra-violet is visible extending in our photographs as far as O,* this fact being no doubt due to the absence of certain members of the xanthophyll group which produce a general obscuration in this region of the spectrum. The solutions also appear to be slightly greener in colour, due to the absence an alcoholic extract of chlorophyll of such a strength as to show the spectrum to the best advantage, the band in the red lies between A 6850 and λ 6400, and in the weaker solutions between a 6750 and x 6600, so it appears that there is a shifting of this band still further into the red in the living leaf, and the spectrum given by the leaf appears to us to be characterised almost solely by this band, whereas the spectrum of an alcoholic extract is characterised by at least three distinctive bands between B and F. * The examination of the violet and ultra-violet region was accomplished by means of photography, an Iceland spar prism and quartz lenses being used, and the source of light was a Welsbach incandescent gas mantle of 60 candle-power. |