At the end of this time about 50 grms. of amorphous, insoluble matter remained in the flask. The acid liquor was filtered from this residue, neutralised with calcium carbonate, and evaporated to a thick syrup. Upon adding about three volumes of strong alcohol to this and boiling, a dark brown, tasteless gum was precipitated and removed, and the liquid after concentration yielded slowly a mass of dark coloured crystals.

After repeated washings and re-crystallisations, two successive portions of crystals were obtained, which were nearly white, sweet to the taste, and reduced Fehling's solution strongly. In all about 6 grms. of this product were obtained, or nearly 3 per cent of the 200 grms. of gum used. Probably an equal amount of sugar remained in the dark coloured and non-crystallisable mother liquors. The amount of this product obtained was disappointingly small in view of the promisingly preliminary indications. The syrups obtained after hydrolysis, however, contained much unchanged gummy matter, and crystallised only slowly and with great difficulty. It was also found subsequently that the extracted cobs still contained a very large amount of the furfurol-producing bodies. The caustic soda had evidently extracted only a very small amount of the pentans, or pentose gums, which accounts for the small yield of sugar. To characterise the product thus obtained, the specific rotation was determined, and its phenylhydrazine derivative prepared. Of the two portions of crystals, No. 1 showed the specific rotation (a)D=19'4°; No. 2, (a)D=19'7°. These numbers agree with sufficient accuracy with the specific rotation of xylose(a)D=18'5° to 19°-to identify both products as xylose, and show that this was the only pentose present. The phenylhydrazine compound was prepared in the usual way, had a normal appearance and behaviour, and showed the melting-point 159°-the approximate melting-point observed for phenylxylosazon (156°-160°). In distinction from the phenylarabinosazon, which has the same melting-point, this compound was distinctly lævo-rotatory, a property of the xylose derivative as shown by Fischer.

These data add a new and, as we hope to show by a repetition of the above work, a fruitful source of xylose to those already known, and emphasize once more the reliability of the furfurol test for the recognition of such materials.-American Chem. Journal, Vol. xiii., No. 5.

IN the CHEMICAL NEWS, vol. lxii., p. 230, November 14th, 1890, I gave a brief preliminary notice of these spectra in a paper on the "Phosphorescence of Lithium Compounds," and as I believe they have never previously been mentioned I now propose to describe them more fully.

Terminal spectra are of the same order as spark spectra obtained in air, their interest mainly lying in the method by which they are obtained, which is usually as follows:

A number of small vacuum tubes are prepared, 2 or 3 inches long and half an inch diameter, one terminal being a very small plate of aluminium, not more than in. square, and usually covered on one side by a mica screen, the other terminal being either the same or a small spiral of fine wire. Some metallic salt or compound is introduced, and the tube exhausted to the required degree, and then sealed up. To use it, a little of the contents is made to rest on one of the terminals by shaking or jerking the

* Ber. d. Chem. Ges., 23, 355.

tube, and the terminal (made negative) is examined with the spectroscope while coil discharge is passing, lines due to the metal being then seen superposed upon the spectrum due to the gas.

I at first hoped that the method would enable me to prepare standard tubes containing salts of various metals, which would give metallic spectra as readily and permanently as vacuum tubes containing gases, and with a few metals this is possible, but with the majority the method is inferior to the ordinary spark process in air. The metallic lines are easily distinguished from the gaseous lines seen at the same time. The latter are steady, extend the whole width of the spectrum, and are of equal brilliancy throughout; while the former flicker more or less, and often appear in flashes, and usually only extend partly across the spectrum. The C and F lines of hydrogen are almost always very strong.

It will be seen that these spectra are quite distinct from the phosphorescent spectra first studied by Mr. Crookes, which are of a totally different character, and are only visible at a degree of exhaustion much beyond the limit at which these spectra disappear.

Some fifty vacuum tubes of various designs were made and used, experiments being made with carbon terminal soaked in solutions of salts, with movable terminals with fused coatings, and with terminals of different metals. The following results were obtained, using a coil giving about 1 inch spark, which was hardly powerful enough for the work.

Lithium.-These spectra were first noticed while examining a specimen of LiNO3 in vacuo. The salt was heated with a lamp until partly volatilised, a portion condensing again on the aluminium terminals of the tube. On passing the discharge, the negative terminal shone with a deep red light, which was found to give the red and orange lines very strongly. All the salts of this metal give very good results.

Sodium.-The D lines are as common an impurity in these spectra as they are in spark and flame spectra, and the reaction seems almost as delicate. Using any sodium salt, the terminal is covered by an intense yellow glow, and the D lines are always strong. A slight increase of strength in the discharge brings out the double green line, and the double red line can be seen occasionally. NaCl has a bright greenish phosphorescence in a high vacuum. Casium.-The two blue lines (a and B) are seen occasionally when the chloride is used, but not very readily.

Calcium.-Ca(NO3)2 was used. The terminal shone points broke out here and there. The orange and green with a strong red glow, and brilliantly white incandescent bands came out very distinctly, and the bluish violet line 4226 was well seen in the flashes.

Strontium.-The chloride and nitrate were used, but with rather poor results. The red lines were almost invisible, green and blue lines were glimpsed in flashes, blue line 4607, which was easily seen. but the only one I could identify with certainty was the

Magnesium.-Using the sulphate or nitrate, the triple green lines came out strongly at times, but the effect was rather uncertain. Very intense green points occasionally seen on the positive terminal.

Zinc.-The oxide and sulphate were used. The red line 6360 was very strong, and three green lines, 4810, 4721, 4679, were also seen. ZnO phosphoresces in a high vacuum with a peculiar greenish yellow glow, and ZnSO gives a bright bluish phosphorescence.

CONTRIBUTIONS TO MICRO-CHEMICAL

ANALYSIS.*

By H. BEHRENS. (Continued from p. 5).

If a loss of precipitate is not an object we may have recourse to the method of filtration recommended by Streng by means of narrow slips of paper. The port object with the drop which is to be filtered is laid upon a small board having a slope of about 1: 20; a second is pushed up to the margin of the board so that the vertical distance of the two from each other is about 5 m.m. They are then connected by means of a slip of filter-paper of the shape of a Y, the arms of which are moistened so as to adhere to the upper glass, and the liquid to be filtered is brought into contact with the paper. The dimensions proposed by Streng, 2 m.m. in breadth, and 25 m.m. in length, may be reduced to 15 m.m. and 10 m.m. The quantity of liquid which is retained in and below the slip of paper amounts then to 5 m. grms. The precipitate does not fare so well, as much of it-if washing is necessary the greater part-soaks into the paper. This loss is diminished if an approximate separation can be effected by means of a wire as described above.

There remains, undeniably, room for improvements which will be very welcome if we do not lose what is good in our quest for something better. I have kept in view the small bulk of the apparatus, and the consequent possibility of working with it in all places, and at the same time the economy of time by avoiding unnecessary preparations. It must not be thought that especial genius is needed in order to work expeditiously and well with the simple means above described. Correct observation and independent thought are certainly needed and the operator must learn to work with little apparatus. All this applies, however, just as well to blowpipe analysis. If a fair attempt is made I am convinced that an experienced analyst who has made occasional use of the microscope will encounter no more serious difficulty with micro-chemical analysis than he once did with the blowpipe.

Reactions.

I.-Potassium.

1. With platinum chloride. Lemon-yellow sharply developed octahedra of K2PtCl6, refracting light strongly, sometimes assuming a hexagonal structure by flattening and combination with ∞∞. Size 10 to 50 micro. If rapidly deposited sometimes combined in threes and fours in the shape of a clover leaf or of a cross. Solubility 1: 100. Limit of the reaction o'0005 m.grm. K.

Use a strong solution of platinum chloride, which when drying up must not deposit octahedral crystals. A drop of this liquid of about o'2 m.grm. is placed on the specimen drop, which should be neutral or faintly acid. In dilute solutions the yellow octahedra only form after a few minutes along the margin. An excellent reaction, if the presence of cæsium, rubidium, and ammonium is excluded. An excess of strong acids, especially of sulphuric acid, reduces the sensitiveness of the reaction, but may be counteracted by an addition of sodium or magnesium

acetate.

2. With phosphomolybdic acid. Minute crystals very similar to those described under 1; combinations of O with O mostly globular, refracting light strongly, principally accumulated at the margin of the drop. Limit of the reaction about o'0003 m.grm. K.

In using phosphomolybdic acid it must be noted that its solution on evaporation yields fine yellow octahedra, even if no alkali is present. The octahedra of the pure acid are sharply developed with straight edges, measuring 30-60 micro., whilst those of the potassium compound are rounded and about three times smaller. The reagent must be added freely; the sample drop is acidified

* Zeit. Anal. Chemie.

with hydrochloric or nitric acid, and a large drop of the reagent is added. It is useful for strongly acid solutions, but it must be used with caution, as besides potassium it precipitates cæsium, rubidium, ammonium, and lithium.

3. With bismuth sulphate. Colourless hexagonal discs which are slowly developed to stellar, rhombohedric structures (R-ROR of 30-60 micro.), limit of the reaction 0'0002 m.grm. K.

Basic bismuth nitrate may be used, dissolved in nitric acid with the addition of a little sulphuric acid. The solution to be tested must first be concentrated. Nitric acid delays the reaction, but does not prevent it except in very large excess. Ammonium behaves like potassium with bismuth sulphate. Sodium is also precipitated, but in a quite different form.

4. With silicofluoric acid. Very small cubes with faint outline 10-20 micro.

This reaction proposed by Boricky is not recommended on account of the smallness and paleness of the crystals, but it sometimes makes its appearance unintentionally during micro-chemical work.

II.-Sodium.

1. With uranylacetate (Streng). Pale yellow sharply developed tetrachloride of about 50 micro. Limit of the reaction o'0008 m.grm. Na.

As a reagent we use a solution of uranylacetate in acetic acid. The sample liquid is to be strongly concentrated as the reaction is very faint at a dilution of 50. Strong acids should be expelled by evaporation, ammonium acetate being used only in cases of necessity, as an excess of ammonium salts interferes with the reaction. Streng adds that platinum chloride also interferes, which must be remembered with reference to potassium. 2. With uranylacetate and magnesium acetate (Streng). Rhombohedric crystals almost colourless; chiefly OR.R.-2R. P2,

3

according to the position and the predominance of tetrahedral, dodekahedral, rhombic, or hexagonal structure. Size up to 120 micro. Limit of the reaction o'0004 m.grm. Na.

This reaction is produced under the same conditions as 1, if along with sodium metals of the magnesium group are present in excess.

3. With silicofluoric acid (Boricky). Hexagonal discs and short columns of the combination oPP, measuring as much as 70 micro. From less dilute solutions there are thrown down elegant, six-leaved rosettes, of 80-120 micro. in size. Solubility 1: 150. Limit at o'00016 m.grm. Na.

Instead of silicofluoric acid may be used ammoniumfluosilicate if a granule is placed in the moderately acid sample. More massive crystals of sodium fluosilicate

have a faint rose colour. Platinum chloride does not interfere. In highly diluted solutions the reaction appears only as evaporation proceeds.

4. Bismuth sulphate. Colourless hexagonal rods 2080 micro. Limit of the reaction o'00004 m.grm. Na.

Compare potassium No. 4. The reaction is serviceable for the simultaneous detection of potassium and sodium. To this end a solution of bismuth sulphate in dilute nitric acid is spread out in a layer of o'3 m.m. in depth, and the sample in question is placed in the middle of the extended drop either as a solid or as a thick paste (sulphate). A short heating to 50-60° accelerates the precipitation of the sodium double salt which collects round the specimen. The discs of the potassium double salt appear later and diffuse themselves over the whole preparation.

Haushofer's test, potassium pyroantimoniate, did not give good results. The reaction was tedious and the crystals not well defined.

(To be continued).

THE SEVENTH INTERNATIONAL CONGRESS OF HYGIENE AND DEMOGRAPHY.

NEXT month will witness a gathering which promises to be the leading scientific event of the season, and which will, we hope, leave its mark behind in the shape of permanent results. Looking at the papers already announced, we may venture to doubt whether they can receive due discussion in the time allotted. We must hope, to this end, that the readers of papers and other speakers will not indulge in self-advertisement, such as is often introduced on such occasions; and that they will, above all things, eschew the introduction of political considerations, for which two, at least, of the sections might supply temptation.

Our attention is naturally drawn chiefly to Section V., "Chemistry and Physics in Relation to Public Health." This section will hear and consider the following papers :On Tuesday, August 11, a memoir on "Smoke and its Effects," by Dr. W. J. Russell, followed by a paper on "The Air of Large Towns, and Methods for its Analysis," by the Manchester Field Naturalists' Society, and a paper on the Means at our Disposal for Preventing the Emission of Smoke from Factories and Dwelling-Houses," by Mr. A. E. Fletcher, H.M. Chief Inspector of Alkali Works, &c. Dr. C. Merdon Williams will take part in the discussion with regard to "fog from a hygienic point

of view."

On Wednesday, August 12, there will be read papers on that much-disputed subject the "Treatment of London Sewage," by Dr. Dupré, F.R.S.; "Outline of the Various Chemical Processes for Purifying Sewage," by Dr. J. C. Thresh; "The Duties of a Locality to Secure the Nitrogen of its Sewage for the Benefit of the Nation," by Dr. A. Carpenter. It need scarcely be said that if the utilisation of such nitrogen is proved to be a duty, all those processes which aim at carrying sewage or sewage precipitates into the sea, and no less all those processes in which its constituents are converted into cements or otherwise destroyed as such, are at once condemned. Dr. A. Carpenter further contributes a paper on the power of soil and vegetation combined to destroy disease germs, and to prevent the possibility of a spread of enteric disease in consequence of sewage farms.

Herr Margraf, of Berlin, promises a paper, though he may possibly be unable to be present and take part in the discussion.

On Thursday, August 13, Prof. Dr. Lehmann, of Wurzburg, will bring forward a paper on "The Hygienic Importance of Copper," a subject as yet far from having been fully elucidated. There follow two other papers on subjects of great importance bearing upon public watersupply: "The Importance to be Attached to Magnesia in Drinking-Waters," by Prof. P. F. Frankland, F.R.S., and "The Action of Water on Lead," by Dr. J. H. Garrell.

Next follows a paper on "The Antiseptic Treatment of Food," by Mr. O. Hehner, and one on "The Effects of the Respiration of Carbonic Acid on Man," by Dr. W. Marcet, F.R.S.

THE long chemical war against phlogiston had been fought and won, and the thought of experimenters was turning in a new direction which was to institute a new war lasting only a little less long than the last, but the difference between the two cases was that whereas the phlogistic theory hung like a pall over the whole science, obscuring during its continuance the entire field; in this case the question in dispute was as to the ultimate constituents of matter, and none of the many views entertained on these questions interfered with the classification and assimilation of the myriads of facts which experiment and research were eliciting. This war, therefore, while it will serve to illustrate that the most eminent chemists share with the rest of the world the weaknesses of our common humanity, did not materially retard the progress of the theory of chemistry.

Proust (1755-1826) announced the unchangeable proportions by weight in which substances combine together; and that if they combine in more than one proportion it is by leaps and not gradually, as the water of the ocean becomes little by little more charged with salts brought down to it by the rivers. This was a great and pregnant discovery which at once led the way to the new field of battle, but the strangest thing about this announcement is that it was vehemently attacked by Berthollet (like Proust, a native of France), in a proposition which a little later seemed nothing but a stupid blunder or obstinate opposition, and yet in Berthollet's contention lay a precious truth only recently recognised and placed in its proper place.

Briefly, the skirmish between these two men was this. Proust discovered that the relative proportion to each other by weight of carbonic acid and copper in carbonate of copper was constant; no matter in how great excess one or the other of these bodies was present, the weight of the carbonate of copper was the same, and the weight of each constituent in it was invariable. For instance, substituting the accurate weights which better methods and apparatus have enabled chemists to obtain for the inaccurate approximations then made, in 123'4 grms. of carbonate of copper (there were always 634 grms. of copper, and 60 grms. of carbonic acid). It made no difference whether these weights of the two elements respectively, or whether two or three times as much of one with the above weight of the other were made to combine; the result was always that 123'4 grms. of the compound were found, and the excess of either element remained uncombined.

It is further expected that papers for this section will also be contributed by Prof. Liebreich, of Berlin, and Prof. Lunge, of Zürich.

We see with regret that much of the matter to be taken into consideration in Section VII.-Engineering in Relation to Hygiene-is in its character distinctly chemical and physical, such as the chemical and electrical treatment of sewage, the quality of potable water, its purification and the pollution of rivers-subjects which can only be discussed by chemists, and on which chemists alone are qualified to adjudicate.

Berthollet (1748-1822) maintained, on the contrary, that if different masses of two elements are brought together, there will be found in the compound more of that constituent which was in greater quantity before the union. On account of the high position which Berthollet held in the chemical world this view received respectful, though silent attention, for few of the masters of the science were won over by it; because Richter, Klaproth, Vauquelin, and Wenzel had placed the constancy of acid and base in a compound of the two beyond all question.

* Introduction to the Chemical Lecture Course at the Franklin Institute, November 10, 1890.

Proust, however, took up the gauntlet and followed each separate publication of Berthollet by a refutation based upon careful experiment. This lasted for eight years, or from 1799 to 1807, and was settled apparently for ever when Proust, by repeating some of Berthollet's own experiments on the successive stages of oxidation showed that his opponent had mistaken a percentage of water for a percentage of oxygen.

But Berthollet's main idea that the mass and the affinity were inseparable factors in the formation of a compound, after having been crushed to earth, was to rise again in more recent times by the labours of his countryment of almost similar name, Berthelot and St. Gilles, and by Guldberg and Waage; but they showed, not that the proportion by weight of the compound, but that the rapidity of the reaction was affected by the masses of the constituents.

This dispute and the rise and fall of a theory was only a slight skirmish, which was preliminary to the general engagement. It had an admirable effect on the science, widened men's views, proved that the weapon of the future was to be a carefully conducted experiment; and without doubt ripened the next great discovery which was then about to be announced.

John Dalton (1766-1844) was led to the happy thought of taking the data of the weights which Proust had announced as those in which tin, iron, oxygen, &c., combined, and reducing them to their simplest proportions. Proust had found that some arbitrary number of grms. of tin (say, for example, 119), combined either with 16 or with 32 grms. of oxygen, and with no other weights. Dalton showed that the weights of oxygen in these two compounds were to each other as 1 to 2.

In the same way the different weights of sulphur which entered into combination with a given weight of iron were to each other as I to 2. And he found that this held for all cases where two constituents combined with each other in more than one proportion.

Thus, if the amount of hydrogen in olefiant gas or ethylene and marsh gas or methane are compared, they are to each other as I to 2. By numerous examinations of this kind, in all of which he found this simple relation, he was led to formulate his atomic theory, some of the more important propositions of which may be thus condensed. (1) Every element consists of similar atoms of fixed weight; (2) Chemical combinations are made by the union of the atoms in the simplest proportions. The atomic weight of a compound is equal to the sum of the atomic weights of its constituents. He supposed all the atoms to be spherical and to be surrounded by heat spheres (!)

It should be mentioned in passing that Higgins had said in 1789 that chemical smallest particles were united to form compounds in simplest proportions, but as he never adduced any proof of this, the merit of the discovery belongs to Dalton by the law of possession already alluded to before, viz., that in natural science not only must a truth be announced, but some reason for it must be given.

The immediate result of his postulates was that Dalton set out to establish a scale of atomic weights for the elements. Among minor postulates of his which have not lived till our day, but which were natural enough at a time when there were no means of obtaining certainty as to the questions of the number of atoms entering into combination, &c., was this, that if only one proportion by weight of a combination between two elements were known, it must be supposed that the number of atoms entering into this combination was one from each element. If two were known, then that in which the least weight of one combined with the least weight of the other must be considered I to I; when with double this weight of the other the proportion must be 1 to 2, &c. In Dalton's time only one combination between oxygen and hydrogen was known, viz., water, and he assumed this to be composed of one atom of H to one atom of O. As H was

NEWS

and is yet the lightest element known, he assumed its weight as one. By the imperfect methods then available, he determined the weight of O which combined with it to form water as 6'5 (in reality it is 7.98 if H=1).

Ammonia, which was the only compound of H and N known to him, and in which he also assumed one atom of each element, gave him the number 5 for the atomic weight of N. By accurate methods it should be 4'66. In the lowest compound of carbon and oxygen known, carbonous oxide, he found the atomic weight of carbon 5'4 calling oxygen 6.5 (the right figure is 6).

All his figures were wrong as we now believe because of his false assumption of the constitution of water (not to speak of his imperfect methods of analysis), yet the accuracy which he attained was surprising for his epoch and the invigorating effect on the science was as great as if all his numbers had been absolutely correct.

Humphry Davy (1778–1829), the great discoverer of the alkaline metals and earths, who first announced the elemental character of chlorine, and by his discovery of the halogen acids seemed to have overthrown Lavoisier's dictum regarding the invariable presence of O in all acids -Davy, the discoverer of the safety-lamp for miners, first announced his belief that chemical affinity and electricity were the same force. This idea was erected by Berzelius later into the splendid structure which he called the electro-chemical theory. Neither Davy nor Wollaston believed that Dalton's experiments had succeeded in establishing the nature and characters of atoms, but contented themselves with Wollaston's theory of "equivalents," without seeking to define how much matter entered

into combination.

Their theory was that the atomic weights of Dalton were merely a series of arbitrary numbers, showing the respective quantities of different elements which were equivalent to each other in combining each with a third.

Wollaston's name of "equivalents" took root later after the apparent failure of Berzelius's theory to account for all the facts, and was the shibboleth of a long period of timidity and vacillation in chemical theory, which marked the reaction of thought when it was feared that the allurements of a beautiful system and the powerful influence of a great authority had drawn the representatives of the science away from sure ground. This period of intellectual cowardice was very tantalising and very confusing to those who pursued their studies during this period, but in the end it was an advantage to the science by letting the field lie fallow for a time, and making it thus the fitter to receive and develop the seed which finally was sown upon it.

In all cases where the development of a science has been rapid, it is found that the great minds are clustered together, and that the great discoveries occur in succession to supplement each other. It was stated that the discovery by Black originated pneumatic chemistry or the chemistry of the gases. In this field the discoveries of Cavendish, Priestley, and Scheele were made, but with the wider view given by Lavoisier to the science, the study of the gases was abandoned for the study of other solid and liquid compounds. But Gay-Lussac (1778— 1850) devoted himself to pneumatic chemistry and accomplished in it what supplemented the work of Dalton, and prepared the way and assisted the researches of Berzelius.

In 1805, in conjunction with Alex. von Humboldt, GayLussac established the fact that exactly two volumes of H combine with one volume of O to form water.

He showed the simple relations of the volumes of combining gases to each other and to their compound; he showed the effects of temperature on gases, and how it must be considered in connection with the Boyle-Mariotte law of pressure. His conclusion was that "The specific gravities of gases are proportional to their atomic weights, or are simple multiples of them."