NEWS

Name.

Two vegetables on this list are not as well known as they ought to be, the names are even unknown by greengrocers doing a fairly good business; salsify is one of them. The roots can be left in the ground until late in the season, and it comes into use as a variety in the winter. It is often called the oyster plant, because the flavour suggests that of oysters. The other vegetable is celeriac, which is also known as "knot celery" and "turnip rooted celery"; the roots only are eaten. It also comes into the market during the winter. Its flavour is something like celery.

In another class of vegetable foods the change in the percentage of water is still more marked; that is the pulses. In dealing with these foods one of the most important factors to consider is the amount of protein present. We often see the statement that the pulses are rich in this nutrient, and for this reason they are called "poor man's beef." This is true only in the uncooked condition, but not as we eat them. This idea arises from the incorrect method of describing the results of analysis. In most tables the percentage of water is left out, and only those of fat, fibre, protein, and carbohydrates are given, thus making these percentages appear higher than they really are; also the comparison is between meat and pulses in the uncooked state. Here we find for protein, beef 22, veal 20, mutton 20, lentils 22, peas (dried) 21. Now, in the process of cooking, meat has lost water; therefore the percentages of fat and protein are higher in the cooked condition, while with the pulses the reverse is the case, so that all the percentages of the nutrients are lower :-Beef (boiled), 34; veal (roasted), 29; lentils, 9; peas, 9; that is, the percentage of protein in the natural moist condition as served at table.

With cereals, such as oatmeal, we have no loss of nutrients; all the water used in cooking is absorbed. Starting with 10 lbs. of each of the following articles we find provost barley and oats 60, frame-food 156, plasmon arrowroot 113, rice (prepared by a Conservative method) 36 lbs.

These, then, are the chief changes in cooking vegetable foods, and it cannot be satisfactory that so much of the valuable nutrients should be lost by bad methods of pre paration. It may be said vegetables are cheap, but the great object of cooking is to make food attractive and wholesome, and it is the salts which give the flavour. It is a pity the subject of the scientific value of food is not studied in this country in the same way that it is in America. Much can be learnt from the publications of the Department of Agriculture in that country,

INFLUENCE OF COMPOSITION UPON
THE CORROSION OF STEEL."

By LESLIE AITCHISON, M.Met.
(Concluded from p. 137).

Manner of Experiments and Results Obtained (continued). THE structurally free carbides in the series show the greatest possible regularity. Throughout the list of steels under review each carbide acts in the same way, being entirely unattacked. Whenever these carbides stand by themselves, whether as crystal boundaries and globules as the result of segregation or in relatively large masses, the result is the same-they remain quite unattacked. This may be seen very well in Figs. 8 and 9 (supra) as illustrating the case of crystal boundaries, in Fig. 10 (No. 37, having 12:45 per cent of vanadium), in Fig. 4 (supra) for the isolated globules, and in Fig. 11 (No. 46, having 5'37 the segregated carbides is seen very well in the nickel The case of per cent of tungsten) for the large masses. steels, particularly Nos. 55 and 56, for there the pearlite and ferrite are corroded distinctly, but where there has been a separation of the pearlite into ferrite and cementite the latter is quite unattacked. Also in Fig. 12 (No. 33, with 0.6 per cent carbon and 0.71 per cent vanadium) a certain amount of segregation has taken place, the result being that the carbide boundaries stand out alone and un

attacked.

The solid solution which occurs in most of the alloy steels is also quite uniform in its behaviour. In general it is attacked fairly evenly, leaving the free carbide which accompanies it, and which may be safely assumed to act as its electrical cathode, standing up quite alone and unattacked. The even attack is seen very well in Fig. 13 (No. 47, having o'74 per cent of carbon and 14'96 per cent of tungsten), and the more uneven, due to the extra carbide dotted about in the solid solution, in Fig. 14 (No. 44, having 0.73 per cent of carbon and 21.5 per cent the steels containing solid solution and free carbides. of tungsten). These two structures are quite typical of above is seen in several of the samples. The effect of the The influence of the inter-granular material referred to corrosion upon it is evident in Fig. 4. There may be noticed here relatively broad black lines surrounding the grains. These are not optical effects at all, and in the actual specimen show up quite plainly and quite definitely. It seems fairly certain that this layer of material has been attacked preferentially at the commencement of the corrosion, or perhaps has been the centre of the action between the two adjacent crystals. Whatever the action may have been it has resulted in the production of a narrow pit, the sides of which have been exposed. These sides are relatively rough, and consequently offer a fairly large surface to the attack of the solution. In conse. quence there has been a relatively intense action in these regions. This would account for the slightly greater attack on the bar iron over that sustained by the steels with a higher carbon content. The inter-granular layer naturally does not appear in those steels having carbide boundaries to the crystals, but does appear in some of those steels having solid solution with very little free carbide, and that as globules scattered about within the grains (cf. No. 45, with 0.70 per cent of carbon and 9'74 per cent of tungsten, also in several of the chromium steels). The steel examined for the action of twinning was corroded in the same manner as the other steels, but it was found that the corrosion had gone on to a very great extent, the structure being entirely obliterated. (This is not due to the twinning alone, as a nickel steel possessing an almost identical structure did not corrode at all). By corroding for a shorter period (three days) the structure shown in Fig. 15 was obtained. This shows that the parts of every steel orientated in different ways, corroded

* A Contribution to a General Discussion on "The Corrosion o Metals" before the Faraday Society, December 8, 1915.

solution-which is to be found as a consequence of the chemical composition only, coupled with the thermal and mechanical treatment. Apparently iron itself is not capable of offering real resistance to corrosion because its solution pressure is never sufficiently low. It is hoped that the results of work in support of this view may be communicated shortly.

4. That carbides in steels act in two directions-(a) as deterrents, in that they resist corrosion in themselves, being quite unattacked by the majority of reagents; (b) as aids in most cases, by providing a suitable cathode to act as counter to the anodic action of the ferrite in solid solution. In this latter way the presence of pearlite, which of course contains the carbide in considerable quantities, tends to increase the corrodibility of the steels. Although the pearlite appears in most cases to be attacked, it is most probable, in view of the persistence of the free carbides, that the attack is confined to the ferrite or solid solution. This latter of course constitutes the matrix in which the small rods or lenticles of carbide are embedded, and the disappearance of this support results in the mechanical removal of the carbide and not in the solution or decomposition of it. This is supported by the fact that in every case of a metal containing any quantity of carbide (either

MOST of the separations made in analytical chemistry depend on pronounced differences in solubility and the use of an excess of reagent, but in geochemistry we have to consider associated substances whose solubility may differ but slightly and whose formation was due to fractional precipitation. A fractional precipitation is one in which only a part of the dissolved substances pass from solution to solid.

The theoretic side of fractional precipitation was treated extensively by Berthollet over a century ago. As regards fractional precipitation Berthollet was correct in contending that the composition of the precipitate might be indefinite, and Prout's view as to the definite composition of the precipitate was correct in regard to pure simple

ree or in the pearlite) the sample after corrosion was covered with a loose layer of carbide that had to be removed before weighing. In the case of those pieces of carbide that are relatively large it is quite conceivable that they are rooted so deeply in the ferrite or solid solution that the corrosion has not proceeded to the depth required to cause them to be loosened sufficiently to be removed by mechanical means. Similarly of course the boundaries of the grains are firmly fixed in the material and are not likely to be removed.

Action of Mercury Salts on Aluminium Foil. St. Minovici and Em. Grozea.-When some drops of a solution of corrosive sublimate are placed on aluminium foil and allowed to dry they leave an efflorescent residue like that of calcium nitrate which forms on damp walls. This reaction occurs very rapidly, and is very sensitive. No other salt prevents its occurrence, but it does not take place if the aluminium foil with the mercury salt is heated in a flame, nor is it produced in vacuo. All salts of mercury give it, and it appears probable that it might be used in general analysis and in toxicological researches.Bulletin de la Section Scientifique de l'Académie Roumaine, [6], iv., 227.

FIG. 15.-C 12 per cent, Mn 12.0 per cent. chemical compounds. Berthollet held that the fundamental factors in chemical reactions are cohesion, elasticity, and mass relations, governing respectively solubility, volatility, and precipitation.

Fractional precipitation of the rare earths has long been familiar to chemists. In this way the constituents of the gadolinite earths were separated by Mosander ("On Yttria, Terbium, and Erbium," Phil. Mag., 1843, 3rd ser., xxiii., 252), praseodymium and neodymium were separated from the cerite earths by Welsbach, the constituents of the yttria earths were separated by Crookes ("On the Methods of Chemical Fractionation," CHEMICAL NEWS, 1886, liv., 131, 155), and the sulphide of polonium by Curie (" Radio-active Substances," CHEMICAL NEWS, 1903, lxxxviii., 146). Debus, in 1853, made quantitative experiments on the fractional precipitations of barium and calcium carbonate ("Ueber Chemische Verwandtschaft," Liebig's Ann., 1853, lxxxv., 124), and Chizynski later studied the phosphates of these two metals ("Ueber die Chemische Massenwirkung," Ann. Chem. Pharm., 1865, Suppl. Bd. 4, P. 226). Chroustchoff and Martinoff studied the precipitation of sulphates and chromates by a barium salt

Bulletin 609, United States Geological Survey.

"Des Coefficients d'Affinité Chimique," Comptes Rendus, 1887, civ., 571), and Küster and Thiel the fractionation of bromides and thiocyanates by silver nitrate ("Ueber Gleichgewichtserscheinungen bei Fällungsreaktionen," Zeit. Anorg. Chem., 1902, xxxiii., 129). Recently Goldblum and Stoffella have studied the system involving lead carbonate and chromate (" Contribution à l'étude de l'Affinité Chimique," Journ. Chim. Phys., 1910, viii., 135).

The advantage of studying fractional precipitation for a theory of ore deposits is that the results show the net effect of all the factors that enter into play, including even unsuspected factors. Thus, if the salt of one metal is more hydrolysed in solution than another, or if a metal has a tendency to form basic salts, the effect will be mani

fest in the fractionation.

An objection that may be made to many laboratory experiments is that chemical precipitates are formed instead of natural minerals. For some purposes this

objection is valid; for others not. In general, such experiments will shed light on the relations to be expected in the formation of minerals. In fact, in many experithe precipitate alters even at ordinary temperature into the mineral, as does lead sulphide (O. Weigel. "Die Löslichkeit von Schwermetallsulfiden in Reinem Wasser," Zeit. Phys. Chim., 1907, lviii., 293). In other experiments the

ments it has been shown that in the course of a few hours

rate of alteration is very slow, as Allen and Crenshaw have shown for zinc sulphide ("The Sulphides of Zinc, Cadmium, and Mercury," Am. Journ. Sci., 1912, 4th ser., xxxiv., 358), and solubility determinations are needed for the minerals as well as for the amorphous precipitates. I have, however, found in every experiment in which the effect of time has been noted, that when equivalent quantities of two metallic salts are taken the experiment shows almost at once which of the two metals forms the more insoluble compound. There may be a decrease in solubility with time or a relative change in the solubilities on account of the very gradual development of the most stable compounds, but in no experiment did the metal first precipitated in excess later become the more soluble.

Method of Experimentation.

The metallic compounds will be discussed in the order of their solubility-sulphides, hydroxides and oxides, carbonates, bicarbonates, and silicates.

The procedure in the experiments on fractional preA dilute solution containing two cipitation was simple. metallic salts in equivalent quantities was precipitated by reagent enough for one metal only. After a time an aliquot part of the mother-liquor was analysed, and the composition of the precipitate was determined by diffference. In some experiments the proportions were varied slightly, as noted.

Table for Computing "Equivalents." For weight given of Divide by For weight given of Mg..

Mn"

Na..

Divide by

72.16

27'46

23'00

29'34 103'59

Ag ΑΙ

107.87

9'033

Pb

thus calculated correspond to the usual chemical equi. valents of salt radicles in aqueous solutions referred to eight parts of oxygen. Of course, equivalent weights of salts may be taken in mgrms., grms., pounds, or tons, but wherever no qualification is expressed grm. equivalents are always understood. SULPHIDES.

Solubility of Sulphides.

of solubility of the compounds studied, it will be adAs these experiments were made to determine the order vantageous to present first such data as already exist on the subject.

The sulphides of the heavy metals are very insoluble. Although they are ordinarily considered quite insoluble their solubilities can be determinated approximately by physico-chemical methods, and, in fact, it has been shown that mercuric sulphide is many thousandfold more insoluble than manganous sulphide. The solubility of the sulphides in water at 18° C. was determined by Weigel Schwermetallsulfiden in Reinem Wasser," Zeit. Phys. by the conductivity method ("Die Löslichkeit von Chem., 1907, lviii., 293), but the determination of the solubility by this method is complicated by the fact that the exact nature of the dissolved substances must be known, and as the sulphides produce solutions differing in alkalinity the method could hardly be expected to yield more than mere approximations to the true solubility.

The solubility products of sulphides have been determined by other investigators from time to time.

(I. Bernfeld, "Studien über Schwefelmetallelektroden." Zeit. Phys. Chem., 1898, xxv., 46; Joseph Knox, "A Study of the Sulphur Anion and of Complex Sulphur Anions, Faraday Soc. Trans., 1908, iv., 48; C. Immerwahr, "Beiträge zur Kenntnis der Löslichkeit von Schwermetallniederschlägen auf Elektrochemischen Wege," Zeit. Elektrochem.. 1901, vii., 478; S. Glixelli, "Zur Theorie der H2S Fällung der Metalle-Die Einwirkungen von Schwefelwasserstoff auf Zinksalze," Zeit. Anorg. Chem., 1907, lv., 306; R. Lucas, "Gleichgewichte Zwischen Silbersalzen." Zeit. Anorg. Chem., 1904, xli., 211; L. Bruner and J. Zawadzki, “Ueber die Gleichgewichte bei der Schwefelwasserstofffällung der Metalle," Zeit. Anorg. Chem., 1910, lxv., 136).

For a univalent, bivalent, and trivalent metal, respectively, the solubility products (S. P.) would be:S.P. = [M+]2 [S--]

The values of the solubility products already published were collected by Bruner and Zawadzki, who also calculated them from the heat of formation of the sulphide,

and the electrolytic potential of the metals and of sulphur. In the table below I have placed their calculated values of the solubility products, those found by the observer noted, and finally the "solubility of the sulphide in water" calculated from the solubility product. It will be noted that the solubilities in this table are far smaller than those obtained by Weigel.

Solubility Products of several Sulphides and their
Solubility in Water.

(Based on determined solubility product, if known; otherwise on solubility product calculated by Bruner and Zawadski).

Solubility in

water. Mols. per litre (calculated). 2.6 × 10-8 2.2 X 10-8 I'9 X 10-II

2.2 X 10-12

I'2 X 10-13

1.8 X 10-14

4'2 X 10-15 2.3 × 10-18 19 X 10-17

Experiments with the Nitrates of Copper and Lead. The first experiments on this point showed that the composition of the precipitate depends a great deal on the ture of the nitrates of lead and copper, partly precipitated manner of precipitation and time of standing. A mixwith a solution of hydrogen sulphide in the cold and filtered within a short time, yielded a precipitate containing the sulphides of both metals. although copper was slightly in excess. On warming, however, copper sulphide was the chief product, being precipitated even at the expense of the lead sulphide first formed.

(To be continued).

THE REFORM OF THE MAN OF SCIENCE.

2.2 X 10-17 2'4 X 10-21 I'IXIO 21

1'7 x 10-27 2'4 X 10-24

The order of solubility obtained from experiments on fractional precipitation should agree with the order of solubility given in the last column of the table. As a matter of fact, the agreement is excellent except for two metals, silver and cobalt. The discrepancy for silver requires explanation; that for cobalt is not surprising in view of the small difference between its solubility and those of the adjacent sulphides. The discrepancy for silver may be due to a failure of the principle of the solubility product, as silver sulphide is a unibivalent compound, or possibly to the precipitation of silver by sulphides in part as free metal. A study of the fractional precipitation of thallous sulphide would assist in deciding between these two possibilities. Weigel's series seems to place silver in its proper position, but it shows several discrepancies for other metals.

While the solubility of sulphides is under consideration attention may be paid briefly to the enormous effect of changes in acidity on the precipitation of sulphides. This effect is due to the nature of the ionisation of hydrogen sulphide, the sulphide ion concentration varying inversely as the square of the hydrogen ion concentration for a given concentration of total sulphide. As is well known, the precipitability of the sulphides varies so greatly that a complete separation of certain groups is possible. Thus in an acid solution hydrogen sulphide added to a dilute mixture of the sulphates of iron and copper precipitates almost wholly copper sulphide. This is not strictly a fractional precipitation, since in an acid solution of hydrogen sulphide the concentration of sulphide is probably not sufficient to exceed the solubility product of ferrous sulphide. A fractional precipitation would be obtained with a small amount of precipitant under conditions which would permit all the metals to be pre cipitated with an excess. With iron and copper sulphates, for example, a fractionation might be made with sodium sulphide, since an excess of sodium sulphide would presipitate both sulphides completely.

It is not proposed here to enter upon a discussion of the legitimate use of the term science. We may be all for brotherhood, but the circumstances of life compel us largely to separate into groups for purposes of action, and there can be no real complaint if the word science is used in a restricted sense for what is perhaps better called natural science. This should not prevent men of science from recognising their kinship with all faithful workers for the elucidation of truth in whatever sphere of action. Let us avoid a controversy about mere words. Lieut.Colonel Barret's complaint is a more substantial one-not one of terminology. It is essentially this, that when operations relating to the forces of nature transcend a certain scale they are no longer recognised as science, and that men of science in the limited sense thus lose a great companionship and an invaluable link with the greater world. He gives as an illustration the work of a railroad president whose operations "involve the placing of towns and even cities in new positions, the reorganisation of the agricultural education of districts, the estimation of future markets, and other complicated actions involving scientific imagination of the first order."

was

It is probable that most men of science would readily admit that some solid advantages would be gained by having in their camp these great operators, with all their intellectual energy, their enterprise, and their influence, and perhaps many would admit their claim to inclusion. There is undoubtedly a tendency for an increased scale of operations to remove a man from the scientific class if he once in it, or to prevent his accession if he did not originally enter through the usual portal. The case may be well illustrated from engineering. A scientifically trained engineer who betakes himself to great problems of engineering, constructing some almost impossible railway or irrigating a whole parched province of India, seems to be moving away from science. An engineer who has acquired such powers without having received the hall. mark of formal scientific training will find it hard to get his place acknowledged in the ranks of science.

We may ask, What is really at the bottom of this? Is it merely narrow-mindedness, or is there something more