The first addition causes the loss of three single bonds, and the conversion of two C.C links into two C:C links. The energy therefore required amounts to—

3×14'7+2(14'7-13.3)

46'9 cals.

NEWS

Thus it is clear that the evidence given by this abnormal addition beat does not justify the assumption of the nine bonds and the use (without more) of aliphatic thermical values for benzene.

Also, the other additions calculated in this way give differences with the values found from about 4 to 7 cals. So the influence of the aromatic character (closed ring structure) seems to be at least not inconsiderable for these cases.

Briefly resuming this discussion we write :

1. The deduction made on p. 155-156 (CHEMICAL NEWS, Vol. cvi.) has no argumentive value whatever for Kekulé's formula. It is, indeed, as Mr. Stanley Redgrove has observed, a vicious circle.

2. The combustion heat of benzene may be explained by Kekulé's formula when we ascribe to the influence of ring closing about

56.3-14'741.6 cals.

OH

OH OH ANN... B

y-Oxime.

In this method of formulation the two hydroxyl groups are nearer together in the y-compound than in the a-compound. Thus the former the more readily yields di-phenyl furazan.

Chemical Department, The University, Edgbaston.

VALENCY AND THE EVOLUTION OF THE ELEMENTS.

By M. D'A. ALBUQUERQUE.

FOR Some time I have regarded certain facts in chemistry, such as similarity, affinity, instability, &c., as depending upon the internal constitution of the atoms, undoubtedly bearing some relation to the genetic formation of these atoms. I could, however, find no explanation of these facts until I heard of the phenomena of vacuum tubes, of radiant matter, of cathodic projections, and finally of radio-active substances bringing about "in vitro" the remarkable phenomenon of the generation of new elements. From the consideration of these phenomena, and in particular of the fact of the similarity of the products obtained from a number of elements, I deduced the conception and the proof of evolution in chemistry, and of the origin of the atoms. Spectroscopic researches and researches with the rare earths have confirmed my views, and have shown that evolution is the clue to the chemistry of the future. All my recent work has been in this direction. I began to look upon chemical affinity as the proof and explanation of atomic evolution, and saw in the similarity of the elements, or rather in the greater or less distance between the reacting atoms of a common ancestral type, the true cause of affinity and of the phenomena allied to it.

From the theory of a common origin of all the elements and especially of those belonging to the same family, and from the consideration of the resemblances existing between different atoms having the same valency, I have been led to look upon the elements of a certain valency as in some way representing the prototypes which are common to them with other atoms (with identical properties and the same valency charge). That is to say, an element with a given valency is quite a different atom from the same element with a different valency. Thus I am led to believe that the atomic properties vary when one passes from one valency to a different valency (in the same atom) as they vary with the valency as one passes from one family or group to another. I think I can foresee that all the properties in any way connected with the existence of "actual" or potential electric charges will follow the same variation.

Basing my conclusions on the differences observed in the atomic magnetic susceptibility of the diado-ferric (Fe'') salts and the ferric (Fe"?) salts, diado-cobaltic (CO") and cobaltic salts, I regard the ferric and cobaltic salts as salts of a hexavalent dimetallic radical, a possible type of the formation of at least some atoms. (Perhaps these researches on polymetallic radicles will be the subject of a future note. I have just heard that the forms of elements entering into combination change with the valency charge, Pope, Journ. Chem. Soc., 1900, lxxix., 828).

CHEMICAL NEWS,

Oct. 25, 1912

In conclusion I wish to urge all who are interested in researches in physical chemistry to continue the investigation of elements like Sm, Ta, &c., which give salts of two different valencies, and to publish their results for the advancement of science and the establishment of a firm basis for the theories of evolution.

THE BROWNIAN MOVEMENT AND THE SIZE OF THE MOLECULES.*

By Prof. JEAN PERRIN, D.Sc., Faculté des Sciences à la Sorbonne, Paris. (Continued from p. 191).

Law of Avogadro.—I have just shown you how the study of chemical substitutions gives, for example, the ratio of the masses of the molecules of benzene and water, which ought to be to one another as 78: 18. Two masses of benzene and water which are in the ratio of 78: 18 each contain the same number of molecules. Now, if we measure the volumes occupied in the gaseous state by these two masses, at the same temperature and pressure, we find that these two volumes are equal. This is not chance, and one invariably finds that two gaseous masses which contain the same number of molecules occupy the same volume when they have the same temperature and pressure. This is Avogadro's law, which may be stated as follows:

"In the gaseous state equal numbers of molecules of different substances, contained in equal volumes at the same temperature produce in it the same pressures."

(NOTE.-Avogadro's law, once well established, will give us by extrapolation molecular weights not obtained by the methods of chemical substitutions. For example, when we have found that 32 parts of oxygen or 2 parts of hydrogen produce in the same volume and at the same temperature the same pressure as 18 parts of water vapour, we know that the masses of these three kinds of molecules are to one another as 32: 2: 18).

You have learnt elsewhere (Boyle's law) that the pressure thus developed varies inversely as the volume occupied.

These gas laws have been extended by van't Hoff to dilute solutions. We must, of course, in that case con sider, not the total pressure exerted on the walls, but only the part of that pressure which is due to the blows of the dissolved molecules, the part called the osmotic pressure of the dissolved substance (which can be measured only by means of a "semi-permeable" wall, which stops the molecules of the solute but not those of the solvent). The laws of Avogadro and Boyle thus become : "In the dilute state (gaseous or dissolved) equal numbers of any molecules whatever, contained in equal volumes at the same temperature, produce in them the same pressure. This pressure inversely as the volume occupied." These laws are applicable equally to all molecules, large or small. The heavy molecules of sugar or of sulphate of quinine produce neither greater nor less effect than the molecule of hydrogen. Yet the molecule of sugar contains 45 atoms, and that of sulphate of quinine more than 100, and it would be easy to find other more complex molecules which obey the laws of van't Hoff (or those of Raoult which follow from them).

varies

Is it not then conceivable that there is no limit of size for the assemblage of atoms which obeys these laws; is it not conceivable that even visible dust also obeys them exactly, so that a granule agitated by

⚫ A Discourse delivered before the Royal Institution, February 24, 1911.

203

the Brownian movement has neither more nor less effect than a molecule of hydrogen, as regards the action of its bombardments on a wall which stops it? Or, more briefly, is it unreasonable to think that the laws of perfect gases may be applied also to emulsions composed of visible grains?

I have made this assumption, and it is in this direction that I have sought a crucial experiment which would decide the origin of the Brownian movement, and at the same time would either provide or preclude an experimental basis for the molecular theories.

The following appears to me to be the simplest :-The Distribution of Equilibrium in a Vertical Column of Diluted Matter.-You know that the air is more rarefied on mountains than at sea-level, and that in a general way a column of gas is compressed under its own weight, the state of equilibrium resulting from the conflict between the force of gravity, which makes the molecules fall, and their movement which scatters them incessantly. The law of rarefaction formulated by Laplace (to show how the altitude can be deduced from the barometer) follows of necessity from Boyle's law, and can be enunciated as follows.

Every time that one ascends a fixed distance, the density is divided by the same number. Or, more briefly, equal vertical elevations are accompanied by equal rarefactions.

For example, in air at the ordinary temperature the density diminishes by one-half each time one ascends 6 kilometres (it is immaterial from what level).

But the elevation which produces a double rarefaction would not be the same in hydrogen. A simple process of reasoning shows that the way in which the nature of the gas influences the result is a necessary consequence of Avogadro's law, and may be enunciated as follows: The elevations which produce the same rarefaction for two different gases (at the same temperature) are inversely proportional to the weights of the molecules of these two gases.

For instance, if in oxygen at o° it is necessary to rise 5 km. to make the density twice as small, it would be necessary to rise 80 km. in hydrogen at o°, since the molecule of hydrogen is 16 times as light as the molecule of oxygen.

You see here (Fig. 1) a diagram showing three huge vertical cylinders (the largest is 300 km. high) into which the same number of molecules of hydrogen, helium, and oxygen have been placed. At a supposed uniform temperature, the molecules would distribute themselves as the diagram shows, collecting more towards the bottom the heavier they are.

Now we have been led to believe that the laws of perfect gases are possibly applicable to emulsions. If this is so, and if we make an emulsion in which the grains are equal, the distribution of matter in a vertical column of this emulsion ought to be the same as in a gas. In other words, once arrive at the distribution of equilibrium and then equal elevations will be accompanied by equal rarefactions. But if it is necessary to rise only 1/20 mm. i.e., 100 million times less than in oxygen, for the concentration to become twice as small, then we must conclude that each grain of the emulsion weighs 100 million times more than one molecule of oxygen. This last weight could be ascertained if we could weigh the grain, which would be a stage between the molecular dimensions and those which are in our scale.

Of course the effectual weight of this grain would be the difference between its real weight and the thrust it would undergo in the liquid (according to Archimedes' principle). If the granules were lighter than the intergranular liquid, they would accumulate in the upper layers (equal depres sions would produce equal rarefactions). They would distribute themselves uniformly if they had the same density as the liquid.

Practical Realisation.—To test these results I used the

emulsions which are obtained by precipitating alcoholic, solutions of resins with water. Thus, as you see, we get with gamboge a beautiful yellow liquid, with mastic a white liquid like milk. The microscope reveals in these liquids the resin precipitated in the form of solid round granules, which do not agglutinate when the chances of the Brownian movement bring them into contact (which is what happens with other resins which give soft granules).

But the diameters of these granules are very varied, and I had to sort them so as to get emulsions in which all the grains were nearly the same size. The method I employed may be compared to the fractionation of a liquid mixture by distillation. Just as during distillation the parts first vaporised are relatively richer in volatile constituents, so during the centrifugation of an emulsion the portions first deposited are relatively richer in large granules. Thus you will see that it is easy to find a practical method of sorting the grains according to their size by fractional centrifugation. The operation is long (I spent some months over it), but it only requires patience).

Once get an emulsion with sufficiently uniform granules, and then the mean weight of the granules must be determined. Their density is measured like that of any other powder (the weight of resin in suspension in the flask used is determined by simply drying). The only difficulty is then to determine the diameter. The obvious way seems to be to measure it with a microscope by the clear chamber method by means of a micrometer objective. But the granules used are so small that errors of 20 per cent or more might easily be made. A method which is almost as direct consists in allowing a droplet of the emulsion to evaporate on the micrometer objective; it is found, as you can see here in a projection (Fig. 2), that the granules then arrange themselves in regular lines, the length of which could be measured with a fair degree of accuracy. Dividing this length by the number of granules gives the diameter.

Another longer but more accurate method consists in counting how many granules there are in a known volume of the standardised emulsion, which gives the mass of a granule, and hence its radius, since we know the density. | For this method I used the fact, accidentally observed, that in a feebly acid medium the grains of gamboge adhere to the glass. At an appreciable distance from the walls the Brownian movement is not modified, but as soon as the chances of this movement bring a grain into contact with a wall the grain becomes motionless. The emulsion thus gets progressively weaker, and after some hours all the granules it contained are fixed. One can then count at one's leisure all those which come from a cylinder of arbitrary base (measured by the clear chamber).

Finally, a third method which I shall not explain to you in detail is based on the observation of the time necessary for the upper portion of a vertical column of emulsion (several centimetres high) to clarify to a given height. This time, required by the granules to descend on the average from this height, gives the diameter by applying the law of the fall of a sphere in a viscous fluid (Stokes).

These three processes agree, and this must be so to give us confidence in the accuracy of the measurements of a radius which is less than a thousandth of a millimetre.

Method of Observation.-We must now get an arrange ment which will enable us to ascertain the distribution of equilibrium as a function of the height. For this purpose a drop of emulsion is placed in a shallow tank, the depth of which is 1/10 mm.; the drop is at once flattened out by a cover-slip which closes the tank and the edges of which are paraffined to prevent evaporation (see Fig. 3).

As the figure shows, the tank can be arranged vertically, the body of the microscope being horizontal, and thus the vertical column throughout which the emulsion can distribute itself may be made some millimetres high. It will be seen that the grains accumulate in the lower layers, and tend to give rise to a distribution of equilibrium (practically reached after one or two days), in which the progressive rarefaction as a function of the height is manifested as

you can see in the projection, which plainly recalls the law of rarefaction of heavy gases.

But however small our granules are, they are so heavy that the rarefaction is very rapid, and the height over which measurements can usefully be made is less than 1/10 mm.

As the figure also shows, the tank can be made horizontal, the body of the microscope being vertical, in which case only about a quarter of an hour is necessary for equilibrium to be established. The objective of the microscope, of great enlarging power, has only a small depth of field, and we can only see clearly at the same instant the granules situated in a very thin horizontal layer, the thickness of which is only a little over 1/1000 mm. If the microscope is raised or lowered one sees the granules of another layer. The distance between these two layers may be deduced from the vertical displacement read on the screw of the microscope; the ratio of the number of granules perceived gives the rarefaction corre sponding to a known elevation. It was with this apparatus that I made my first experiments.

I was not sure that there would be the least rarefaction, and, moreover, not even sure whether, on the contrary, all the granules would not group themselves quite against the bottom. But I saw that a permanent state of uniform rarefaction was established. This rarefaction is specially striking when, keeping the eye fixed on the preparation, the observer rapidly raises the microscope by means of its micrometer screw. The granules are then seen to be rapidly rarefied, like the atmosphere round an aerostat which is rising. When

It now remains to make accurate measurements. one sees in the field some hundreds of granules which are moving in all directions or which disappear while new ones make their appearance, one soon gives up trying to count them. Luckily we can take instantaneous photographs of the different layers, and then at our leisure count the number of granules in these layers on the plates. It is thus easy to verify that equal elevations are accompanied by equal rarefactions. For instance, for granules of radius equal to o°212 μ three successive rises of 30 μ practically lower the concentration to one-half, onequarter, and one-eighth of its value. With other grains of radius equal to o 367 u an elevation of 6 μ is enough to make the density about twice as small. You see here a drawing obtained by placing one above the other five sections made at distances of 6 μ in this emulsion (Fig. 4).

To obtain the same rarefaction in air we have seen that it would be necessary to rise 6 km., a thousand million times as much. If our theory is correct the weight of a molecule of air would be the thousand-millionth part of the weight of one of our granules in water. The weight of the atom of hydrogen would be obtained in the same way, and now our interest centres on finding out if we shall thus obtain the same numbers as those given by the kinetic theory,

So I was much elated when I found at the first attempt numbers which were the same as those obtained by the kinetic theory; i.e., by a fundamentally different method. I also varied the conditions of the experiment as much as I could. For instance, the mass of my granules had a series of values ranging between limits which were to one another as 1 : 40; I changed the nature of the granules, using different resins (especially mastic); by the addition of glycerin I increased the viscosity of the intergranular liquid in the ratio of 120: 1, at the same time changing the nature of the liquid; finally, I made a considerable change in the apparent density of the granules, which in water varied from the same to five times as much, and which became negative for gamboge in glycerin with 10 per cent of water (in this last case the granules being lighter than the liquid accumulated in the upper layers).

(NOTE.-Quite recently under my direction M. Bruhat made the temperature vary from -10° to + 58°, and still found the same weight).

I always obtained concordant results, giving for the

atom of hydrogen a weight very nearly equal to the value 1.6 X 10-24, given by the kinetic theory.

I do not think that this agreement can leave any doubt as to the origin of the Brownian movement. To understand how striking it is it must be remembered that before the experiment one could not have dared to affirm that the fall of concentration would not be negligible for such a small height as a few microns, or that all the granules would not collect in the immediate neighbourhood of the bottom of the tank. The first eventuality would give a zero value, and the second an infinite value for the weight of the hydrogen atom. That one should have hit exactly upon a value so close to that foretold, with each emulsion, in the enormous interval which seems à priori possible, could obviously not be a chance coincidence.

But there is more to be said. While the kinetic theory, because of the simplifications permissible in its calculations, gives results of an uncertain degree of approximation, even from perfect experiments, the numbers given by emulsions correspond to a true measure, such that there is no limit to its accuracy. By this method we can really weigh the atoms, and not only roughly estimate their weight.

The two series of experiments which I regard as the most accurate have thus given me for the weight of the atom of hydrogen (after the enumeration of about 30,000 granules) the value

I'47 1,000,000,000,000,000,000,000,000'

IN sealing any part of a mine certain precautions are observed, for it is a matter of common knowledge that a fire continues to burn for some time after stoppings are built, and the heated gases, unless they can escape readily, exert considerable pressure on stoppings. Moreover, an inrush of air after the cooling of the imprisoned atmo sphere must likewise be guarded against or the fire may start afresh; again, dangerous explosione may result if the outflow of gases from a fire is checked too suddenly. At some fires the precaution has been adopted of having a pipe, with an elbow, placed in the upper part of the dam. This pipe dips into a vessel, an open barrel or keg, continuously supplied with water from a pipe fitted with a tap. The overflow from the barrel is allowed to run off. The tap is for shutting off the water when no longer needed. By this arrangement the outflow of gas is checked gradually, and is finally stopped when the pressure of the water over the outlet of the escape pipe equals that of the gas behind the stopping. The relief pipe also serves to indicate the pressure in the fire area. When the temperature rises the pressure is outward, and more gas escapes from the pipe; when the temperature falls the pressure is inward, and water is drawn into the pipe. After all combustion has ceased and the temperature has become normal, the atmosphere in the enclosed area may alternately expand and

contract.

• Technical Paper 13, Bureau of Mines. Washington.

COLLECTION OF SAMPLES-SIGNIFICANCE OF Data.

temperature and pressure of the imprisoned atmosphere a For the purpose of taking samples and determining the straight pipe provided with a valve should be placed in each stopping. Samples of gas can be collected with a small hand-pump, pressures can be read with a waterinserted through the pipe. gauge, and temperatures can be taken with a thermometer

Data obtained thus at regular intervals during the entire period that the fire is sealed are valuable, but their significance is governed more or less by the distance of the stopping from the fire. If this distance be great, changes in the condition of the fire will not be quickly indicated by the atmosphere behind the dam, and a change for the it. Clearly, however, in most cases the fire can burn more worse may occur before the fire fighters become aware of vigorously only by inleakage of air through the stoppings, consequently chemical analyses of the atmosphere just behind them show whether the inleakage of air is sufficient ditions to become worse. to keep the fire burning, and thus, perhaps, permit con

COMPOSITION OF THE ATMOSPHERE IN BURNING MINES.

Samples of gas were collected from behind stoppings that were built to seal off a fire in and close to the mouth of a drift mine working the Pittsburg bed. The first sample was taken through a hole in a concrete stopping, very near the fire, one day after the mine was sealed. The second and third samples were collected at the same place. Shortly after the third sample was taken water which had been forced in rose so high inside the dam that more samples could not be collected (see Table A).

The samples in Table B were collected from the same mine by inserting a tube in a small hole driven through the thin covering over the main heading near the mine mouth. The samples thus collected represent gases from a very hot part of the burning section.

These analyses are interesting because they show a rapid depletion of oxygen after sealing, and the formation of an atmosphere that would check the progress of the fire. In other words, they show that the dams were tight and the fire was being brought under control.

Samples were taken from behind another stopping, which was situated about 2000 feet away from the burning area, by boring a 4-inch hole in the wooden brattice and drawing the gas into the sample container with a small air-pump. Later, by using helmets, entrance was made into the mine at this place, and three samples of air were obtained at points approximately 200, 600, and 800 feet inward from the stopping. Samples could not be collected farther in than 800 feet because of the heavy pall of smoke encountered there. The analyses of the samples are given in Table C.

The deficiency of oxygen in the samples was in part due to the absorption of oxygen by coal not affected by the fire. Except for the stagnant condition of the atmosphere and the heavy accumulation of black damp in the passages between this section and the burning section, the high oxygen content of the samples, especially of Nos. 1 and 2, would have caused apprehension as being sufficient to increase combustion. Because of the stagnation of the air, however, it was felt that little oxygen from outside was reaching the fire. Air was apparently leaking in to some extent at the place where these samples were taken, or a greater oxygen deficiency would undoubtedly have been found. On account of many small openings at different places on the surface, air could not be entirely excluded.

The mine was non-gaseous, and this fact accounts for the small amount of methane that accumulated directly at the fire area and at the brattice 2000 feet beyond. The knowledge that the accumulation of methane was slight was valuable, as it gave assurance that an explosion would not follow an accidental inrush of air.

The samples of gas given in Table D were obtained