498 



NATURE 



{April t^, 1877 



is met with in the central portion of thick bars of Swedish 

 iron, kept for some weeks at a temperature below their 

 melting point, but high enough to give rise to recrystal- 

 lisation. We then get a complete separation of free iron 

 from a compound containing some carbon, and a crystal- 

 line structure which, as far as mere form is concerned, 

 . most closely corresponds with that of meteoric iron, as 

 may be at once seen on comparing them. 



These facts clearly indicate that the Widmanstatt's 

 figuring is the result of such a complete separation of the 

 constituents and perfect crystallisation as can occur only 

 when the process takes place slowly and gradually. They 

 appear to me to show that meteoric iron was kept for a 

 long time at a heat just below the point of fusion, and that 

 we should be by no means justified in concluding that 

 it was not previously melted. Similar principles are 

 applicable in the case of the iron masses found in Disco, 

 and it by no means follows that they are meteoric be- 

 cause they show the Widmanstatt's figuring. Difference 

 in the rate of cooling would serve very well to explain the 

 difference in the structure of some meteoric iron, which do 

 not differ in chemical composition ; but, as far as the 

 general structure is concerned, I think that we are quite at 

 liberty to conclude that all may have been melted, if this 

 will better explain other phenomena. On this supposi- 

 tion we may account for the separation of the iron from 

 the stony meteorites, since under conditions which brought 

 into play only a moderate gravitative force, the melted 

 iron would subside through the melted stone, as happens 

 in our furnaces ; whilst at the same time, as shown in my 

 paper read at the meeting of the British Association in 

 1864, where the separating force of gravitation was small, 

 they might remain mixed together, as in the Pallas iron, 

 and others of that type. 



In conclusion I would say that though from want of 

 adequate material for investigation I feel that what I 

 have so far done is very incomplete, yet I think that the 

 facts I have described will, at all events, serve to prove 

 that the method of study employed cannot fail to yield 

 most valuable' results, and to throw much light on many 

 problems of great interest and importance in several dif- 

 ferent branches of science. 



MENDELEEF'S RESEARCHES 

 MARIOTTE'S LAW^ 



ON 



T^ROM researches on the depression of the mercury results the 

 possibility of introducing a precise correction relative to the 

 volume of gas contained between the surface of the mercury 

 and the horizontal plane which touches the summit of the 

 meniscus. In all my researches I introduce each time a cor- 

 rection relative to this volume. 



The volume of the reservoir which contains the mercury and 

 the gas under various pressures undergoes two kinds of varia- 

 tions ; first, those which are due to the difference between the 

 pressures which act on the two sides of the vessel, and second, 

 those which depend on differences in the volume of mercury. The 

 compressibility of the reservoirs employed in the researches has 

 been always determined by experiment, and their change of 

 volume produced by the introduction of mercury can be de- 

 termined by surrounding the vessel filled with mercury by 

 another filled with the same material. When the height in the 

 two vessels is the same, the capacity of the vessel is that which 

 exists at the time of equality of pressure on the external and 

 internal surfaces of the vessel. If we empty a part of the 

 external vessel the capacity of the vessel changes in the same 

 manner as when we fill or when we empty the vessel. Experi- 

 ments of this kind have shown the possibility of determining the 

 changes of capacity depending on the quantity of mercury. The 

 relative corrections have in each case been introduced into the 

 calculations. 



All the practical side of the subject — the desiccation of the 

 gas, the complete abstraction of the remains of the gas from the 

 apparatus, the hermetical junction of the parts of the apparatus 

 by means of mastic and mercury stop-valves, the means of main- 



' Continued from p. 457. 



taining the gases and the mercury at a constant temperature, the 

 calibration of the tubes, and a number of other details have had to 

 be elaborated more or less anew. All this will be found 

 described in my work " On the Elasticity of Gases." I have 

 published this work only in Russian, not having means suffi- 

 cient to publish a translation of a work so voluminous, and 

 desiring to conform to the custom existing among savants of all 

 countries of describing their labours in their mother-tongue, in 

 order to present to the scientific literature of the country where 

 they live and work a gift in proportion to their powers. 



My desire was to investigate the subject in its minutest details in 

 order to eliminate every possibility of doubt as to the causes which 

 determine the deviations observed from the Boyle-Mariotte Law. 

 I know that that law is firmly established, and I believe it will 

 remain so. Not less great is the certainty in the mind that rare- 

 fied gases approach the perfect state. That certainty I had also 

 on commencing my experiments. It was necessary then to deter- 

 mine as completely as possible all the circumstances on which 

 depend the facts contrary to the opinion generally held. This is 

 why I have modified the apparatus, improved the methods, and 

 employed in this work more than three years without inter- 

 ruption. Now so far as regards low pressures the work is 

 finished, and I have obtained definitely certain proofs of the 

 rigorous accuracy of my first observations. 



The experiments which I have made with Kirpitchoff have 

 proved that not only for air, but also for hydrogen, and even for 

 carbonic acid, the deviations are posiiive when the gas is sub- 

 jected to a very small pressure ; it is found, moreover, that these 

 deviations increase in proportion to the variation from the normal 

 pressure. The same thing has been found in a new series of 

 experiments undertaken by me with M. Hemilian. The experi- 

 ments are described in tome ii. of my work on the "Elasticity 

 of Gases," which I have just published. A brief extract on this 

 subject is published in the Arm. de Chiimc et de Physique, 

 October, 1876. I shall quote only the results obtained by us 

 from the experiments made in 1875 and in the beginning of 

 1876. 



Into a new apparatus we have introduced several further im- 

 provements, of which the chief are : — (i) The baromanometer, 

 the metre, and the reservoir, containing the gas and the mercury, 

 have been placed in the same bath full of water ; (2) We have 

 succeeded in producing a complete vacuum in the barometric 

 chamber ; (3) The bath was maintained at an almost uniform 

 temperature by means of an agitator, and the small differences 

 in the temperatures of the var.ous layers have been determined 

 by a differential thermometer ; (4) The junction between the 

 air reservoir and the baromanometer has been made, not only 

 without the aid of a tap, but also without the use of mastic.^ 

 Thus the gas was surrounded only by the glass and the mercury. 

 We shall confine ourselves to a summary of the results of our 

 experiments, made between 650 and 20 millimetres' pressure, 

 with four gases — H, air, CO-, and 30^. 



1. If, starting with a certain small pressure, we arrive at pres- 

 sures smaller still, we find for all gases positive deviations, viz. , 



-~ — - > o ; the gases, then, are in this case less compressed 



than Mariotte's Law requires. Similar deviations were also 

 observed for hydrogen by M. Regnault between I and 30 atmo- 

 spheres, and M. Natterer for all gases between 100 and 3,000 

 atmospheres. 



2. Under small pressures and for all gases, the value of the 



positive deviations, i.e., the numerical quantity (or magnitude) 



d( ■pvS 



-~^ , mcreases when the initial pressure diminishes. Thus, 

 dp 



for example, for hydrogen at 400 millimetres — 



'■f ' — + O '000002, 

 dp 



_^^_i = + O'OOOOIO, 



and at 120 millimetres- 



d(pv) _ 



dy 



3. For gases like CO^ and SO* we find near the atmospheric 

 pressure, negative deviations ; ^.^.,'for CO^, /^ = ^ZS> Pi — 20O, 

 /o 5^'p = 10,000, p-^ v-^ = 10,029 ; but, under less pressures still, the 

 deviations become positive even for CO^ and SO^. For example, 

 for CO'^, /„ = 190, /i = 64, p2 = 22, /o ^'o = 10,000, /i v^ - 

 9,996, p.2 f2 = 9,983 ; for SO^ /o = 190, /i = 60, /2 = 22, 



PqZ''^ — 10,000, /i 2/, = 10,010, /a 572 = 9,996- 



4. The existence of positive and negative deviations for the 



' To attain this end the gas-vessel and the branch of the baromanometer 

 are soldered together by a capillary tube made of a single piece. 



