18-t8.] 



THE CIVIL ENGINEER AND ARCHITECT'S JOURNAL. 



265 



THEORY OF STEAM-ENGINES. 



Account of the experiments to determine the principal laws and 

 numerical data which enter into the calculation of Steam-Engines. 

 By M. V. Regnault. 



fContirMedfiom page 240.J 



Second Memoir. — the deteemination of the density op oases. 



The common method of determining the density of a gas consists 

 in weighing a glass globe of great size : — 



1st. When the globe contains perfectly dry air of a known tem- 

 perature, and under a known atmospheric pressure : 



2nd. After having exhausted it by means of the air-pump, so 

 that the inclosed air exerts only a very feeble pressure, which, as 

 well as the corresponding temperature, is noted : 



3rd. After having filled the globe with the gas perfectly purified, 

 the atmospheric pressure and temperature being again noted : 



4th. After having again exhausted the globe, noting, as before, 

 the pressure and temperature. 



By which four weigliings, and their accompanying observations 

 of pressures and temperatures, all the data necessary for deter- 

 mining the density of the gas are given. But this method requires 

 the exact knowledge of several elements, the determinations of 

 which present generally great uncertainty. 



In the first place, the temperatures which the air and the gas 

 present at the moments of closing the globe, must be very exactly 

 known ; but the method generally used of having a thermometer 

 placed near the globe is very defective : the temperature indicated 

 by the thermometer may be totally different from that of the air 

 in which it is bathed, and still more so from that of the gas which 

 fills the globe. 



MM. Dumas and Boussingault, who have lately {Ann. de Chim. 

 et de Phys., 3d sirie, tome iii., p. 270) very successfully endea- 

 voured to determine the densities of certain gases, place the ther- 

 mometer intended to indicate the temperatures of the gases in the 

 very centre of the globe; and for more security, they cause that 

 temperature to be almost constant, by placing the globe in an 

 inclosure formed by a large cylindrical vessel of zinc with double 

 walls. The annular space left between the walls is filled with 

 water, at a temperature differing but little from that of the sur- 

 rounding air. AVith this arrangement, we may admit that the 

 temperature of the gas is known with sufficient accuracy. 



But the greatest uncertainties exist in the weighings of the 

 globe ; for we must weigh the globe in the air, and, to have its 

 true weight, we must add to its apparent weight that of the air 

 which it displaces. And, in certain cases, this latter weight is 

 greater than that of the gas which fills the globe, so that it must 

 be known with at least equal exactness. We are, up to a certain 

 point, masters of the gas which we pass into the globe. We may 

 prepare it so as to be sure of its purity ; but it is not so with the 

 external atmosphere ; we are obliged to take it as it is. In a close 

 chamber the air may change its composition very perceptibly ; its 

 temperature, and the quantity of moisture which it contains, very 

 incessantly. MM. Dumas and Boussingault thought that they had 

 completely avoided the errors arising from this source, by placing 

 below their balance a large chest lined with lead, in which the 

 globe suspended from one of the scales of the balance floats. A 

 very sensitive thermometer is placed in this chest, and gives the 

 temperature of the air. This arrangement is certainly far prefer- 

 able to allowing the globe to hang freely in the air of the room ; 

 the globe suspended in the chest is kept from the currents of air, 

 which render the weighings very uncertain, and the temperature 

 of the air in which it is placed changes but slowly ; but it does 

 not do away with the errors arising from the changes in the com- 

 position of the air, and these are by no means negligible, especially 

 when we are working with very light gases — such, for instance, as 

 hydrogen. 



At the same time that MM. Dumas and Boussingault were 

 weighing gases, M. Regnault was also engaged in the determination 

 of the density of steam under different pressures, and especially 

 under very feeble pressures. He was struck with the uncertainties 

 which the ordinary methods of weighing gases present, especially 

 owing to the alteration of the density of the surrounding air, which 

 seems to have attracted but little attention from men of science, 

 and he was led to a method which presents a degree of certainty 

 and precision which those heretofore used do not offer. 



He avoids completely, and by a very simple artifice, the uncer- 

 tainties which arise from the changes in the air in which the globe 



is weighed. In place of equipoising it by means of weights on the 

 opposite scale, he balances it by means of a second hermetically- 

 sealed globe of the same kind of glass, hung from the opposite 

 scale. All the variations which take place in the air then affect 

 the two globes in the same way, whether they arise from changes 

 of temperature, barometric pressure, or composition of the atmo- 

 sphere. It is not, therefore, required at the moment of weighing 

 to watch the thermometer, barometer, and hydrometer ; it is sufii- 

 cient to wait until the two globes are in equilibrium of temperature, 

 and when tliis is the case, it continues indefinitely. We liave, in 

 consequence, a very definite character by which to know when to 

 read the weighings. Tliis method presents also another advantage, 

 that is, of avoiding the error arising from the different amount of 

 moisture deposited upon the globe during difi'erent weighings. As 

 the globes are made of the same glass, and equally dried before the 

 commencement of the experiments, they may be assumed to con- 

 dense the same amount of moisture when bathed in the same air, 

 and consequently will remain in equilibrium. 



The globes had a capacity of about 10 litres (2-2 gallons). The 

 globe in which the gas is to be weighed has a stop-cock adjusted to 

 it, so that it may be exposed to the temperature of boiling water 

 without leakage. It is weighed when full of water, first in the air, 

 and afterwards in water of the same temperature as that which it 

 contains ; thus is obtained the weight of water displaced by the 

 globe. 



The balancing globe is so selected that the weight of water dis- 

 placed by it is rather less than that of the first, even after the 

 addition of its metallic mounting by which it is hermetically sealed, 

 and hung from the opposite balance scale ; there is then added to 

 it a glass tube of such capacity as that the weight of water dis- 

 placed by it shall just make up the difference. 



Before closing the second globe, a quantity of mercury was intro- 

 duced into it so as to render it about 10 grammes (154-38 grains 

 troy) heavier than the other. The two globes thus adjusted were 

 submitted to several tests, in order to be sure that they satisfied the 

 required conditions ; they were left hanging for fifteen days under 

 the balance scales, and the equilibrium was rigorously maintained 

 all this time, although in the interval the temperature of the air 

 had changed from 0° to 17°, (32° to 62-5° Fahi.,) and the barome- 

 tric pressure from 741 to 771 millimetres (29-6 to 30-8 inches.) 

 The general mode of operating was as follows : — A vacuum as 

 complete as possible being made in the globe, it is placed in coni- 

 munication with the apparatus for producing the gas whose density 

 is to be determined, and the stop-cock is opened in such a way that 

 the gas in the apparatus preserves a slight excess of pressure. 

 When the globe is filled with gas, it is again placed in communica- 

 tion with tiie air-pump, a very perfect vacuum made, and it filled 

 a second time with the gas. In order to avoid any correction for 

 temperature — a correction which would require the knowledge of 

 the co-efficient of dilatation of the gas, and that of the globe— the 

 globe is placed in a zinc cover and completely enveloped in melting 

 ice. Before closing the globe, it is placed in direct communication 

 with the atmosphere, so that the gas may place itself in equilibrium 

 with the exterior pressure. 



The globe taken out of the ice is carefully washed and dried, and 

 hung from the scale of the balance. It requires a long time (often 

 more than two hours) for the globe to take exactly the temperature 

 of the sun-ounding air, and for its surface to cover itself with its 

 normal amount of moisture. The balance used was able to appre- 

 ciate with certainty a half milligramme when charged with one 

 kiUogramme upon each scale.* "It was placed over a large chest, 

 such as used by MM. Dumas and Boussingault. At the end of the 

 weighing, the observer did not approach the balance, but observed 

 the oscillations of the index at a distance, with a telescope. 



M. Regnault then notices the electric effects produced by wiping 

 the globes, and the effects upon the weighings ; he avoided it by 

 wiping the globes with a napkin dipped in distilled water, and 

 tested them by the gold-leaf electrometer. The pressures were 

 measured by an apparatus which he describes under the name of a 

 barometric-manometer. It consists of two tubes, one of which is 

 an ordinary barometer of 20 millim. (0-8 in. diameter,) made very 

 carefully ; the other is a glass tube of the same diameter, whicJi 

 may, by a lead tube, be connected with the vessel in which pres- 

 sures less than that of an atmosphere are to be measured , they 

 plunge below into a cistern of mercury having a partition, so that 

 the two instruments may be separated at pleasure by drawing off 

 the mercury in the cistern of the manomete r below the top of the 



« That is. sensible to one four-millionth of the total load. Quere, whether this sensi- 

 Mlily was determined alter the globes had beeu hangiug from the scales, aud the balanc* 

 iu actiou for lifttteD days ? 



35 



