MOLECULAK WEIGHTS.] CHEMISTRY 471 mine the atomic weights of elements which do not furnish volatile compounds, it being merely necessary to ascertain the equivalent of the element by analyzing its compounds, and to multiply the equivalent by such a number that the product when multiplied by the specific heat of the element will furnish a number not less than 5 5, and not much greater than 6 5. As an instance of the application of the method, the metal indium, one of the most recently discovered elements, may be cited. Analysis showed that the equivalent of indium was 37 8, but this number was doubled on account of the many points in which indium resembles zinc, the latter metal having the equivalent 32 45, but the atomic weight 64 9. For some time, until Bnnsen determined its specific heat, 75 6 was accepted as the atomic weight of indium, but he found that it was necessary to multiply the equivalent by 3 in order that the product of specific heat into atomic weight should corre spond to that ordinarily observed, thus raising the atomic weight to 113-4. The following table is a list of the elements whose atomic weights have been determined from the observation of their specific heats alone : Name of Element. Specific Heat. Atomic Weight. Atomic Heat. Aluminium .. . 214 07-3 5 8 Cadmium 0567 111 6 6 3 Calcium 170 39 9 6 8 Cerium 0447 141 6 2 Chromium 100 52 4 5-2 1 Cobalt 107 58-6 6 3 Copper 0952 63 3 6 Pidymium 0456 147 6-7 Glucinum 64 9 3 5-9 Gol-1 0324 196 2 6 4 Indium 0570 113 4 65 Indium 0326 1967 6 4 Iron 114 55-9 6 4 Lanthanum 0448 139 6 2 Lithium 9408 7-01 6 6 Magnesium. .. 250 23-94 6 Manganese 122 54 8 67 Nickel 109 58 6 6 4 Palladium 0593 106-2 6 3 Platinum 0324 1967 6 4 Potassium 166 39-04 6 5 Rhodium 0588 104 1 6 1 Ruthenium 0611 103 5 6 3 Silver 0570 107 - 66 6 1 Sodium 293 23 fi-7 Thallium 0335 203 64 fi-S The only elements of which at present the atomic weights have not been ascertained, either by the vapour- density or specific heat method, are erbium, thorium, uranium, and yttrium. Even the equivalent of gallium is not yet known. Often, also, information which is of service in determining the atomic weights of elements may be obtained by observ ing the manner in which their compounds crystallize, for it is found that in many instances a given element in a com pound may be displaced by another without altering the crystalline form ; for example, in ordinary alum, which is a double sulphate of potassium and aluminium, the potas sium may readily be displaced by sodium, or the aluminium by chromium, and yet the resulting compounds crystallize in precisely the same manner as the alum. Substances which thus agree in crystalline form -are said to be iso morphous. A great number of isomorphous substances have been examined by chemists ; and they have been led to infer that, as a rule, when two compounds containing similar elements agree in crystalline form they contain the same number of atoms; and hence the quantity of an element which is capable of displacing an atom of another 1 This number is known to be too low. element in a compound without altering its crystalline- form is regarded as its atomic weight. This is by no means always the case, however, and the occurrence of isomorphism cannot alone be taken as a proof of the atomic weight of an element ; it is of service rather as a check on the determinations made in other ways, and as a stimulus to investigation in the case of elements which have not been sufficiently examined. For instance, certain phosphorous and arsenic minerals of similar composition were known to be isomorphous with a vanadium mineral, vanadinite, the composition of which, according to the received atomic weight of vanadium, exhibited no analogy with them. Roscoe was led by this to investigate anew the compounds of vanadium and to determine its atomic weight, and he found that what had hitherto been regarded as vanadium was really an oxide of the element, and that when the composition of the mineral vanadinite was calculated from the true atomic weight of vanadium, it was precisely similar to that of the minerals with which it was known to be isomorphous. Molecidar Weights. We have pointed out (p. 4 09) that it is to be supposed that the molecule of ^hydrogen consists of two atoms; hence, if the atomic weight of hydrogen be taken as .1 its molecular weight is 2. In order to ascertain the molecular weights of other elements that is to say, the relative weights of their molecules referred to that of hydrogen it is merely necessary to determine their densities referred to hydrogen as unity, and then to multiply the densities by 2. Unfortunately, owing to the high temperature at which most of the elements are converted into vapour, the densities of only very few of them are as yet known. When, however, the molecular weights of the elements are compared with their atomic weights it is found that they are not always, as in the case of hydrogen, double their atomic weights ; hence it is inferred that the molecules of elements do not all contain two atoms. In a few cases the atomic weight and molecular weight agree, which necessitates the conclusion that the molecules are monatomic or con sist of a single atom ; in a few other cases the molecular weight is either 4 or 6 times the atomic weight, and the molecules are therefore regarded as tetratomic or hexatomic. The following table includes all the elements of which the molecular weights have been determined : Xame. Atomic Weight. Molecular Weight . Xo. of Atoms in Molecule. Hydrogen 1 2 2 Chlorine 35-36 79-74 2 Bromine 7975 159-50 2 Iodine 126-53 253 06 9 Nitrogen 14-01 28 02 2 Oxygen 15-96 31-92 2 Selenium 79 158 2 Tellurium 128 256 2 Mercury 199-8 199 8 1 Cadmium 111-6 111*6 1 Phosphorus.... 30-94 1 9 376 4 Arsenic .. 74-9 299 6 4 Sulphur . 31 98 j 63-96 2 | 191-88 6 ] It will be seen that two numbers are given for sulphur This is because at temperatures above 80.0 C. the density of sulphur vapour is such as to indicate that the sulphur molecule consists of 2 atoms, whereas its density at about 500 C. is three times as great, and, consequently, it is to be supposed that the molecules are hexatomic. Selenium, which is closely allied to sulphur, exhibits a very similar be haviour, its vapour at about 1400 C. containing only
diatomic molecules, but as the temperature sinks its density