From what has been said of boiling solutions, it may be seen that the boiling-point of a liquid is the higher, the greater the amount of foreign substance contained in it.
This is. however, not the only factor on which the temperature of a boiling solution depends. The molecular weight of the dissolved substance is another factor. If we were to take two equal quantities, say, of water, heat them to boiling and then add to them. respectively, equal weights of two different substances. we would find that the elevation of temperature is greater in the case of the sample to which we have added the substance of smaller molecular weight. Experi ments out in this manner with many different substances permit us to induce the law, that the difference of the boiling-tem pera t urcs of a solution and of the pure solvent is in rcrscl y proportional to the molecular weight of the substance dissolved. It must, however, be borne in mind that the law is limited to solutions which do not conduct electricity; in solutions of electrolytes relations are not quite so simple. (See articles SOLUTIONS and DissocIA Tiox.) In the case of solutions of non-electro lytes, such as many of the carbon compounds, the law holds with considerable precision; and it has proved especially useful in this, that it permits us to determine the unknown molecular weights of newly discovered substances. To as certain, by this method, the molecular Nveigh t of a new substance. all the chemist has to do is first to observe the rise of temperature produced in a boil ing solvent, when a certain quantity of his substance is add ed to it, and then to compare the rise of boiling-point to the rise produced by the same quantity of some other sub stance, of known molecular weight.
These important de terminations a r e carried out with considerable preci sion by the use of Beckniann's a p p a ratus (see figure).
The a pparatuseon sists essentially of two parts: an inner tube, a, containing the solution, and an outer vessel, b, con taining the pure sol vent, and separated from a near the bottom by a sheet of asbestos. The coolers c and d serve to condense the escaping va pors. To determine the rise of boiling-point caused by a given substance, the operator introduces into a first a few cubic centimeters of the pure solvent, observes the boiling-point, then introduces also a known amount of his substance, and again ob serves the boiling-point. In the ease of most
solvents, the inner tube need not be heated di rectly, and heating only the outer vessel is suffi cient to keep the liquid in the inner tube boiling.
What has been said above of pure boiling water applies equally well to any other com pound in the liquid state: any liquid consisting of one single chemical compound has a definite boiling-point depending on nothing but the ex ternal pressure. Thus, under normal atmospheric pressure, the boiling-point of ether is 35° C.; the normal boiling-point of pure alcohol is 78° ; aniline boils at 183°, and the hydrocarbon picene at 520°. The boiling-temperature is, of course, very easy to determine; and since it is one of the highly characteristic properties of a liquid, a chemist determines that point usually before inquiring into other properties when called upon to identify a liquid compound. Further, since in the case of a pure chemical compound, if the vapors are allowed to escape, the boiling-tem perature remains constant, while in the case of a solution (o• any other homogeneous liquid mixture), it continuously rises: whenever a chemist wants to ascertain whether he is deal ing with a single and pure chemical compound, he subjects his liquid to distillation, to see whether the boiling-point will remain constant. If it does not remain constant he knows that he is dealing with a mixture or that some chemical change is taking place in the liquid.
From the fact that every compound has a char acteristic boiling-point of its own. chemists were naturally led to think: (l ) that some definite relation must surely exist between the chemical nature of compounds and their boiling-points, and hence (2) that in comparing compounds of simi lar constitution, definite numerical relations should be expected to exist between the boiling points of such compounds. As a matter of fact, in the homologous series of organic chemistry, the boiling-point is generally found to increase with the complexity of the molecule (see, for example, the article ; which shows that the boiling-point of compounds is really dependent on their nature and structure. The numerical relations hitherto discovered are, however, far from being definite and general. Probably we do not know as yet how to search for the more definite relations.