Page images

of melting points as a means of identification-often an important matter. It will be found, for example, that many of the felsparsalthough related in composition-can be distinguished at once from one another. As a more particular instance I may mention that in this way I first arrived at the identification of iolite in the Dublin granite when present in microscopical crystals. Indeed, when minute quantities of substances are in question the application of this thermal method of identification in many cases exceeds all other physical tests in readiness and surety.

The scale of fusibility of Van Kobell is still quoted in the most recently published mineralogical works. It constitutes a comparative method, beset with errors, even as a means of identification, save under special circumstances. Thus whether a splinter of a mineral will melt in the flame of the blowpipe depends as much, or more, on the shape, conductivity, and dimensions of the splinter as on its melting point. For example, a filament of the fibrous actinolite is more easily fused than a fragment of the compact orthoclase. The melting point of actinolite is, however, some one hundred degrees centigrade above that of orthoclase. Again, the chemical action of the flame may sometimes interfere. But not alone do these sources of error arise in comparing the behaviour of substances in the blowpipe flame, but the scale of Van Kobell is itself highly irregular in its spacing, and even erroneous in the order of fusibilities. Thus the melting points in centigrade degrees of specimens of Van Kobell's standards, as determined by the method to be described later, are as follows:

[merged small][ocr errors][ocr errors][merged small][ocr errors]

The numbers prefixed give the order of fusibilities as ordinarily given for Van Kobell's scale. In some cases the temperatures found must be considered approximate, owing to the fact that the substances become viscous at high temperatures, and do not suddenly change state. A margin of 10° at each side of the number given will very certainly

" See Proceedings, Royal Dublin Society, vol. v., N. S., pp. 68 and 69.

include a temperature at which melting would, even in the finest dust, fail to occur, or would occur decisively. But notwithstanding the absence of definite temperatures of melting for these substances it is apparent that this scale-although of good service in its time—is unsymmetrical as well as erroneous in the order of melting points.

The “Meldometer” (ueldw, I melt), as I have designated the instrument used in determinations of melting points, is of the following construction :-A ribbon of pure platinum, having a width of about 1.2 mm., is stretched between forceps, furnished each with a binding-screw, and insulated from each other so that on connecting the binding-screws with a battery a current can be passed through the ribbon. Upon the surface of this ribbon the substance to be examined is placed. It is necessary first to reduce the mineral to a fine powder in an agate mortar, and finally grind it with a little water till in the form of a fine fluent paste; a speck of this is spread over a small area of the ribbon. The best condition for observing melting is when a few particles are thinly spread here and there upon the platinum. A quantity invisible to the unassisted eye may be dealt with, for in all cases the phenomena of melting are observed through a microscope. The microscope having been brought to bear upon the thin coating of the powdered mineral upon the platinum, a current of gradually increasing intensity is passed through the ribbon till the mineral melts or volatilizes. In this process the mineral often exhibits very characteristic phenomena. It remains to describe how the temperature of the platinum ribbon at the moment at which the mineral melts may be determined. Commencing, however, with the simplest form of the apparatus, I proceed to describe a form of the meldometer with which observations, directly comparative between the substance being investigated and other substances of known melting points, are effected. In cases in which identification is the object in view, we compare the behaviour of the substance with that of the species to which we think it referable from its other characters.

Fig. 1 shows, to real scale, a plan and side elevation of this form of the meldometer intended to rest on the stage of a microscope. It consists of two forcepsone insulated-attached to a disk-shaped brass plate. This latter may be held down on the stage of the microscope

* This form has been already noticed briefly in Nature, XXXIII. (1885), pp. 15–16. See algo Journal of the Royal Microscopical Society, x., p. 1068. A further notice of the meldometer is reported from a Paper read before the British Association at Bath, 1888, in Industries, vi., p. 20.

[ocr errors]

by the stage clamps, and is placed so that the ribbon stretched between the forceps is brought to traverse the field of view. A one-inch foreign objective of cheap pattern is suitable for this work. I have had one in use for some years, and it seems uninjured. These powers have single lenses. If a compound cemented objective is used the Canada balsam used in the cementing melts if observations be prolonged at the higher temperatures. The forceps holding the ribbon close by their elasticity, and are opened by the small screws threaded into the upper member of each forceps. The ribbon is shown cut away where it is held by the forceps. This is done with a scissors before inserting it. The object is to secure a more uniform temperature over the length of the ribbon, which otherwise diminishes

Fig. 1. rapidly in temperature approaching the forceps by reason of thermal conductivity, the effect being intensified by decreased electrical resistance where the ribbon is colder. The effect of reducing the section is to diminish this loss of heat by conduction, and at the same time the electrical resistance increasing with diminished cross section up to the forceps, causes a development of heat which augments as the cold metal of the forceps is approached. In this way the loss of heat to the forceps is made good, and the central parts of the short strip made more uniform in temperature. Exact uniformity of temperature throughout is not needful, for the method is one of comparison, and if the substances being compared are brought to adjacent portions of the ribbon it may be assumed they are under like conditions as regards temperature.

The length assigned to the ribbon is conveniently about two centimetres. A storage cell or a couple of grove cells will furnish sufficient current to raise to a blinding white heat and finally fuse the ribbon. The current is regulated as follows :—Two rods of carbon, 51 cms. in length, 17 mms. in diameter, such as are used for electric-arc lighting, having each a resistance of about half an ohm, are clamped at their ends upon a piece of board. A sliding-piece consisting of two sprung brass tubes encircling each carbon, and connected by a crosspiece of brass, can be moved along from end to end of the carbons. At

melted befects polar it is neces Lanıp-fla

one end the carbons are each furnished with a binding screw, and are insulated from one another save where connected by the sliding piece. This resistance is placed in circuit with battery and meldometer, so that the current flows through more or less of the carbon rods according to the position of the slider. Thus, the total resistance in circuit may be varied by moving the latter, and in this way the quantity of current traversing the circuit controlled. Having placed the two substances to be compared side by side, we can thus expose both to any temperature up to that at which the platinum breaks. It will be found that quartz may be melted before this point is reached, and reduced to a glass which no longer affects polarized light. At these high temperatures the light is very intense, and it is necessary to shield the eye from its effects. A cover glass smoked over a lamp-flame may be placed above the eye-piece of the microscope, but a small piece of neutraltinted glass, such as is used in snow spectacles, is better for the purpose.

When the platinum is heated in this form of the meldometer, as the forceps are fixed in position, the expansion results in considerable sag of the ribbon. This, however, is only an inconvenience in so far as it necessitates re-focussing the microscope upon the object. By making one of the forceps free to rotate on a vertical axis at the binding screw, and affixing a weak spiral controlling-spring, so that the ribbon is always distended by a small force, this inconvenience may be avoided, or the distance of the forceps apart may be adjustable by a screw. (See figure in the Journal of the Royal Microscopical Society x., p. 1068); but the simpler form of the apparatus is preferable.

If we seek to identify the body of course we ultimately compare it with accredited specimens of the body we suspect it to be; and here, even if infusible below the melting point of the platinum, information is often gained by comparing the behaviour of the substances. Often again, the phenomena which precede fusion, accompany it, or succeed it, are of much determinative value. Thus topaz blisters on the surface, glassy bubbles forming and breaking up which throw out viscous threads and emit a gas (fluorine ?) which attacks the platinum, forming coloured rings upon the bright-red metal : iolite turns milkwhite. The presence of any considerable quantity of calcium oxide produces this latter effect. Labradorite may be distinguished from the other felspars by this phenomenon. Orthoclase at high temperatures develops large bubbles—which seldom break-throughout the clear viscous liquid into which it is transformed. Tourmaline boils easily, often with effervescence at first, ultimately settling into a quiet boil. Pyromorphite flows into an oily-looking liquid part of which runs along the ribbon till it attains a cooler region near the forceps, where it pulsates backward and forward; a surface-tension phenomenon. On cooling it is seen that a double-coloured slag is produced as if a separation had been effected. Molybdenite sublimes, building up a frail skeleton-like structure over the ribbon, which consists of glistening, tabular, colourless crystals. These again resublime to similar crystals if tumbled back upon the platinum.

Some substances show a change of colour when first heated, as realgar, which blackens, and beryl, which bleaches. Some remain viscous to the last and have no definite melting point, but gradually soften. Beryl is a case of this, and the felspars, in a less degree, possess the property. Other minerals break down suddenly and flood the platinum, as garnet, tourmaline, &c. Finally, the colour of the melted mineral—whether a slag or a glass—affords distinctive characters as in the blow pipe, but in a more extended degree inasmuch as the range of temperature is greater, under perfect control, and the method more cleanly in use, uncomplicated by secondary effects, and the substance is under much better conditions for observation.

As regards the determination of melting points in centigrade degrees by the use of this simple form of apparatus, the closeness with which we can approximate to a true estimate depends on our ability to map out the range of temperature at our command with reliable and sufficiently varied standards of reference. As the construction of such a table or scale of melting points should be based upon the widest possible experience, I do not suggest any complete scale here; but in the second part of this paper will endeavour to rectify the omission. For the present, it will be seen from the list given at the conclusion of this paper, that a wide choice exists of comparative substances up to a certain range of temperature, from the determinations of Carnelley and others on metals and pure salts. The scale will have to exclude as far as possible substances which undergo a period of riscosity before decisively melting. To such substances no perfectly definite melting point can, of course, be assigned, and there is much difficulty in determining the temperature at which they begin to soften. It will be best to record in such cases a temperature at which, after a prolonged interval, there are decisive indications of softening, and a second temperature at which the body may be said to melt rapidly, meaning that in a space of four or five minutes there is distinct flooding of the finer dust upon the surface of the platinum. In all cases it is this finer dust

« PreviousContinue »