=To Calibrate Tubes for Measuring Gases.=--Prepare a small gla.s.s tube sealed at one end and ground at the other to a plate of gla.s.s. The tube should hold about as much mercury as will fill 10 mm. divisions of the graduated tube. Fill this tube with mercury, removing all bubbles of air that adhere to the sides by closing the open end of the tube with the thumb, and washing them away with a large air-bubble left for the purpose. If any persistently remain, remove them by means of a fine piece of bone or wood. Then completely fill the tube with mercury, removing any bubbles that may be introduced in the operation, and remove the excess of mercury by placing the ground-gla.s.s plate on the mouth of the tube, and pressing it so as to force out all excess of mercury between the two surfaces. Clean the outside of the tube, and place it on a small stand (this may be a small wide-mouthed gla.s.s bottle), with which it has been previously weighed when empty, and re-weigh. Repeat this operation several times. From the mean of the results, which should differ one from another but very slightly, the capacity of the tube can be calculated.

The purest mercury obtainable should be used. Since the density of pure mercury at 0 C. is 13596, the weight of mercury required to fill the tube at 0 C., taken in grams, when divided by 13596, will give the capacity of the tube at 0 C. in cubic centimetres. If the experiment be not made at 0 C., and if a very exact determination of the capacity of the tube be required, the density of mercury must be corrected for expansion or contraction.

Having now a vessel of known capacity, it can be employed for ascertaining the capacities of the divisions of a graduated tube in the following manner:--The graduated tube is fixed perpendicularly, mouth upwards, in a secure position. The small tube of known capacity is filled with mercury as previously described, and its contents are transferred to the divided tube. The number of divisions which the known volume of mercury occupies is noted after all air-bubbles have been removed. This process is repeated until the divided tube is filled. A table of results is prepared, showing the number of divisions occupied by each known volume of mercury introduced.

In subsequently using the tube the volumes of the gases measured in it must be ascertained from the table of values thus prepared.

In observing the level of the mercury, unless a cathetometer is available, a slip of mirror should be held behind the mercury close to the tube, in such a position that the pupil which is visible on the looking-gla.s.s is divided into two parts by the surface of the mercury.

A correction must be introduced for the error caused by the meniscus of the mercury. As the closed end of the tube was downwards when each measured volume of mercury was introduced, and as the surface of mercury is convex, the volume of mercury in the tube when it is filled to any division _l_ (Fig. 41) is represented by _A_ of 1. But in subsequently measuring a gas over mercury in the same tube, when the mercury stands at the same division _l_, the volume of the gas will be as represented by _B_ of 2, which is evidently somewhat greater than _A_. This will be seen still more clearly in 3, where _a_ represents the boundary of the mercury, and _b_ the boundary of the air, when the tube is filled to the mark _l_ with mercury or a gas over mercury respectively.

[Ill.u.s.tration: FIG. 41.]

It is plain that when the level of the mercury in measuring a gas is read at _l_, the volume of the gas is greater than the volume of the mercury recorded, by twice the difference between the volume _A_ of mercury measured, and that which would fill the tube to the level _l_, if its surface were plane.

The usual mode of finding the true volume of a gas collected over mercury is as follows:--

Place the graduated tube mouth upwards, introduce some mercury, and, after removing all bubbles, note the division at which it stands. Then add a few drops of solution of mercuric chloride; the surface of the mercury will become level, read and record its new position. Then, in any measurement, having observed that the mercury stands at _n_ divisions of the tube, add twice the difference between the two positions of the mercury to _n_, and ascertain the volume which corresponds to this reading from the table of capacities.

=To Calibrate the Tube of a Thermometer.=--Detach a thread of mercury from half an inch to one inch in length from the body of the mercury.

Move it from point to point throughout the length of the tube, and note its length in each position. If in one part it occupies a length of tube corresponding to eight degrees, and at another only seven degrees, then at the former point the value of each division is only seven-eighths of those at the latter position.

From the results obtained, a table of corrections for the thermometer should be prepared.

It is sometimes necessary to join soda gla.s.s to lead gla.s.s. In this case the edge of the lead gla.s.s tube may be bordered with white enamel before making the joint. Enough enamel must be used to prevent the lead and soda gla.s.ses from mingling at any point. The enamel is easily reduced, and must be heated in the oxidising flame. Dr. Ebert recommends _Verre d'urane_ for this purpose. It is supplied by Herr Gotze of Leipzig (Liebigstra.s.se).

CHAPTER VI.

_GLa.s.s TUBING._

The diagrams given below show the sizes and thickness of the gla.s.s tubes most frequently required. In ordering, the numbers of these diagrams may be quoted, or the exact dimensions desired may be stated.

Gla.s.s tubes are usually sold by weight, and therefore the weight of tube of each size that is wished for should be indicated, and also whether it is to be of lead or soda gla.s.s.

[Ill.u.s.tration]

[Ill.u.s.tration]

CHAPTER VII.

_VITREOUS SILICA._

=Introductory.=--Vitreous Silica was made in fine threads by M. Gaudin in 1839,[22] and small tubes of it were made in 1869 by M. A. Gautier, but its remarkable qualities were not really recognised till 1889, when Professor C. V. Boys rediscovered the process of making small pieces of apparatus of this substance, and used the torsion of "quartz fibres" for measuring small forces. More recently the author of this book has devised a process for preventing the "splintering" of quartz which gave so much trouble to the earlier workers, and jointly with Mr. H. G.

Lacell, has produced a variety of apparatus of much larger dimensions than had been attempted =previously=. At the time of writing we can produce by the processes described in the following pages tubes 1 to 15 cm. in diameter and about 750 cm. in length, globes or flasks capable of containing about 50 cc., ma.s.ses of vitreous silica weighing 100 grams or more, and a variety of other apparatus.

[22] A brief summary of the history of this subject will be found in _Nature_, Vol. 62, and in the Proceedings of the Royal Inst.i.tution, 1901.

=Properties of Vitreous Silica.=--For the convenience of those who are not familiar with the literature of this subject, I may commence this chapter with a brief account of the properties and applications of vitreous silica, as far as they are at present ascertained. Vitreous silica is less hard than chalcedony, but harder than felspar. Tubes and rods of it can be cut with a file or with a piece of sharpened and hardened steel, and can afterwards be broken like similar articles of gla.s.s. Its conducting power is low, and Mr. Boys has shown that fine fibres of silica insulate remarkably well, even in an atmosphere saturated with moisture. The insulating qualities of tubes or rods of large cross sections have not yet been fully tested; one would expect them to give good results provided that they are kept scrupulously clean. A silica rod which had been much handled would probably insulate no better than one of gla.s.s in a similar condition. The density of vitreous silica is very near to that of ordinary amorphous silica. In the case of a small rod not absolutely free from minute bubbles it was found to be 221.

Vitreous silica is optically inactive, when h.o.m.ogeneous, and is highly transparent to ultraviolet radiations.

The melting point of vitreous silica cannot be definitely stated. It is plastic over a considerable range of temperature. Professor Callendar has succeeded in measuring the rate of contraction of fine rods in cooling from 1200 to 1500 C., so that its plasticity must be very slight below the latter temperature. If a platinum wire embedded in a thick silica tube be heated from without by an oxy-hydrogen flame the metal may be melted at temperatures at which the silica tube will retain its form for a moderate length of time, but silica softens to a marked extent at temperatures a little above the melting point of platinum.

It has been observed by Boys, Callendar, and others that fine rods of silica, and also the so-called "quartz fibres," are apt to become brittle after they have been heated to redness. But I have not observed this defect in the case of more ma.s.sive objects, such as thick rods or tubes; and as I have repeatedly observed that mere traces of basic matter, such as may be conveyed by contact with the hand, seriously injure the surface of silica, and have found that silica quickly becomes rotten when it is heated to about 1000 in contact with an infusible base such as lime, I am disposed to ascribe the above-mentioned phenomenon to chemical rather than to purely physical causes.[23] It is certain, however, that silica apparatus must never be too strongly heated in contact with basic substances. Silica is easily attacked by alkalis and by lime, less readily by copper oxide, and still less by iron oxide.

[23] In a recent communication Professor Callendar tells me that the devitrification commences at the outside and is hastened by particles of foreign matter.

The rate of expansion of vitreous silica has been studied by H. le Chatelier, and more recently by Callendar. The former found its mean coefficient of expansion to be 00000007 between 0 and 10000,[24] and that it contracted when heated above 700.

[24] The silica blocks used were prepared by fusion in an electric furnace; it is therefore probable that they were not quite pure.

Professor Callendar used rods of silica prepared by the author from "Brazil crystal"; these were drawn in the oxy-gas flame and had never been heated in contact with solid foreign matter, so that they consisted, presumably, of very pure silica. His results differ in some respects from those obtained by Le Chatelier, for he finds the mean coefficient of expansion to be only 000000059, _i.e._ about one seventeenth as great as that of platinum. Callendar found the rods of silica expanded very regularly up to 1000 but less regularly above that temperature. Above 1200 they contracted when heated.

The behaviour of vitreous silica under sudden changes of temperature is most remarkable. Large ma.s.ses of it may be plunged suddenly when cold into the oxy-gas flame, and tubes or rods at a white heat may be thrust into cold water, or even into liquid air, with impunity. As a consequence of this, it is in one respect much more easily worked in the flame than any form of gla.s.s. Difficult joints can be thrust suddenly into the flame, or removed from it, at any stage, and they may be heated unequally in different parts with impunity. It is safe to say that joints, etc., in silica never crack whilst one is making them nor during the subsequent cooling. They may be set aside in an unfinished state and taken up again without any precautions. Therefore it is possible for an amateur to construct apparatus in silica which he would be quite unable to produce from gla.s.s.

The behaviour of vitreous silica with solvents has not yet been fully investigated, but Mr. H. G. Lacell has this subject in hand. If it behaves like the other forms of anhydrous silica it will withstand the action of all acids except hydrofluoric acid. It is, of course, very readily acted upon by solutions of alkalis and alkaline salts.

As regards the use of silica in experiments with gases, it must be remarked that vitreous silica, like platinum, is slightly permeable to hydrogen when strongly heated. One consequence of this is that traces of moisture are almost always to be found inside recently-made silica tubes and bulbs, however carefully we may have dried the air forced into them during the process of construction. Owing to the very low coefficient of expansion of silica, it is not possible to seal platinum wires into silica tubes. Nor can platinum be cemented into the silica by means of a.r.s.enic enamel, nor by any of the softer gla.s.ses used for such purposes.

I have come near to success by using kaolin, but the results with this material do not afford a real solution of the problem, though they may perhaps point to a hopeful line of attack. Possibly platinum wires might be soldered into the tubes (see _Laboratory Arts_, R. Threlfall), but this also is uncertain.

The process of preparing silica tubes, etc., from Lumps of Brazil Crystal may be described conveniently under the following headings. I describe the various processes fully in these pages, as those who are interested in the matter will probably wish to try every part of the process in the first instance. But I may say that in practice I think almost every one will find it advantageous to start with purchased silica tubes, just as a gla.s.s-worker starts with a supply of purchased gla.s.s tubes. The manufacturer can obtain his oxygen at a lower price than the retail purchaser, and a workman who gives much time to such work can turn out silica tube so much more quickly than an amateur, that I think it will be found that both time and money can be saved by purchasing the tube. At the same time the beginner will find it worth while to learn and practise each stage of the process at first, as every part of the work described may be useful in the production of finished apparatus from silica tubes.

This being so, I am glad to be able to add that a leading firm of dealers in apparatus[25] has commenced making silica goods on a commercial scale, so that the new material is now available for all those who need it or wish to examine its properties.

[25] Messrs. Baird and Tatlock.

=Preparing non-splintering Silica from Brazil Pebble.=--The best variety of native Silica is Brazil Pebble, which may be obtained in chips or larger ma.s.ses. These should be thoroughly cleaned, heated in boiling water, and dropped into cold water, the treatment being repeated till the ma.s.ses have cracked to such an extent that they may be broken easily by blows from a clean steel pestle or hammer.

The fragments thus produced must be hand-picked, and those which are not perfectly free from foreign matter should be rejected. The pure and transparent pieces must then be heated to a yellow-red heat in a covered platinum dish in a m.u.f.fle or reverberatory furnace and quickly plunged into a deep clean vessel containing clean distilled water; this process being repeated, if necessary, till the product consists of semi-opaque friable ma.s.ses, very much like a white enamel in appearance. After these have been washed with distilled water, well drained and dried, they may be brought into the hottest part of an oxy-gas flame safely, or pressed suddenly against ma.s.ses of white hot silica without any preliminary heating, such as is necessary in the case of natural quartz. Quartz which has not been submitted to the above preparatory process, splinters on contact with the flame to such an extent that very few would care to face the trouble and expense of working with so refractory a material.

But after the above treatment, which really gives little trouble, all the difficulties which hampered the pioneer workers in silica disappear as if by magic.

=Apparatus.=--Very little special apparatus need be provided for working with silica, but it is absolutely essential to protect the eyes with very dark gla.s.ses. These should be so dark as to render it a little difficult to work with them at first. If long spells of work are undertaken, two pairs of spectacles should be provided, for the gla.s.ses quickly become hot enough to cause great inconvenience and even injury to the eyes.

Almost any of the available oxy-gas burners may be used, but they vary considerably in efficiency, and it is economical to obtain a very efficient burner. The 'blow-through' burners are least satisfactory, and I have long since abandoned the use of them. Some of the safety 'mixed-gas jets' have an inconvenient trick of burning-back, with sharp explosions, which are highly disconcerting, if the work be brought too near the nozzle of the burner. I have found the patent burner of Mr.

Jackson (Brin's Oxygen Company, Manchester) most satisfactory, and it offers the advantage that several jets can be combined in a group easily and inexpensively for work on large apparatus. The large roaring flames such as are used, I understand, for welding steel are very expensive, and not very efficient for the work here described.