Chapter V: Front Matter (5)

The dangers of the mineral oil lamp, which were a grave drawback in
the past, have been very much reduced by improvements in construction
and quality, and if it were possible to abolish the cheap and
dangerous rubbish sold in poor neighbourhoods, and to prevent the use
of side-fillers and glass reservoirs in lamps of better quality, a
still larger reduction in the number of accidents would take place. In
the use of the lamp for domestic purposes only soft well-fitting wicks
should be employed, and the lamp should be filled with oil each day so
as never to allow it to burn too low and so leave a large space above
the surface of the oil in the reservoir. The lamp should never be
moved whilst alight, and it should only be put out by means of a
proper extinguisher or by blowing across the top instead of down the
chimney. By these means the risk of accident would be so reduced as to
compare favourably with other illuminants.

Candles, oil and coal gas all emit the same products of complete
combustion, viz. carbon dioxide and water vapour. The quantities of
these compounds emitted from different illuminants for every candle of
light per hour will be seen from the following table:

Cubic Feet per Candle
Illuminant. Carbon Dioxide. Water Vapour.

Sperm candle 0.41 0.41
Oil lamp 0.24 0.18
Gas--Flat flame 0.26 0.67
Argand 0.17 0.45
Regenerative 0.07 0.19
Incandescent 0.03 0.08

From these data it appears that if the sanitary condition of the air
of a dwelling-room be measured by the amount of carbon dioxide
present, as is usually done, candles are the most prejudicial to
health and comfort, oil lamps less so, and gas least, an assumption
which practical experience does not bear out. The explanation of this
is to be found in these facts: First, where we illuminate a room with
candles or oil we are contented with a less intense and more local
light than when we are using gas, and in a room of ordinary size would
be more likely to use a lamp or two candles than the far higher
illumination we should demand if gas were employed. Secondly, the
amount of water vapour given off during the combustion of gas is
greater than in the case of the other illuminants, and water vapour
absorbing radiant heat from the burning gas becomes heated, and,
diffusing itself about the room, causes great oppression. Also the
air, being highly charged with moisture, is unable to take up so
rapidly the water vapour which is always evaporating from the surface
of our skin, and in this way the functions of the body receive a
slight check, resulting in a feeling of depression.

Oil-spray lamps.

A very successful type of oil lamp for use in engineering is represented by the Lucigen, Doty, and Wells lights, in which the oil is forced from a reservoir by air-pressure through a spiral heated by the flame of the lamp, and the heated oil, being then ejected partly as vapour and partly as spray, burns with a large and highly luminous flame. The great drawback to these devices is that a certain proportion of the oil spray escapes combustion and is deposited in the vicinity of the light. This form of lamp is often used for heating as well as lighting; the rivets needed for the Forth Bridge were heated in trays by lamps of this type at the spot where they were required. The great advantage of these lamps was that oils of little value could be employed, and the light obtained approximated to 750 candles per gallon of oil consumed. They may to a certain extent be looked upon as the forerunners of perhaps the most successful form of incandescent oil-burner.

Oil applied to incandescent lighting.

As early as 1885 Arthur Kitson attempted to make a burner for heating purposes on the foregoing principle, i.e. by injecting oil under pressure from a fine tube into a chamber where it would be heated by the waste heat escaping from the flame below, the vapour so produced being made to issue from a small jet under the pressure caused by the initial air-pressure and the expansion in the gasifying tube. This jet of gas was then led into what was practically an atmospheric burner, and drew in with it sufficient air to cause its combustion with a non-luminous blue flame of great heating power. At the time when this was first done the Welsbach mantle had not yet reached the period of commercial utility, and attempts were made to use this flame for the generation of light by consuming it in a mantle of fine platinum gauze, which, although giving a very fine illuminating effect during the first few hours, very soon shared the fate of all platinum mantles--that is, carbonization of the platinum surface took place, and destroyed its power of light emissivity. It was not until 1893 that the perfecting of the Welsbach mantle enabled this method of consuming the oil to be employed. The Kitson lamp, and also the Empire lamp on a similar principle, have given results which ought to ensure their future success, the only drawback being that they need a certain amount of intelligent care to keep them in good working order.

Incandescent table-lamps.

Oil gas and oil vapours differ from coal gas merely in the larger proportion and greater complexity of the hydrocarbon molecules present, and to render the oil flame available for incandescent lighting it is only necessary to cause the oil gas or vapour to become mixed with a sufficient proportion of air before it arrives at the point of combustion. But with gases so rich in hydrocarbons as those developed from oil it is excessively difficult to get the necessary air intimately and evenly mixed with the gas in sufficient proportion to bring about the desired result. If even coal gas be taken and mixed with 2.27 volumes of air, its luminosity is destroyed, but such a flame would be useless with the incandescent mantle, as if the non-luminous flame be superheated a certain proportion of its luminosity will reappear. When such a flame is used with a mantle the superheating effect of the mantle itself very quickly leads to the decomposition of the hydrocarbons and blackening of the mantle, which not only robs it of its light-giving powers, but also rapidly ends its life. If, however, the proportion of air be increased, the appearance of the flame becomes considerably altered, and the hydrocarbon molecules being burnt up before impact with the heated surface of the mantle, all chance of blackening is avoided.

On the first attempts to construct a satisfactory oil lamp which could
be used with the incandescent mantle, this trouble showed itself to be
a most serious one, as although it was comparatively easy so to
regulate a circular-wicked flame fed by an excess of air as to make it
non-luminous, the moment the mantle was put upon this, blackening
quickly appeared, while when methods for obtaining a further air
supply were devised, the difficulty of producing a flame which would
burn for a considerable time without constant necessity for regulation
proved a serious drawback. This trouble has militated against most of
the incandescent oil lamps placed upon the market.

It soon became evident that if a wick were employed the difficulty of
getting it perfectly symmetrical was a serious matter, and that it
could only be utilized in drawing the oil up to a heating chamber
where it could be volatilized to produce the oil gas, which on then
being mixed with air would give the non-luminous flame. In the earlier
forms of incandescent oil lamps the general idea was to suck the oil
up by the capillarity of a circular wick to a point a short distance
below the opening of the burner at which the flame was formed, and
here the oil was vaporized or gasified by the heat of the head of the
burner. An air supply was then drawn up through a tube passing through
the centre of the wick-tube, while a second air current was so
arranged as to discharge itself almost horizontally upon the burning
gas below the cap, in this way giving a non-luminous and very hot
flame, which if kept very carefully adjusted afforded excellent
results with an incandescent mantle. It was an arrangement somewhat of
this character that was introduced by the Welsbach Company. The lamps,
however, required such careful attention, and were moreover so
irregular in their performance, that they never proved very
successful. Many other forms have reached a certain degree of
perfection, but have not so far attained sufficient regularity of
action to make them commercial successes. One of the most successful
was devised by F. Altmann, in which an ingenious arrangement caused
the vaporization of oil and water by the heat of a little oil lamp in
a lower and separate chamber, and the mixture of oil gas and steam was
then burnt in a burner-head with a special arrangement of air supply,
heating a mantle suspended above the burner-head.

The perfect petroleum incandescent lamp has not yet been made, but the
results thus obtained show that when the right system has been found a
very great increase in the amount of light developed from the
petroleum may be expected. In one lamp experimented with for some time
it was easy to obtain 3500 candle hours per gallon of oil, or three
times the amount of light obtainable from the oil when burnt under
ordinary conditions.

Air-gas.

Before the manufacture of coal-gas had become so universal as it is at present, a favourite illuminant for country mansions and even villages where no coal-gas was available was a mixture of air with the vapour of very volatile hydrocarbons, which is generally known as "air-gas." This was produced by passing a current of dry air through or over petroleum spirit or the light hydrocarbons distilled from tar, when sufficient of the hydrocarbon was taken up to give a luminous flame in flat flame and Argand burners in the same way as coal-gas, the trouble being that it was difficult to regulate the amount of hydrocarbon held in suspension by the air, as this varied very widely with the temperature. As coal-gas spread to the smaller villages and electric lighting became utilized in large houses, the use of air-gas died out, but with the general introduction of the incandescent mantle it again came to the front. In the earlier days of this revival, air-gas rich in hydrocarbon vapour was made and was further aerated to give a non-luminous flame by burning it in an atmospheric burner.

One of the best illustrations of this system was the Aerogene gas
introduced by A. I. van Vriesland, which was utilized for lighting a
number of villages and railway stations on the continent of Europe. In
this arrangement a revolving coil of pipes continually dips into
petroleum spirit contained in a cylinder, and the air passed into the
cylinder through the coil of pipes becomes highly carburetted by the
time it reaches the outlet at the far end of the cylinder. The
resulting gas when burnt in an ordinary burner gives a luminous flame;
it can be used in atmospheric burners differing little from those of
the ordinary type. With an ordinary Welsbach "C" burner it gives a
duty of about 30 candles per foot of gas consumed, the high
illuminating power being due to the fact that the gas is under a
pressure of from 6 to 8 in. With such a gas, containing a considerable
percentage of hydrocarbon vapour, any leakage into the air of a room
would give rise to an explosive mixture, in the same way that coal-gas
would do, but inasmuch as mixtures of the vapour of petroleum spirit
and air are only explosive for a very short range, that is, from 1.25
to 5.3%, some systems have been introduced in which by keeping the
amount of petroleum vapour at 2% and burning the gas under pressure in
a specially constructed non-aerating mantle burner, not only has it
been found possible to produce a very large volume of gas per gallon
of spirit employed, but the gas is itself non-explosive, increase in
the amount of air taking it farther away from the explosive limit. The
Hooker, De Laitte and several other systems have been based upon this
principle.

2. GAS LIGHTING

In all measurements of illuminating value the standard of comparison used in England is the light yielded by a sperm candle of the size known as "sixes," i.e. six to the pound, consuming 120 grains of sperm per hour, and although in photometric work slight inequalities in burning have led to the candle being discarded in practice, the standard lamps burning pentane vapour which have replaced them are arranged to yield a light of ten candles, and the photometric results are expressed as before in terms of candles.

When William Murdoch first used coal-gas at his Redruth home in 1779, he burnt the gas as it escaped from the open end of a small iron tube, but soon realizing that this plan entailed very large consumption of gas and gave a very small amount of light, he welded up the end of his tube and bored three small holes in it, so arranged that they formed three divergent jets of flame. From the shape of the flame so produced this burner received the name of the "cockspur" burner, and it was the one used by Murdoch when in 1807 he fitted up an installation of gas lighting at Phillips & Lee's works in Manchester. This--the earliest form of gas burner--gave an illuminating value of a little under one candle per cubic foot of gas consumed, and this duty was slightly increased when the burner was improved by flattening up the welded end of the tube and making a series of small holes in line and close together, the jets of flame from which gave the burner the name of the "cockscomb." It did not need much inventive faculty to replace the line of holes by a saw-cut, the gas issuing from which burnt in a sheet, the shape of which led to the burner being called the "batswing." This was followed in 1820 by the discovery of J. B. Neilson, of Glasgow, whose name is remembered in connexion with the use of the hot-air blast in iron-smelting, that, by allowing two flames to impinge upon one another so as to form a flat flame, a slight increase in luminosity was obtained, and after several preliminary stages the union jet or "fishtail" burner was produced. In this form of burner two holes, bored at the necessary angle in the same nipple, caused two streams of gas to impinge upon each other so that they flattened themselves out into a sheet of flame. The flames given by the batswing and fishtail burners differed in shape, the former being wide and of but little height, whilst the latter was much higher and more narrow. This factor ensured for the fishtail a greater amount of popularity than the batswing burner had obtained, as the flame was less affected by draughts and could be used with a globe, although the illuminating efficiency of the two burners differed little.

Regenerative burner.

In a lecture at the Royal Institution on the 20th of May 1853, Sir Edward Frankland showed a burner he had devised for utilizing the heat of the flame to raise the temperature of the air supply necessary for the combustion of the gas. The burner was an Argand of the type then in use, consisting of a metal ring pierced with holes so as to give a circle of small jets, the ring of flame being surrounded by a chimney. But in addition to this chimney, Frankland added a second external one, extending some distance below the first and closed at the bottom by a glass plate fitted air-tight to the pillar carrying the burner. In this way the air needed for the combustion of the gas had to pass down the space between the two chimneys, and in so doing became highly heated, partly by contact with the hot glass, and partly by radiation. Sir Edward Frankland estimated that the temperature of the air reaching the flame was about 500°F. In 1854 a very similar arrangement was brought forward by the Rev. W. R. Bowditch, and, as a large amount of publicity was given to it, the inception of the regenerative burner was generally ascribed to Bowditch, although undoubtedly due to Frankland.

The principle of regeneration was adopted in a number of lamps, the best of which was brought out by Friedrich Siemens in 1879. Although originally made for heating purposes, the light given by the burner was so effective and superior to anything obtained up to that time that it was with some slight alterations adapted for illuminating purposes.

Improvements followed in the construction and design of the regenerative lamp, and when used as an overhead burner it was found that not only was an excellent duty obtained per cubic foot of gas consumed, but that the lamp could be made a most efficient engine of ventilation, as an enormous amount of vitiated air could be withdrawn from the upper part of a room through a flue in the ceiling space. So marked was the increase in light due to the regeneration that a considerable number of burners working on this principle were introduced, some of them like the Wenham and Cromartie coming into extensive use. They were, however, costly to install, so that the flat flame burner retained its popularity in spite of the fact that its duty was comparatively low, owing to the flame being drawn out into a thin sheet and so exposed to the cooling influence of the atmosphere. Almost at the same time that Murdoch was introducing the cockscomb and cockspur burners, he also made rough forms of Argand burner, consisting of two concentric pipes between which the gas was led and burnt with a circular flame. This form was soon improved by filling in the space between the tubes with a ring of metal, bored with fine holes so close together that the jets coalesced in burning and gave a more satisfactory flame, the air necessary to keep the flame steady and ensure complete combustion being obtained by the draught created by a chimney placed around it. When it began to be recognized that the temperature of the flame had a great effect upon the amount of light emitted, the iron tips, which had been universally employed, both in flat flame and Argand burners, were replaced by steatite or other non-conducting material of similar character, to prevent as far as possible heat from being withdrawn from the flame by conduction.

In 1880 the burners in use for coal-gas therefore consisted of flat flame, Argand, and regenerative burners, and the duty given by them with a 16-candle gas was as follows:--

Candle units
per cub. ft.
Burner. of gas.
Union jet flat flame, No. 0 0.59
" " 1 0.85
" " 2 1.22
" " 3 1.63
" " 4 1.74
" " 5 1.87
" " 6 2.15
" " 7 2.44
Ordinary Argand 2.90
Standard Argand 3.20
Regenerative 7 to 10

The luminosity of a coal-gas flame depends upon the number of carbon particles liberated within it, and the temperature to which they can be heated. Hence the light given by a flame of coal-gas can be augmented by (1) increasing the number of the carbon particles, and (2) raising the temperature to which they are exposed. The first process is carried out by enrichment (see GAS: _Manufacture_), the second is best obtained by regeneration, the action of which is limited by the power possessed by the material of which burners are composed to withstand the superheating. Although with a perfectly made regenerative burner it might be possible for a short time to get a duty as high as 16 candles per cubic foot from ordinary coal-gas, such a burner constructed of the ordinary materials would last only a few hours, so that for practical use and a reasonable life for the burner 10 candles per cubic foot was about the highest commercial duty that could be reckoned on. This limitation naturally caused inventors to search for methods by which the emission of light could be obtained from coal-gas otherwise than by the incandescence of the carbon particles contained within the flame itself. A coal-gas flame consumed in an atmospheric burner under the conditions necessary to develop its maximum heating power could be utilized to raise to incandescence particles having a higher emissivity for light than carbon. This led to the gradual evolution of incandescent gas lighting.

Incandescent gas light.

Long before the birth of the Welsbach mantle it had been known that when certain unburnable refractory substances were heated to a high temperature they emitted light, and Goldsworthy Gurney in 1826 showed that a cylinder of lime could be brought to a state of dazzling brilliancy by the flame of the oxy-hydrogen blowpipe, a fact which was utilized by Thomas Drummond shortly afterwards in connexion with the Ordnance Survey of Ireland. The mass of a lime cylinder is, however, relatively very considerable, and consequently an excessive amount of heat has to be brought to bear upon it, owing to radiation and conduction tending to dissipate the heat. This is seen by holding in the flame of an atmospheric burner a coil of thick platinum wire, the result being that the wire is heated to a dull red only. With wire of medium thickness a bright red heat is soon attained, and a thin wire glows with a vivid incandescence, and will even melt in certain parts of the flame. Attempts were accordingly made to reduce the mass of the material heated, and this form of lighting was tried in the streets of Paris, buttons of zirconia and magnesia being heated by an oxy-coal-gas flame, but the attempt was soon abandoned owing to the high cost and constant renewals needed. In 1835 W. H. Fox Talbot discovered that even the feeble flame of a spirit lamp is sufficient to heat lime to incandescence, provided the lime be in a sufficiently fine state of division. This condition he fulfilled by soaking blotting-paper in a solution of a calcium salt and then incinerating it. Up to 1848, when J. P. Gillard introduced the intermittent process of making water-gas, the spirit flame and oxy-hydrogen flame were alone free from carbon particles. Desiring to use the water-gas for lighting as well as heating purposes Gillard made a mantle of fine platinum gauze to fit over the flame, and for a time obtained excellent results, but after a few days the lighting value of the mantle fell away gradually until it became useless, owing to the wire becoming eroded on the surface by the flame gases. This idea has been revived at intervals, but the trouble of erosion has always led to failure.

The next important stage in the history of gas lighting was the discovery by R. W. von Bunsen about 1855 of the atmospheric burner, in which a non-luminous coal-gas flame is obtained by causing the coal-gas before its combustion to mix with a certain amount of air. This simple appliance has opened up for coal-gas a sphere of usefulness for heating purposes as important as its use for lighting. After the introduction of the atmospheric burner the idea of the incandescent mantle was revived early in the eighties by the Clamond basket and a resuscitation of the platinum mantle. The Clamond basket or mantle, as shown at the Crystal Palace exhibition of 1882-1883, consisted of a cone of threads of calcined magnesia. A mixture of magnesium hydrate and acetate, converted into a paste or cream by means of water, was pressed through holes in a plate so as to form threads, and these, after being moulded to the required shape, were ignited. The heat decomposed the acetate to form a luting material which glued the particles of magnesium oxide produced into a solid mass, whilst the hydrate gave off water and became oxide. The basket was supported with its apex downwards in a little platinum wire cage, and a mixture of coal-gas and air was driven into it under pressure from an inverted blowpipe burner above it.

The Welsbach mantle was suggested by the fact that Auer von Welsbach had been carrying out researches on the rare earths, with constant use of the spectroscope. Desiring to obtain a better effect than that produced by heating his material on a platinum wire, he immersed cotton in a solution of the metallic salt, and after burning off the organic matter found that a replica of the original thread, composed of the oxide of the metal, was left, and that it glowed brightly in the flame. From this he evolved the idea of utilizing a fabric of cotton soaked in a solution of a metallic salt for lighting purposes, and in 1885 he patented his first commercial mantle. The oxides used in these mantles were zirconia, lanthania, and yttria, but these were so fragile as to be practically useless, whilst the light they emitted was very poor. Later he found that the oxide of thorium--thoria--in conjunction with other rare earth oxides, not only increased the light-giving powers of the mantle, but added considerably to its strength, and the use of this oxide was protected by his 1886 patent. Even these mantles were very unsatisfactory until it was found that the purity of the oxides had a wonderful effect upon the amount of light, and finally came the great discovery that it was a trace of ceria in admixture with the thoria that gave the mantle the marvellous power of emitting light.

Certain factors limit the number of oxides that can be used in the
manufacture of an incandescent mantle. Atmospheric influences must not
have any action upon them, and they must be sufficiently refractory
not to melt or even soften to any extent at the temperature of the
flame; they must also be non-volatile, whilst the shrinkage during the
process of "burning off" must not be excessive. The following table
gives the light-emissivity from pure and commercial samples of the
oxides which most nearly conform to the above requirements; the effect
of impurity upon the lighting power will be seen to be most marked.

Pure. Commercial.
Metals--
Zirconia 1.5 3.1
Thoria 0.5 6.0
Earth metals--
Cerite earths--Ceria 0.4 0.9
Lanthania 6.0
Yttrite earths--Yttria 3.2
Erbia 0.6 1.7
Common earths--Chromium oxide 0.4 0.4
Alumina 0.6 0.6
Alkaline earth metals--
Baryta 3.3 3.3
Strontia 5.2 5.5
Magnesia 5.0 5.0

Of these oxides thoria, when tested for shrinkage, duration and
strength, stands pre-eminent. It is also possible to employ zirconia
and alumina. Zirconia has the drawback that in the hottest part of the
flame it is liable not only to shrinkage and semi-fusion, but also to
slow volatilization, and the same objections hold good with respect to
alumina. With thoria the shrinkage is smaller than with any other
known substance, and it possesses very high refractory powers.

The factor which gives thoria its pre-eminence as the basis of the
mantle is that in the conversion of thorium nitrate into thorium oxide
by heat, an enormous expansion takes place, the oxide occupying more
than ten times the volume of the nitrate. This means that the mass is
highly spongy, and contains an enormous number of little air-cells
which must render it an excellent non-conductor. A mantle made with
thoria alone gives practically no light. But the power of
light-emissivity is awakened by the addition of a small trace of
ceria; and careful experiment shows that as ceria is added to it
little by little, the light which the mantle emits grows greater and
greater, until the ratio of 99% of thoria and 1% of ceria is reached,
when the maximum illuminating effect is obtained. The further addition
of ceria causes gradual diminution of light, until, when with some 10%
of ceria has been added, the light given by the mantle is again almost
inappreciable. When cerium nitrate is converted by heat into cerium
oxide, the expansion which takes place is practically nil, the ceria
obtained from a gramme of the nitrate occupying about the same space
as the original nitrate. Thus, although by weight the ratio of ceria
to thoria is as 1:99, by volume it is only as 1:999.

Manufacture of mantles.

The most successful form of mantle is made by taking a cylinder of cotton net about 8 in. long, and soaking it in a solution of nitrates of the requisite metals until the microscopic fibres of the cotton are entirely filled with liquid. A longer soaking is not advantageous, as the acid nature of the liquid employed tends to weaken the fabric and render it more delicate to handle. The cotton is then wrung out to free it from the excess of liquid, and one end is sewn together with an asbestos thread, a loop of the same material or of thin platinum wire being fixed across the constricted portion to provide a support by which the mantle may be held by the carrying rod, which is either external to the mantle, or (as is most often the case) fixed centrally in the burner head. It is then ready for "burning off," a process in which the organic matter is removed and the nitrates are converted into oxides. The flame of an atmospheric burner is first applied to the constricted portion at the top of the mantle, whereupon the cotton gradually burns downwards, the shape of the mantle to a great extent depending on the regularity with which the combustion takes place. A certain amount of carbon is left behind after the flame has died out, and this is burnt off by the judicious application of a flame from an atmospheric blast burner to the interior. The action which takes place during the burning off is as follows: The cellulose tubes of the fibre are filled with the crystallized nitrates of the metals used, and as the cellulose burns the nitrates decompose, giving up oxygen and forming fusible nitrites, which in their semi-liquid condition are rendered coherent by the rapid expansion as the oxide forms. As the action continues the nitrites become oxides, losing their fusibility, so that by the time the organic matter has disappeared a coherent thread of oxide is left in place of the nitrate-laden thread of cotton. In the early days of incandescent lighting the mantles had to be sent out unburnt, as no process was known by which the burnt mantle could be rendered sufficiently strong to bear carriage. As the success of a mantle depends upon its fitting the flame, and as the burning off requires considerable skill, this was a great difficulty. Moreover the acid nature of the nitrates in the fibres rapidly rotted them, unless they had been subjected to the action of ammonia gas, which neutralized any excess of acid. It was discovered, however, that the burnt-off mantle could be temporarily strengthened by dipping it in collodion, a solution of soluble gun-cotton in ether and alcohol together with a little castor-oil or similar material to prevent excessive shrinkage when drying. When the mantle was removed from the solution a thin film of solid collodion was left on it, and this could be burned away when required.

After the Welsbach mantle had proved itself a commercial success many
attempts were made to evade the monopoly created under the patents,
and, although it was found impossible to get the same illuminating
power with anything but the mixture of 99% thoria and 1% ceria, many
ingenious processes were devised which resulted in at least one
improvement in mantle manufacture. One of the earliest attempts in
this direction was the "Sunlight" mantle, in which cotton was
saturated with the oxides of aluminium, chromium and zirconium, the
composition of the burnt-off mantle being:--

Alumina 86.88
Chromium oxide 8.68
Zirconia 4.44
------
100.00

The light given by these mantles was entirely dependent upon the
proportion of chromium oxides present, the alumina playing the part of
base in the same way that the thoria does in the Welsbach mantle, the
zirconia being added merely to strengthen the structure. These mantles
enjoyed considerable popularity owing to the yellowish pink light they
emitted, but, although they could give an initial illumination of 12
to 15 candles per foot of gas consumed, they rapidly lost their
light-giving power owing to the slow volatilization of the oxides of
chromium and aluminium.

Another method of making the mantle was first to produce a basis of
thoria, and, having got the fabric in thorium oxide, to coat it with a
mixture of 99% thoria and 1% ceria. This modification seems to give an
improvement in the initial amount of light given by the mantle. In the
Voelker mantle a basis of thoria was produced, and was then coated by
dipping in a substance termed by the patentee "Voelkerite," a body
made by fusing together a number of oxides in the electric furnace.
The fused mass was then dissolved in the strongest nitric acid, and
diluted with absolute alcohol to the necessary degree. A very good
mantle having great lasting power was thus produced. It was claimed
that the process of fusing the materials together in the electric
furnace altered the composition in some unexplained way, but the true
explanation is probably that all water of hydration was eliminated.

The "Daylight" mantle consisted of a basis of thoria or thoria mixed
with zirconia, dipped in collodion containing a salt of cerium in
solution; on burning off the collodion the ceria was left in a finely
divided condition on the surface of the thoria. In this way a very
high initial illuminating power was obtained, which, however, rapidly
fell as the ceria slowly volatilized.

Perhaps the most interesting development of the Welsbach process was
dependent upon the manufacture of filaments of soluble guncotton or
collodion as in the production of artificial silk. In general the
process consisted in forcing a thick solution of the nitrated
cellulose through capillary glass tubes, the bore of which was less
than the one-hundredth of a millimetre. Ten or twelve of the expressed
fibres were then twisted together and wound on a bobbin, the air of
the room being kept sufficiently heated to cause the drying of the
filaments a few inches from the orifice of the tube. The compound
thread was next denitrated to remove its extreme inflammability, and
for this purpose the skeins were dipped in a solution of (for
instance) ammonium sulphide, which converted them into ordinary
cellulose. After washing and drying the skeins were ready for the
weaving machines. In 1894 F. de Mare utilized collodion for the
manufacture of a mantle, adding the necessary salts to the collodion
before squeezing it into threads. O. Knöfler in 1895, and later on A.
Plaissetty, took out patents for the manufacture of mantles by a
similar process to De Mare's, the difference between the two being
that Knöfler used ammonium sulphide for the denitration of his fabric,
whilst Plaissetty employed calcium sulphide, the objection to which is
the trace of lime left in the material. Another method for making
artificial silk which has a considerable reputation is that known as
the Lehner process, which in its broad outlines somewhat resembles the
Chardonnet, but differs from it in that the excessively high pressures
used in the earlier method are done away with by using a solution of a
more liquid character, the thread being hardened by passing through
certain organic solutions. This form of silk lends itself perhaps
better to the carrying of the salts forming the incandescent oxides
than the previous solutions, and mantles made by this process, known
as Lehner mantles, showed promise of being a most important
development of De Mare's original idea. Mantles made by these
processes show that it is possible to obtain a very considerable
increase in life and light-emissivity, but mantles made on this
principle could not now be sold at a price which would enable them to
compete with mantles of the Welsbach type.

The cause of the superiority of these mantles having been realized,
developments in the required direction were made. The structure of the
cotton mantle differed widely from that obtained by the various
collodion processes, and this alteration in structure was mainly
responsible for the increase in life. Whereas the average of a large
number of Welsbach mantles tested only showed a useful life of 700 to
1000 hours, the collodion type would average about 1500 hours, some
mantles being burnt for an even longer period and still giving an
effective illumination. This being so, it was clear that one line of
advance would be found in obtaining some material which, whilst giving
a structure more nearly approaching that of the collodion mantle,
would be sufficiently cheap to compete with the Welsbach mantle, and
this was successfully done.

By the aid of the microscope the structure of the mantle can be
clearly defined, and in examining the Welsbach mantle before and after
burning, it will be noticed that the cotton thread is a closely
twisted and plaited rope of myriads of minute fibres, whilst the
collodion mantle is a bundle of separate filaments without plait or
heavy twisting, the number of such filaments varying with the process
by which it was made. This latter factor experiment showed to have a
certain influence on the useful light-giving life of the mantle, as
whereas the Knöfler and Plaissetty mantles had an average life of
about 1500 hours, the Lehner fabric, which contained a larger number
of finer threads, could often be burnt continuously for over 3000
hours, and at the end of that period gave a better light than most of
the Welsbach after as many hundred.

It is well known that plaiting gave the cotton candle-wick that power
of bending over, when freed from the binding effect of the candle
material and influenced by heat, which brought the tip out from the
side of the flame. This, by enabling the air to get at it and burn it
away, removed the nuisance of having to snuff the candle, which for
many centuries has rendered it a tiresome method of lighting. In the
cotton mantle, the tight twisting of the fibre brings this torsion
into play. When the cotton fibres saturated with the nitrates of the
rare metals are burnt off, and the conversion into oxides takes place,
as the cotton begins to burn, not only does the shrinkage of the mass
throw a strain on the oxide skeleton, but the last struggle of torsion
in the burning of the fibre tends towards disintegration of the
fragile mass, and this all plays a part in making the cotton mantle
inferior to the collodion type.

If ramie fibre be prepared in such a way as to remove from it all
traces of the glutinous coating, a silk-like fabric can be obtained
from it, and if still further prepared so as to improve its absorbent
powers, it can be formed into mantles having a life considerably
greater than is possessed by those of the cotton fabric. Ramie thus
seemed likely to yield a cheap competitor in length of endurance to
the collodion mantle, and results have justified this expectation. By
treating the fibre so as to remove the objections against its use for
mantle-making, and then making it into threads with the least possible
amount of twist, a mantle fabric can be made in every way superior to
that given by cotton.

The Plaissetty mantles, which as now manufactured also show a
considerable advance in life and light over the original Welsbach
mantles, are made by impregnating stockings of either cotton or ramie
with the nitrates of thorium and cerium in the usual way, and, before
burning off, mercerizing the mantle by steeping in ammonia solution,
which converts the nitrates into hydrates, and gives greater density
and strength to the finished mantle. The manufacturers of the
Plaissetty mantle have also made a modification in the process by
which the saturated fabric can be so prepared as to be easily burnt
off by the consumer on the burner on which it is to be used, in this
way doing away with the initial cost of burning off, shaping,
hardening and collodionizing.

Intensifying systems.

Since 1897 inventions have been patented for methods of intensifying the light produced by burning gas under a mantle and increasing the light generated per unit volume of gas. The systems have either been self-intensifying or have depended on supplying the gas (or gas and air) under an increased pressure. Of the self-intensifying systems those of Lucas and Scott-Snell have been the most successful. A careful study has been made by the inventor of the Lucas light of the influence of various sizes and shapes of chimneys in the production of draught. The specially formed chimney used exerts a suction on the gas flame and air, and the burner and mantle are so constructed as to take full advantage of the increased air supply, with the result that the candle power given by the mantle is considerably augmented. With the Scott-Snell system the results obtained are about the same as those given by the Lucas light, but in this case the waste heat from the burner is caused to operate a plunger working in the crown of the lamp which sucks and delivers gas to the burner. Both these systems are widely used for public lighting in many large towns of the United Kingdom and the continent of Europe.

The other method of obtaining high light-power from incandescent gas burners necessitates the use of some form of motive power in order to place the gas, or both gas and air, under an increased pressure. The gas compressor is worked by a water motor, hot air or gas engine; a low pressure water motor may be efficiently driven by water from the main, but with large installations it is more economical to drive the compressor by a gas engine. To overcome the intermittent flow of gas caused by the stroke of the engine, a regulator on the floating bell principle is placed after the compressor; the pressure of gas in the apparatus governs automatically the flow of gas to the engine. With the Sugg apparatus for high power lighting the gas is brought from the district pressure, which is equal to about 2½ in. of water, to an average of 12 in. water pressure. The light obtained by this system when the gas pressure is 9½ in. is 300 candle power with an hourly consumption of 10 cub. ft. of gas, equivalent to 30 candles per cubic foot, and with a gas pressure equal to 14 in. of water 400 candles are obtained with an hourly consumption of 12½ cub. ft., which represents a duty of 32 candles per cubic foot of gas consumed. High pressure incandescent lighting makes it possible to burn a far larger volume of gas in a given time under a mantle than is the case with low pressure lighting, so as to create centres of high total illuminating value to compete with arc lighting in the illumination of large spaces, and the Lucas, Keith, Scott-Snell, Millennium, Selas, and many other pressure systems answer most admirably for this purpose.

Inverted burners.

The light given by the ordinary incandescent mantle burning in an upright position tends rather to the upward direction, because owing to the slightly conical shape of the mantle the maximum light is emitted at an angle a little above the horizontal. Inasmuch as for working purposes the surface that a mantle illuminates is at angles below 45° from the horizontal, it is evident that a considerable loss of efficient lighting is brought about, whilst directly under the light the burner and fittings throw a strong shadow. To avoid this trouble attempts have from time to time been made to produce inverted burners which should heat a mantle suspended below the mouth of the burner. As early as 1882 Clamond made what was practically an inverted gas and air blowpipe to use with his incandescent basket, but it was not until 1900-1901 that the inverted mantle became a possibility. Although there was a strong prejudice against it at first, as soon as a really satisfactory burner was introduced, its success was quickly placed beyond doubt. The inverted mantle has now proved itself one of the chief factors in the enormous success achieved by incandescent mantle lighting, as the illumination given by it is far more efficient than with the upright mantle, and it also lends itself well to ornamental treatment.

Burners.

When the incandescent mantle was first introduced in 1886 an ordinary laboratory Bunsen burner was experimentally employed, but unless a very narrow mantle just fitting the top of the tube was used the flame could not be got to fit the mantle, and it was only the extreme outer edge of the flame which endowed the mantle fabric with the high incandescent. A wide burner top was then placed on the Bunsen tube so as to spread the flame, and a larger mantle became possible, but it was then found that the slowing down of the rate of flow at the mouth of the burner owing to its enlargement caused flashing or firing back, and to prevent this a wire gauze covering was fitted to the burner head; and in this way the 1886-1887 commercial Welsbach burner was produced. The length of the Bunsen tube, however, made an unsightly fitting, so it was shortened, and the burner head made to slip over it, whilst an external lighting back plate was added. The form of the "C" burner thus arrived at has undergone no important further change. When later on it was desired to make incandescent mantle burners that should not need the aid of a chimney to increase the air supply, the long Bunsen tube was reverted to, and the Kern, Bandsept, and other burners of this class all have a greater total length than the ordinary burners. To secure proper mixing of the air and gas, and to prevent flashing back, they all have heads fitted with baffles, perforations, gauze, and other devices which oppose considerable resistance to the flow of the stream of air and gas.

In 1900, therefore, two classes of burner were in commercial existence for incandescent lighting--(1) the short burner with chimney, and (2) the long burner without chimney. Both classes had the burner mouth closed with gauze or similar device, and both needed as an essential that the mantle should fit closely to the burner head.

Prior to 1900 attempts had been made to construct a burner in which an
incandescent mantle should be suspended head downwards. Inventors all
turned to the overhead regenerative gas lamps of the Wenham type, or
the inverted blowpipe used by Clamond, and in attempting to make an
inverted Bunsen employed either artificial pressure to the gas or the
air, or to both, or else enclosed the burner and mantle in a globe,
and by means of a long chimney created a strong draught. These burners
also were all regenerative and aimed at heating the air or gas or
mixture of the two, and they had the further drawback of being
complicated and costly. Regeneration is a valuable adjunct in ordinary
gas lighting as it increases the actions that liberate the carbon
particles upon which the luminosity of a flame is dependent, and also
increases the temperature; but with the mixture of air and gas in a
Bunsen regeneration is not a great gain when low and is a drawback
when intense, because incipient combination is induced between the
oxygen of the air and the coal-gas before the burner head is reached,
the proportions of air and gas are disturbed, and the flame instead of
being non-luminous shows slight luminosity and tends to blacken the
mantle. The only early attempt to burn a mantle in an inverted
position without regeneration or artificial pressure or draught was
made by H. A. Kent in 1897, and he used, not an inverted Bunsen, but
one with the top elongated and turned over to form a siphon, so that
the point of admixture of air and gas was below the level of the
burner head, and was therefore kept cool and away from the products of
combustion.

In 1900 J. Bernt and E. Cérvenka set themselves to solve the problem of making a Bunsen burner which should consume gas under ordinary gas pressure in an inverted mantle. They took the short Bunsen burner, as found in the most commonly used upright incandescent burners, and fitted to it a long tube, preferably of non-conducting material, which they called an isolator, and which is designed to keep the flame at a distance from the Bunsen. They found that it burnt fairly well, and that the tendency of the flame to burn or lap back was lessened, but that the hot up-current of heated air and products of combustion streamed up to the air holes of the Bunsen, and by contaminating the air supply caused the flame to pulsate. They then fixed an inverted cone on the isolator to throw the products of combustion outwards and away from the air holes, and found that the addition of this "deflecting cone" steadied the flame. Having obtained a satisfactory flame, they attacked the problem of the burner head. Experiments showed that the burner head must be not only open but also of the same size or smaller than the burner tube, and that by projecting it downwards into the mantle and leaving a space between the mantle and the burner head the maximum mantle surface heated to incandescence was obtained. It was also found that the distance which the burner head projects into the mantle is equivalent to the same amount of extra water pressure on the gas, and with a long mantle it was found useful under certain conditions to add a cylinder or sleeve with perforated sides to carry the gas still lower into the mantle. The principles thus set forth by Kent, Bernt and Cérvenka form the basis of construction of all the types of inverted mantle burners which so greatly increased the popularity of incandescent gas lighting at the beginning of the 20th century, whilst improvements in the shape of the mantle for inverted lighting and the methods of attachment to the burner have added to the success achieved.

The wonderful increase in the amount of light that can be obtained from gas by the aid of the incandescent gas mantle is realized when one compares the 1 to 3.2 candles per cubic foot given by the burners used in the middle of the 19th century with the duty of incandescent burners, as shown in the following table:--

_Light yielded per cubic foot of Gas._

Burner. Candle power.
Low pressure upright incandescent burners 15 to 20 candles
Inverted burners 14 to 21 "
Kern burners 20 to 24 "
High pressure burners 22 to 36 "

(V. B. L.)

3. ELECTRIC LIGHTING.

Electric lamps are of two varieties: (1) _Arc Lamps_ and (2) _Incandescent_ or _Glow Lamps_. Under these headings we may briefly consider the history, physical principles, and present practice of the art of electric lighting.

1. _Arc Lamps._--If a voltaic battery of a large number of cells has its terminal wires provided with rods of electrically-conducting carbon, and these are brought in contact and then slightly separated, a form of electric discharge takes place between them called the _electric arc_. It is not quite certain who first observed this effect of the electric current. The statement that Sir Humphry Davy, in 1801, first produced and studied the phenomenon is probably correct. In 1808 Davy had provided for him at the Royal Institution a battery of 2000 cells, with which he exhibited the electric arc on a large scale.

The electric arc may be produced between any conducting materials maintained at different potentials, provided that the source of electric supply is able to furnish a sufficiently large current; but for illuminating purposes pieces of hard graphitic carbon are most convenient. If some source of continuous electric current is connected to rods of such carbon, first brought into contact and then slightly separated, the following facts may be noticed: With a low electromotive force of about 50 or 60 volts no discharge takes place until the carbons are in actual contact, unless the insulation of the air is broken down by the passage of a small electric spark. When this occurs, the space between the carbons is filled at once with a flame or luminous vapour, and the carbons themselves become highly incandescent at their extremities. If they are horizontal the flame takes the form of an arch springing between their tips; hence the name _arc_. This varies somewhat in appearance according to the nature of the current, whether continuous or alternating, and according as it is formed in the open air or in an enclosed space to which free access of oxygen is prevented. Electric arcs between metal surfaces differ greatly in colour according to the nature of the metal. When formed by an alternating current of high electromotive force they resemble a lambent flame, flickering and producing a somewhat shrill humming sound.