of the flame as to innumerable points of high temperature created on the surface of the mantle by the catalytic action of the ceria upon the hydrogen and carbon monoxide, when once the temperature necessary to carry this on in the best way is reached, then the rise in temperature due to the combustion of a gas of high calorific value in the inner zone of the flame would merely increase the glow due to the effect of the slight extra heat on the mantle fabric. How small this is can be judged from the light emitted by a pure thoria mantle heated by the combustion of different grades of gas, and experiment shows that this is about the difference found in the experiment recorded with the 17.7 and 14 candle gas.

The results upon which Messrs. White, Russell, and Travers based their views on the effect of the calorific value of the gas on the mantle may be arranged in order in the following Table, interpolating the value in B.T.U.'s as well as calories::

COMPARISON OF THE HEATING POWER OF GASES WITH THE LIGHT GIVEN BY THE INCANDESCENT MANTLE.

The fact of 5.53 being put forward as what may be expected from blue water gas under the best conditions of burning, makes one feel disinclined to put any great degree of confidence in Messrs. White, Russell and Travers' results, which are manifestly wrong.

A point which cannot be too strongly insisted upon is that the mesh of the mantle and the size of the opening at the top are important factors in the light emitted by different grades of gas, and should be attended to with the same care as the degree of aeration. Experiments soon show that the mantle best fitted for use with a rich carburetted water gas, such as is used in America, would give very poor results with a 10 or 12 candle gas, whilst the gas ejector nipple suited to a rich gas is useless with a low grade gas. It may be neglect of such conditions as these that has led to such manifestly erroneous deductions.

The result of Herr Oechelhaeuser's experiment with Dessau gas, which showed that a reduction in candle power, as tested with a batswing burner, of from 13 to less than 2 candles, caused an increase of from 66.5 to 73.5 candles with the Welsbach mantle, may have been partly due to the conditions of air supply to the burner being better fitted for the combustion of the lower grade of gas; but there is another factor that must not be lost sight of, as it has an important bearing on the use of blue water gas as a diluent for coal gas.

It is well known that the increase in volume of hydrogen and carbon monoxide, caused by diluting coal gas with blue water gas, shortens the flame and increases the intensity, and in this way makes the light of the flame whiter, due to the higher incandescence of the carbon particles. With the incandescent mantle the same thing is noticeable, and I have many times obtained the same result as Herr Oechelhaeuser.

In the case of the Dessau gas, the lowering in candle power was brought about by "stripping," i.e., removing absorbable hydrocarbons, and this would have the same effect, as it increased the percentage of hydrogen present, and either this gas or carbon monoxide have the same shortening effect on the flame as the mixture of the two (water gas).

An even more accentuated case of the same character came under my notice last year. I was testing the gas in a northern town, and found that in the standard flat flame burner it gave an illuminating value of 20 candles. The gas was made from coal, and had an illuminating value of 15 to 14 candles before enrich

ment, and it was brought up to 20 candle gas by benzol vapour. I was struck by the poor light given by the mixture in the incandescent lamps of the town, and on testing it with mantles burnt both on the ordinary 66 "C" and Kern burners (No. 4), could only get a total illuminating value of 46 to 50 candles. I then tested the unenriched gas, which at that time had an illuminating value in the London argand of 14.11 candles, with the same mantles and burners as before, and got 80 to 90 candles without any trouble. The 20 candle benzolised gas had a calorific value of 608-8 nett B.T.U.'s, the 141 candle coal gas 510 B.T.U.'s.

I could only assume that benzolised coal gas was unfitted for use with incandescent burners with ordinary adjustments I repeated the experiment with the Kern burner, carefully measuring the rate of flow, and also trying the same coal gas enriched with petroleum spirit to 21.5 candle power.

KERN BURNER No. 4 WITH MANTLE.

1 and 4 inches pressure, this increase is more apparent than real. Mr. Walter Grafton last year read a most interesting paper before the Gas Institute on the effect of quality and pressure on the efficiency of incandescent gas lighting, aud gave a table of the results obtained with various grades of coal gas and coal gas enriched by carburetted water gas, as well as a number of results obtained from mixtures of these gases with air.

Taking his results for the coal gas and enriched gas only, and calculating the results at 1, 2, and 3 inches pressure to candles per cubic foot, the following Table is obtained

34

10.6

13.3

spirit 21.5 1.5 Benzol.... 20.0 1.5 3:45 46 None 14'1 1.5 4:45 90

20.2

In all probability the explanation of these results is to be found in the fact that although the air holes of the burners were full open, they had been made for a 16 candle coal gas, and gave too small an aeration for the enriched mixtures, and that if the nipples of the gas injectors had been changed more normal results would have been obtained with the rich gas. I give these results, not as showing that enrichment reduces the lighting value that can be obtained from a mantle, but rather that the consumer is more likely to get his money's worth with a poor gas than with a doctored up high candle value gas, as in nine cases out of ten the burner vendors never take the trouble to adjust the burners to the quality of gas with which they are to be used.

Another question, which is an important one, is the effect of the pressure under which the gas is supplied to the burner upon the light emitted by the mantle. When the pressure of the gas rises the visible effect is an increase of light from the mantle, but within the ordinary limits of gas supply, i.e., between

This shows that the general tendency of such variations of pressure is to decrease the candles per cubic foot of gas consumed and that any increase of the total candle power is due to a very largely increased consumption.

When, however, high pressures are used, i.e., pressures of 8 to 10 inches, such as are employed in high intensity lighting on the Sugg, Keith, or other systems, the high pressure of the gas enables some 10 cubic feet of gas to be consumed per hour in place of the 3 cubic feet supposed to be used by the "C" burner, so that nearly three times the amount of gas is burnt in the same time and area, and results in a great increase in calorific intensity, with the result that over 30 candles per cubic foot can be obtained with London gas, and, of course, higher powers with a richer gas.

I have long since pointed out the importance of a thorough mixing of the gas and the air drawn in by the atmospheric burner in increasing the light given by the mantle, and it was to do this that Chemin, Denayrouse, Bandsept, and Kern devised the special forms of burners associated with their names.

The late Mr. Frank Livesey also showed that by allowing the gas to take in the needed air

some distance away from the burner head, so that a length of travel should bring about a more thorough mixing, an enhanced effect could be produced. In the paper by Mr. Grafton, already alluded to, he shows by carefully made photometric experiments that by making a mixture containing 15 to 30 per cent. of air with the coal gas after leaving the gas meter, the light emitted by the mantle is increased to such an extent that it far exceeds the ordinary results obtained for the same consumption of gas by other methods.

This method of burning gas for incandescence is in practical use on the Continent, and is giving such excellent results that its use is rapidly spreading.

It was as early as 1894 that Daus took out a patent for the production of an intensely hot flame by the use of a mixture of gas and air, made by a suction air drum driven by the gas meter, which secured a constant mixture of the gas and air in the required proportions, and used the flame for heating Fahnehjelm combs. Later on Doller, Fuller, and finally the "Selas" Company took out further patents on improvements of the same idea, and under these combined patents have made a great success, the last installation being in the Thiergarten in Berlin, where the results attained are said to be exceedingly good.

The so-called " Selas" system consists of a small apparatus for the mixing of two volumes of air with one of gas immediately after leaving the consumer's meter, and it is worked by a small water motor using 12 litres or a little over a quarter of a gallon per burner per hour, or if desired it may be worked very cheaply by an electric motor.

The results obtained by this method, which is intended for use in the house, are that with a consumption of 3.53 cubic feet of coal gas in admixture with air a light of 134 English candles is obtained, whilst with a consumption of 424 of gas the light rises to 167.9 candles, or 39'5 per foot.

These results are obtained with Berlin, Stockholm, and Copenhagen gas, the illuminating power of which varies from 14 to 15 candles, whilst with a high pressure system on the same lines, but with a compressor as well as a mixer, over 45 candles per foot are obtained with the same low power gas. I have not tested the results myself, but the fact that they have been obtained by Mr. F. D. Marshall at the Copenhagen works is a sufficient guarantee of their correctness, and I hope Mr. Marshall will be able to give a

paper and demonstration of these remarkable results at the next meeting of the combined Institutions.

A very interesting phase of high pressure illuminations which tends in the same direction, is to be found in some experiments made by Mr. R. G. Shadbolt on the effect of using air under pressure to suck in coal gas instead of the gas sucking in air. In order to do this, he removed the gas nipple from an ordinary injector burner, and supplied air instead of gas through it, and sucked in the gas through the side holes. He found that by passing air in at a pressure of 32 tenths (3.2 inches) he got a consumption of 18 cubic feet of 17 to 17.5 candle gas, and a duty from the mantle of 678.5 candle, or 37.69 candles per foot.

High pressure incandescent lighting is a means to attain a definite end, and that end is to create centres of high total illuminating value in order to compete with arc lighting for out-door work and large buildings. There is no doubt that the amount of light which can be obtained from the surface of a single mantle can be enormously increased thereby, but I do not think carefully made tests would reveal so great an economy in the units of light obtained per cubic foot of gas consumed as one would expect. I frankly admit that that the high pressure incandescent companies have not shown any burning desire to submit their systems to my tender mercies, so that I have no personal knowledge of the actual results they are obtaining, or the cost at which it is being done, but Continental practice has gone so far ahead of ours, that one can take a comparison of the candle units per cubic foot of gas consumed.

Probably the greatest volume of light obtained by high pressure incandescent gas lighting is in the "Millenium burners, which, with a double mantle, give a light of 1,800 Hefner candles, obtained by a consumption of 1,566 litres of Berlin gas, i.e., 28.5 candles per cubic foot.

In the Millenium burner, I understand, the gas is compressed under a pressure of over four feet of water, but one knows that over 30 candles a foot can be obtained with Berlin gas at lower pressures, such as 10 inches by the forms of apparatus in use here, but then the total volume of light is much smaller.

The value of mixing the coal gas with some of the air before compression, so as to get more thorough admixture, and then using this to suck in the extra air required is also clear, as Mr. Marshall found with the

"Selas high-pressure system, that with Berlin gas and using a pressure of 24 inches of water he could obtain a light of 766 candles with a consumption of 16.7 cubic feet of gas, or 45.8 candles per cubic foot: and it is clear that by using lights obtained under this system over a large area, a great economy as well as a far better illuminating effect would be produced, as compared with half the number of lamps of double the candle power obtained by merely using gas compressed to 6 feet.

In considering the economy and utility of high pressure lighting, it must be borne in mind that, although it may be politic and necessary to use what are practically blowpipes rather than burners in order to create light which in volume shall compete with the electric arc, yet up to the present the mantle has not been made that will stand such usage for any great length of time. It is claimed for the Selas method of lighting that the life of the mantle is enhanced by 10 per cent. instead of being shortened, and if this be so, it would certainly accentuate its claim to rank before all other systems.

After all is said and done, however, the great and most important factor in incandescent lighting is the mantle, and we must now turn to that portion of the subject and see what improvements have been made and what improvements are likely to be made, as upon this so much depends.

Since the introduction of the present composition of 99 per cent. thoria and 1 per cent. ceria it is safe to say that no improvement has been made in the composition itself, and in this country but little improvement in the physical condition of the mantle, and although the introduction of the beautiful artificial silk processes of mantle making of Knofler, Plaissetty, and Lehner showed that it was possible to obtain a very considerable increase in the life and light emissivity of the mantle, yet with the cheapening of mantles that has taken place during the past year, it would not be possible for a mantle made on this principle to be sold at a price which would compete with mantles of the Welsbach type, such as now can be bought at 3d., or even less.

In a previous course of Cantor lectures I showed that the structure of the cotton mantle differed widely from that obtained by the various collodion processes, and that it was this alteration in structure that was mainly responsible for the increase in life. Whereas the average of a large number of Welsbach mantles tested, only showed a useful

some

life of 700 to 1,000 hours, the collodion type would average about 1,500 hours, 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 at the same time be sufficiently cheap to be able to compete in the open market with the Welsbach mantle, and this I think has now been successfully done.

By the aid of photo-micrography the structure of the mantle can be clearly defined, and on 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 Knofler and Plaisetty mantles had an average life of about 1,500 hours, the Lehner fabric, which contained a larger number of finer collodion threads, could often be burnt continuously for over 3,000 hours, and at the end of that period gave a better light than most of the Welsbachs after as many hundred.

It is well known that the plaiting of the cotton candle wick gave it that power of bending over when freed from the binding influence of the candle material and influenced by heat, that brought the tip of the wick 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 had rendered it a tiresome metho! of lighting. In the cotton mantle, the tight twisting of the fibre brings this principle of 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 I think this all plays a part in making the cotton mantles inferior to their collodion brethren.

It has long been known that if ramie fibre be prepared in such a way as to remove from it all traces of the glutinous coating, a beautiful