When water meets calcium carbide, acetylene is released. The gas is colourless, and pure acetylene has almost no smell. The stench everybody associates with these lamps comes from impurities in the carbide. Makers sold deodorised carbide, and users complained that it clogged the gas pipe and the burner tip with a sticky residue.

The chemistry was known long before the lamp. Edmund Davy described producing the gas by dripping water onto carbide in 1836. The French chemist Marcellin Berthelot studied it and named it in the early 1860s. What was missing was cheap carbide.

That came in 1892. The Canadian Thomas Willson found a production method during a failed attempt to make aluminium. The material from his furnace was dropped into a pail of water to cool, and it fizzed and gave off a gas that burned very brightly. French sources credit Henri Moissan instead. Moissan worked out a simple and inexpensive preparation in 1894, and received the Nobel Prize in chemistry in 1906. Priority was disputed at the time, and the two accounts have never been reconciled.

The light was about twelve times as bright as ordinary coal gas, and was widely advertised as artificial daylight. Montgomery Ward rated the Solar at 100 candle power in 1916 and claimed a beam of 100 to 150 feet. In practice a bicycle lamp burned for five to six hours and threw light some thirty to fifty metres. It depended on the burner fitted and on how hard it was run.

Burners were sold separately and rated by gas consumption. Riemann of Chemnitz listed his in 1911 at 5, 8, 15, 20, 25 and 35 litres an hour.

This is not settled. A claim that E. H. Fuller made one in 1895 under the Illinois Acetylene Company is widely repeated. It has never been substantiated, and the standard works say so themselves.

The first commercially successful lamp was the American Solar. The accounts of that one do not line up either. One puts it in late 1896 with the Solar Acetylene Lamp Company, later renamed Badger Brass of Kenosha. Another gives 1897 and places the Solar Acetylene Lamp Company in Chicago. The lamps themselves carry patent dates of February 1896. All three versions stand here. No source among them gives a document that settles it.

After that the format spread quickly, with large numbers of makers in France, Germany and England.

Montgomery Ward sold carbide on the same page as the lamps in 1916, as No. 61B655, in 2 lb cans at 18 cents a can. That is nine cents a pound, against $1.68 for the cheapest acetylene lamp on that page.

Carbide could not be sent by parcel post, and the catalogue said so in the entry. The lamp arrived by mail. The fuel had to come from a local dealer or by freight.

Riemann's price list of 1911 gives the other half of the sum. He states the carbide capacity and the burning time for each model. The Phänomen motorcycle lamp took about 90 grams and ran three and a half hours. The Pfadfinder took about 55 grams and ran two hours. The largest carriage lantern took about 340 grams and ran seven hours.

Using one properly took some care. Open the water valve too far and you got a flame you could not manage. Twenty to thirty drops a minute gave a steady flame of the right size. From about 1902 carbide was sold with a thin paraffin coating that slowed the reaction down.

After 1900 most makers fitted a horsehair filter inside the upper part of the generator, to catch impurities before they reached the burner. Restorers often take it out, and the burner then blocks within a few rides. Cotton wool works as a replacement.

Using one also took planning. You had to judge in advance how long the ride would last and how much carbide it needed. At home the lamp came off the bicycle and went outside, so the carbide left in it could gas off into the air. Spent carbide reacted, baked solid and stuck to the housing, so the lamp had to be cleaned after every ride.

Makers later sold pre-packaged carbide cartridges for rides of different lengths. A paper wrapper let you lift the slurry out cleanly, even when it was wet. A blocked burner bore could be cleared with an ordinary bicycle pump.

There was a real hazard as well. Gas that had built up could be set off by a flame nearby, and there are reports of copper acetylene lamps exploding when struck during use.

Not everyone was won over. Acetylene lamps dazzled oncoming traffic. The organisers of the December 1898 shows in London refused to allow them to be demonstrated, citing the risk of bangs and fumes.

Price went against the format later on. On the Montgomery Ward page of 1916 the acetylene lamps ran from $1.68 to $2.40. The cheapest electric outfit on the same page was 90 cents without a battery. The acetylene lamp stayed in ordinary use until the Second World War.

Riemann warned his customers at length in 1911 against imitation push-fit burners that did not seat properly in his lanterns. All his original burners carry the stamp H.R. CH.-G. D.R.G.M.SCH. A burner with that stamp is a Riemann part.

What acetylene displaced is covered under bicycle oil lamps and bicycle candle lamps. What displaced acetylene in turn is under battery and dynamo bicycle lamps. For brackets, glass colours and dating marks, see identifying and dating an antique bicycle lamp.

Primary: Montgomery Ward catalogue, 1916; Herm. Riemann illustrated price list, 23rd edition, Chemnitz-Gablenz, February 1911; Sears, Roebuck and Co. Consumers Guide, catalogue no. 104, spring 1897; lamps in my own collection.

Secondary: American Chemical Society, National Historic Chemical Landmark on the commercialisation of calcium carbide and acetylene, used for the chemistry and its dating; Peter W. Card, Early Cycle Lighting 1868–1948; Tony Hadland and Hans-Erhard Lessing, Bicycle Design: An Illustrated History; Bernard Mahot, Les lanternes, phares et fanaux. Mahot's account of the discovery of acetylene gives Humphry Davy and 1806, Berthelot and 1863, and the name Léopold Wilson. The chemical literature gives Edmund Davy and 1836, the early 1860s, and Thomas Willson. He has not been followed on those points.