At Halloween, it has become traditional to carve pumpkins into grotesque faces lit from within by candles, known as “jack-o-lanterns”. However the carved pumpkins associated with Halloween take their name from a mysterious, and nowadays seldom seen, natural phenomenon which has so far defied all attempts at a satisfactory explanation. Also known as will-o’-the-wisp, ignis fatuus (“foolish fire”) and corpse candles, it is an eerie, flickering flame or light seen hovering over marshy ground at night. In times past these were thought to be malevolent spirits intent on leading lonely travellers astray, but at least since the 19th century, attempts have been made to find a rational explanation.
A number of mysterious light phenomena are reported from time to time, including ball lightning, associated with thunderstorms, and “earthlights”, glowing spheres which can appear in any conditions and seem to be electrical discharges associated with piezoelectric activity in pressurised rocks. The jack-o’-lantern is quite distinct from these, taking the form of a relatively stationary flame or diffuse light, usually bluish in colour, and generally seen in marshy areas. It is said to recede when approached, giving the impression that it is leading the observer on, perhaps into danger on the marshy ground.
In contrast to other light phenomena, jack-o’-lantern is seldom reported nowadays, probably due to widespread draining of marshes, light pollution and increasingly urbanised populations, but there are numerous eyewitness accounts from the past. Unfortunately, many of these come from uneducated or superstitious people who were perhaps too frightened to investigate or take note of any details which might help towards an explanation. However, two relatively detailed reports have survived from the 19th century; interestingly they appear to contradict one another in at least one important respect. They are summarised below:
1. Recorded: L Blesson, Engineer, Forest of Gorbitz, Germany, 1833
The phenomenon was observed over several days and nights in a marshy valley. During the day, bubbles were seen rising from the water and at night blue or bluish purple flames were seen. Blesson relates that the flames receded as he moved towards them, but returned to their original locations when he was still, attributing this behaviour to air currents caused by his movements. He also describes how he was able, with difficulty (due to his breathing disturbing the flame), to singe and set light to paper, describing the singed paper as being “covered with a viscous moisture”. On a subsequent visit he describes a small explosion taking place. Blesson describes the water in the marsh as “ferruginous” (rust coloured or rich in iron) and the surface as having an iridescent crust (possibly bacterial colonies).
2. Recorded: E Knorr, Physicist, Herzeberg-Elster, Germany, 1853
Knorr observed a 12cm flame rising from a swampy stream by a roadside. The flame was yellow in the centre, but an intense violet at the edges and bright enough to illuminate the surrounding vegetation. He describes holding the brass ferrule of his walking stick in the flame for approximately fifteen minutes and finding it still cool when he withdrew it.
The first account suggests that the flame is due to the combustion of gasses released by the decay of organic matter in the low oxygen conditions of marshes and bogs, and this forms the basis of most attempts at an explanation. Methane, which is flammable and burns with a blue flame, is often produced in these environments. Hydrogen sulphide may also be produced in smaller quantities and also burns with a blue flame. However, it is not clear how the gasses might be ignited. One possibility is a mixture of two phosphorus-hydrogen compounds, phosphine and diphosphine, which ignites on contact with air. At first sight, this seems plausible: waterlogged, marshy areas create a low oxygen “reducing” environment in which elements present in organic matter tend to form compounds with hydrogen rather than oxygen. Thus carbon will form methane, while sulphur and nitrogen form hydrogen sulphide and ammonia, respectively. It seems reasonable, therefore, that phosphorus, which is present in significant quantities in organic matter, might combine with hydrogen to form phosphine and diphosphine. However, there is a problem.
The phosphorus in organic matter is in the form of phosphates, which are extremely stable and difficult to break down. There is no known chemical process which will generate phosphine/diphosphine directly from phosphates; normally it would be necessary to first reduce the phosphate to elemental phosphorus (a process requiring sustained high temperatures) then heat the product with a strong alkali. Because of this difficulty, the phosphine/diphosphine theory was initially dismissed as impossible by many chemists.
Nevertheless, nature can sometimes find a way. Production of phosphine and diphosphine from phosphates by microorganisms has now been confirmed in the laboratory, and these compounds have been found in the digestive tracts of various mammals, including humans. It has even been speculated that they could be involved in some cases of alleged spontaneous human combustion! Exactly how these microorganisms produce phosphine/diphosphine remains a mystery, but there is a precedent for nature finding a low energy way of achieving something which normally requires high energies: nitrogen fixation. Nitrogen molecules in the atmosphere consist of two nitrogen atoms held together by a very strong triple bond; for this reason, nitrogen will normally only combine with other elements when subjected to high temperatures which break this bond. However, certain soil bacteria are able to capture atmospheric nitrogen and combine it with hydrogen to form molecules which are essential to life. The precise chemical mechanism for nitrogen fixation was only fully unravelled relatively recently. Possibly something similar happens with phosphates. Another possibility is the release of phosphine from the action of microbially produced chemicals on iron compounds which contain phosphorus; this also has been demonstrated in the laboratory.
How does all this square with actual accounts of the jack-o’-lantern? The “viscous” deposit on the singed paper mentioned in Bresson’s account is consistent with the combustion of phosphine: this would produce a form of phosphoric acid, which is indeed viscous. He also mentions the “ferruginous” appearance of the water, indicating high iron content. However the seemingly cold flame observed by Knorr (and mentioned in some other accounts) appears to indicate some form of chemiluminescence rather than combustion. Laboratory experiments have created a cold, pale greenish luminescence, accompanied by white smoke, by introducing small amounts of phosphine into methane mixed with various other, non-combustible gasses, however these results seem to differ from most reports of jack-o’-lantern phenomena, both in the flame colour and in the fact that smoke is not generally mentioned.
So the mystery remains. Combustion and/or chemiluminescence of marsh gasses seems to be involved in some way, but clearly, while research into this phenomenon has provided us with some clues, we have yet to arrive at a complete explanation.