Too Many Definitions of Candle

I have been doing some reading about photometry lately and I noticed that the unit of lighting called the candle has had quite a history. I used to work for a metrology company and I have always been interested in how standards and units are established. The candle has had a more volatile history than I am used to seeing. I thought I would document some of the history here to see how much the unit has changed over time. In fact, the candle has now been replaced in most situations by the candela. However, you still see flashlights rated in "candlepower".

Table 1 summarizes a few of the early versions of the candle that ran into while doing a bit of googling. I became curious about the candle when I saw a number of forum chats that were struggling with determining the proper conversion factors for the different types of candles. As I looked into the matter, it seemed the early experimenters had trouble with these units as well.

Table 1: Definitions of Candle that I Encountered During a Google Search
Unit Name Definition Definition Source
Candela (1948) The candela is the luminous intensity, in a given direction, of a source that emits monochromatic radiation of frequency 540E12 hertz (555 nm wavelength in a vacuum) and that has a radiant intensity in that direction of 1⁄683 watt per steradian. Candela
New Candle (1946) The luminous intensity of a square centimeter of a blackbody radiator at the temperature at which molten platinum solidifies as 60 new candles. New Candle
International Candle (1909) The International Candle was replaced by the New Candle. The International Candle is equivalent to 58.9 international candles. International Candle
US Candle (Date Unknown)Spermaceti candle burning 120 grains per hour. Weights, Measures Dict.
British Candle (1860) A spermaceti candle of 1/6 lb at 2 grains per minute. Electrical Age
Weights, Measures Dict.
Carcel candle (before 1882) A standard Carcel lamp burning colza oil at the standard rate and producing a standard flame. Carcel candle

Decimal Candle (1889) Candle that burns 8.5 grams of wax per hour. It put out the one-tenth the light of a Carcel candle, which I have found little information on. Decimal Candle
Hefner Candle (1884 -- Germany) Burns amyl acetate. Flame height of 40 mm, with a very specifically defined wick. Hefner Candle

Figure 1 shows some unit conversions that I put together. Note that these are not solid conversions. The early candles were very poorly defined and the early experimenters appear to have had a difficult time coming up with a consistently reproducible standard.

Figure 1: Summary of Candle Unit Conversions.

Figure 1: Summary of Candle Unit Conversions.

I am including a link to a web page that has a good set of luminosity conversions.

This entry was posted in General Science, History of Science and Technology. Bookmark the permalink.

2 Responses to Too Many Definitions of Candle

  1. Elvira says:

    I was also doing some research about history of candela and found the same things you have written. But one thing is still making me curious and I would like to hear your opinion. I have noticed that all units after the first candle are calibrated in such way they are more or less 1 candle. What do you think, why is that?

    • mathscinotes says:

      I noticed the same thing. The need to preserve a standard, even a bad one, is very strong. Commercial interests are the main force behind this resistance to change. When the optics industry began to move away from combustion-based light standards, there were millions of light bulbs being purchased every year. The industry needed an easy to understand equivalent rating for the light bulbs that people had been buying. I still see flashlights rated in the old "candlepower."

      In the US, light bulbs even go as far to give their power rating today (e.g. 8 W) and the equivalent incandescent wattage they emit (e.g. 60 W ). People want to be able to compare the cost of the light they are buying now to what they bought in the past.

      Preserving bad standards is very distasteful. The worst case I know of was done by Microsoft with Excel. Their standard date system is based on the flawed assumption that their reference year of 1900 was a leap year. Rather than fix it, Microsoft decided to perpetuate the error because so many spreadsheets were already in existence. It was better to be in error than to fix the problem. Of course, they did give you the option of the correct 1904-based system. However, if you use the correct system you are now no longer compatible with everyone else.


Comments are closed.