74 JOURNAL OF THE WASHINGTON ACADEMY OF SCIENCES VOL. 13, No. 5 
lambert may therefore be defined as the brightness of a perfectly 
diffusing surface emitting or reflecting one lumen per square centi- 
meter. This represents a high degree of brightness for a secondary 
source. A white surface in the brightest direct sunlight would have 
a brightness of only 10 lamberts, and for most cases of illuminated 
surfaces, the millilambert, or 0.001 lambert, is more convenient. The 
millilambert is the brightness of a perfectly white surface with 10 
meter-candles illumination, and since a foot-candle is somewhat over 
10 meter-candles (the actual ratio being 10.76), the millilambert 
represents almost exactly the brightness of an actual white surface, 
such as magnesium oxide or carbonate, under an illumination of 
1 foot-candle. 
A special case in which a still smaller unit, the microlambert (one- 
millionth of a lambert), is used, is the measurement of radio-active 
self-luminous materials. The brightness of these materials as actually 
used is a few microlamberts. That is, they are as bright as a white 
surface receiving a few thousandths of a foot-candle; for comparison, 
it may be noted that full moonlight is several hundredths of a foot- 
candle, giving to a white surface a brightness ten times as great as these 
materials show. 
Referring to the table of units we have constructed (Table I), it 
may be noted that the light output of a source can be specified in 
two ways. We can either state the flux in lumens or give theaverage 
intensity or candlepower over a specified solid angle, usually the 
complete sphere. From the point of view taken in making up this 
table it appears that the former practice is simpler, but on the other 
hand the historical development favors the second, and in other 
countries than this the flux rating has gained little use in practice. 
In this country, however, the lumen has rapidly attained an extensive 
use as a practical unit for two reasons. One is that for rough and 
ready calculations of illumination the lumen is more convenient, 
For example, if one has to provide an area of 2000 square feet with 
an average illumination of 5 foot-candles the product 10,000 imme- 
diately gives the net amount of flux required. A more potent 
reason, however, is the fact that so long as candlepower is considered 
to indicate light output, misunderstandings are bound to occur, and 
misrepresentation is facilitated. For instance, when a lamp rated 
at 21 candles is put in a headlight reflector and gives a candlepower 
of many thousands, it is not surprising that the non-technical man 
is puzzled. Even a technically trained man is likely to overlook 
the fact that watts per candle for one kind of lamp do not mean the 
