84 JOURNAL OF THE WASHINGTON ACADEMY OF SCIENCES VOL. 13, No. 5 
temperatures in a vacuum. These equations, known as the Middle- 
kauff-Skogland equations, have been very useful and still constitute 
the best established procedure for determining relative values for 
vacuum tungsten lamps at various efficiencies. Moreover, the experi- 
ence gained at the Bureau since these equations were established 
indicates that the values were very fortunately chosen. They appar- 
ently agree well with the National Physical Laboratory results over 
the range where they overlap, and the values have been rather closely 
checked by successive groups of observers at the Bureau during the 
undesirably rapid change of personnel which we have suffered during 
the last few years. It should be emphasized, however, that groups 
of observers of at least equal experience in other laboratories have not 
agreed so well with these results. 
Some years ago very careful comparative measurements on lamps 
and glasses representing the step from carbon lamps to tungsten were 
made by groups of five or six observers in each of the three laboratories 
in the country best fitted for this work. Measurements made a 
number of months apart indicated that each laboratory adhered very 
closely to its average result, the consistency of performance in this 
respect being within one-quarter of one per cent. However, the 
difference between two of the laboratories became as large as 1.9 per 
cent, although the range covered was only that from carbon lamps 
to vacuum tungsten and did not approach the color of gas-filled 
lamps. The Bureau of Standards was at one extreme in the results 
obtained, but other data have been in general more consistent with 
the Bureau’s own measurements in this intercomparison, and conse- 
quently these values as represented in the Middlekauff-Skogland 
curves have been retained. The rather imperfect agreement obtained 
in the intercomparisons has been accepted as a practical corroboration 
of these values, but again it is somewhat disturbing to have differences 
of the order of 2 per cent in checking values which are supposed to 
be maintained to a fraction of 1 per cent. 
The result of such comparisons as have been referred to has been 
to establish agreement on more or less arbitrary values for standards 
of successively higher and higher temperature and efficiency, which, 
when once established, can be maintained as practically independent 
standards for future use. In fact in the case of the National Physical 
Laboratory it is definitely stated that the values of the higher effi- 
ciency standards are ‘‘assigned to them once for all.’ Such multi- 
plication of standards, or determination of curves, while serving to 
meet the more urgent commercial needs, is limited in its application. 
