ROSA AND CRITTENDEN: FLAME STANDARDS 281 



adopted by the x'arious governments. In Germany preference 

 is given to the Hefner lamp, in England to the pentane lamp, 

 and in France to the Carcel lamp. Each of these serves in its 

 own country both as the primary standard and as a working 

 standard, but for the photometry of electric lamps and generally 

 in accm*ate photometric work standardized electric incandescent 

 lamps are used in all countries. In America a group of such lamps 

 kept at the Bureau of Standards is considered as provisional 

 primary standards serving to maintain the unit until a better 

 primary standard shall have been devised. It is believed that 

 the unit which has been agreed upon can be so maintained with 

 an accuracy considerably above that with which it can be re- 

 produced by reference to any of the so-called reproducible stand- 

 ards at present in use. In other words, the incandescent lamps 

 have really been employed as primary standards, and the flame 

 standards, which logically should play the part of primary stand- 

 ards, have been relegated to a subordinate position. 



There is, however, a possibility of an appreciable drift in the 

 value of the unit if there is no photometric standard accurately 

 reproducible from its specifications which is capable of serving 

 as a reliable check upon the electric standards. It has therefore 

 appeared worth while to make a study of the best types of flame 

 lamps to see how closely they would reproduce, in the Bureau 

 laboratory, the A^alues adopted by international agreement and 

 also to find whether their reliability as primary standards could 

 be increased by any changes in construction or in operation. 



The Carcel lamp is by far the least reliable of the three types, 

 and cannot be considered as a competitor for general accept- 

 ance. The Hefner and the pentane lamp as made at present 

 divide honors; the latter is markedly superior as a practical 

 standard, but individual pentane lamps do not agree, and until 

 lamps can be independently made which shall give the same 

 value the type can hardly be said to be reproducible. The 

 Hefner lamp is so simple in construction that reproduction 

 of lamps is relatively easy. Lamps now made show small 

 differences due to slight departures from mean dimensions, 

 but these differences can be made negligible by more careful 



