276 CARNEGIE INSTITUTION OF WASHINGTON. 



large negative pressure shifts. But the effect appears not to be due to 

 differences of pressure in the arc, as other Hnes with pressure shifts 

 of 0.009 A per atmosphere show no determinable displacements, while 

 a difference in pressure of the order of an atmosphere is required to 

 account for the large displacements observed. Further, the displace- 

 ments observed do not vary as the cube of the wave-length, as in the 

 case with pressure shift. Provisionally they are considered to be due 

 to changes in density, but the question is still under investigation. 



The remaining 1,300 lines compared show no measurable difference 

 in wave-length depending upon the portion of the arc used or current 

 strength. Their independence of ordinary arc conditions makes them 

 eminently suitable for standards of wave-length. 



STANDARDS OF WAVE-LENGTH. 



The critical study of the iron arc by Mr. St. John and Mr. Babcock 

 has shown what lines can be employed with confidence as standards, 

 what lines must be used with caution, and under what arc conditions 

 they yield consistent results. Unfortunately many of the tertiary 

 standards and some of the international secondaries belong to these 

 groups of questionable lines. Between X 2900 and X 6700 there are 

 1,300 lines of good quality, and approximately 300 lines of the question- 

 able groups. No class of lines is distributed with even approximate 

 uniformity through the spectrum, so that a wide region, from X 5500 to 

 X 6000, is without a line of high quality, and from X 4500 to X 5000 the 

 great majority of the lines belong to the questionable classes. In these 

 two regions both the secondaries and tertiaries must of necessity include 

 questionable lines. It has thus been necessary to establish the limiting 

 arc conditions which will yield wave-lengths of the desired precision. 

 The source recommended by the Committee of the International Union 

 for Cooperation in Solar Research is an arc 6 mm. long between iron 

 terminals 7 mm. in diameter, carrying a current of 6 amperes. The 

 investigation has shown that the current may vary between 5 and 7 

 amperes without producing a variation in wave-length exceeding 0.001 

 or 0.002 A, and that the light should not be taken from a point nearer 

 to the negative pole than 2 mm. nor nearer the positive pole than 

 1 mm. In view of these results it has seemed advisable to Mr. St. John 

 and Miss Ware to redetermine the wave-lengths of the suspected lines 

 under strictly standard conditions, and to omit lines of this character 

 in extending the measurements into the ultra-violet, where lines of high 

 quality are the prevailing type. The measurement of the sun-arc plates 

 for the wave-lengths on Mount Wilson and the pole-center plates for 

 the wave-lengths in Pasadena are in progress. With the installation 

 of an interferometer on the mountain, a series of secondary standards 

 will be determined for use at that elevation. 



The investigation by Mr. St. John and Mr. Babcock of the behavior 

 of lines under laboratory conditions is still in progress — in particular a 



