inches long. An efficiency of 1.4 watts per hefner is claimed by the manufacturer, from whose report these figures are taken. Watts per square foot, 1.39.

Figure No. 8 is a good illustration of lighting by means of Cooper Hewitt tubes. This pier is about 526 feet long by 83 feet wide, and is illuminated by seven 45-inch 3.5 ampere tubes. Watts per square foot, 0.064.

Figure 9 shows the same pier illuminated by three flaming arcs, which consume approximately 5 amperes apiece. Watts per square foot, 0.041.

Figure 10 shows a carpet store lighted by two 3-glower Nernst lamps. Total watts, 500. This room is 125 feet long by 60 feet wide; height of ceiling, 22 feet. Watts per square foot, 0.08. The customer is pleased with the quality of the light, colors apparently having satisfactory values.

Figure 11 is an illustration of the exhibition ring in a sales stable, lighted by incandescents.

In conclusion, I might add that the mere arranging of minor installations does not constitute illuminating engineering. Careful study and wide experience are necessary before difficult problems can be approached.

Whether or not it would pay central-station management to employ illuminating engineers is an open question. The small cases arising could probably be taken care of by the local inspectors, and in the larger cases the customer or prospective customer would undoubtedly desire to employ his own illuminating engineer, as well as his own electrical engineer.

THE PRESIDENT: This paper is now open for discussion.

DISCUSSION

MR. LANSINGH: Relative to the question as to whether or not an illuminating engineer is a valuable adjunct to the central station, I should like to say from my experience with different central stations that such an engineer is a most valuable addition to the commercial force. I have in mind the case of one large department store where gas arcs had been decided upon and where, after comparative costs, values of illumination, and so forth, had been carefully worked by an illuminating engineer, the owners decided to change to distributed electrical units, thus

adding several thousand dollars a year to the income of the central station. In fact, this one installation alone would probably be worth to the central station the entire salary of such an engineer for a whole year.

As to the question of minor installations being taken care of by inspectors, as suggested by Mr. Norman, I would call attention to the fact that the ordinary bedroom, which seems a very simple thing to light properly, is, as a matter of fact, a very difficult thing to light properly in an economical manner, and that the ordinary inspector is entirely inadequate to care for even so simple a proposition as this. When the ordinary inspector is able to handle the lighting of a house from an artistic, economical and hygienic standpoint, he is then probably in a position to designate himself an illuminating engineer.

THE PRESIDENT: As a matter of experience, I may say that just that work to which Mr. Lansingh has referred has proven its commercial value over and over again, and, what is of even more importance, it has given electric lighting under certain conditions a new value to the users.

If there is no further discussion on this paper we shall proceed to the next paper, The Frequencies of Flicker at Which Variations in Illumination Vanish, by Dr. A. E. Kennelly and Mr. S. E. Whiting, of Harvard University.

The following paper was presented by Dr. Kennelly: THE FREQUENCIES OF FLICKER AT WHICH VARIATIONS IN ILLUMINATION VANISH

It is well known that arc lamps operating on 60-cycle alternating-current circuits and subject, therefore, to 120 cycles of flicker per second, produce visible flickering of illumination on moving targets. It is also well known that arc lamps operating on 25-cycle alternating-current circuits produce visible flickering of illumination on stationary targets, and that incandescent lamps operating on such circuits also produce visible flickering in illumination under certain conditions. There is, however, considerable difference of opinion as to the conditions under which flickering appears or disappears in the illumination cast by incandescent lamps operating on low-frequency alternating-current circuits. It is the object of this paper to describe some experiments made by the writers on this matter, and the results thereby indicated.

The measurements were made in a photometer-room with dark walls, by employing a rotating sector-disc immediately in front of a steady and stationary incandescent lamp, thus producing an artificial flicker in its light. It is known that when different observers attempt to balance, in the ordinary photometer, a steady illumination against a flickering illumination, they make different estimates of the photometer setting.* The eye does not seem to be able to integrate or average an illumination that undergoes flickering, or visibly rapid cyclic variation. One observer may estimate the mean balance of illumination relatively high and another relatively low with respect to the mean value. It was found on trial, however, that when several different observers simultaneously watched a stationary target illuminated by a beam of light which was artificially flickered through the intervention of a rotating sector-disc, they substantially agreed upon the speed of rotation at which the flickering disappeared. This method of observation was, therefore, adopted in the tests here described.

The general arrangement of the apparatus used is indicated

*Measurement of Mean Horizontal Candle-Power, by Hyde and Cady, Bulletin of Bureau of Standards, Vol. II, No. 3, 1907.