under which the photometer is used. In an attempt to obviate this error, a so-called compensated test-plate as illustrated in Fig. 12 has been produced.13 In this figure the light incident upon the upper surface of the test-plate P is reinforced by light admitted through a translucent glass ring A, so that at all angles the brightness of the under surface of P due to the combined action of the light trans

Fig. 14. Principle of compensated reflecting test-plate.

mitted from above and the light incident upon it from beneath, corresponds to the theoretical amount. It will be noted that the amount of compensation increases with the angle of incidence and hence can be made practically complete for all angles up to the very

Fig. 15.-Errors of various test-plates viewed normally.

A, Depolished transmitting plate; B, polished transmitting plate; C, depolished glass reflecting plate; D, compensated transmitting plate; E, compensated transmitting plate.

high ones. The intrusion of light to the ring A when the rays are parallel to P is prevented by the screen S. The test-plate is attached. to a photometer as shown in Fig. 13. The same principle is shown applied to a reflecting test-plate in Fig. 14. In this case the reflecting test-plate consists of a sheet of depolished white glass. This is set a small distance from another diffusing white surface C. Compensating light falling upon the plate C is reflected on the lower sur

13 Sharp and Little, I. E. S. Trans., Vol. 10, p. 727, 1915.

face of P, transmitted to the upper surface of P, and reinforces the light reflected from P to exactly the right amount to insure compensation for the deficiencies of P at high angles. The behavior of various test plates is summed up in the curves of Fig. 15. It is evident that where high precision is required in illumination photometry, and where test-plate errors cannot be computed and allowed for, the use of some form of corrective test-plate is of vital importance.

Static Illumination Tester. 14-This instrument provides a means for the ready measurement of illumination to a relatively low degree of

precision by an instrument of extreme simplicity in construction and use. The principle which it involves may be described as follows: The field of uniformly graded brightness produced by the comparison lamp, instead of being formed in the air where it is invisible, is formed on a sheet of diffusing material, which thereby is given a continuously graded brightness. The light which is to be measured falls on a diffusing sheet in juxtaposition to this field, so that the point can be readily seen where the brightness of the one field equals that of the other. The graded field being calibrated, the brightness of the unknown field is determined by finding with the eye the point where the brightness of the unknown field. equals the brightness of the known graded field.

The illumination tester embodying this principle is shown in Fig. 16. This figure shows a rectangular box B approximately 2.5 X 2.5 cm. in cross-section by 20 cm. long, containing at one end a small tungsten filament lamp L behind an opal glass screen. The top of the box over the rest of its length is made up of a sheet of

14 Sharp, Electrical World, September 16, 1916.

clear glass to which is pasted an arrangement of papers P which constitutes what may be described as a continuous photometer disc of the Leeson type, extending from one end of the glass to the other. The interior of the box is painted white, except for the far distant end which is black. The photometric element P consists of a sheet of fairly heavy paper with a slit cut out of it having saw tooth edges. Over the entire arrangement is then pasted a sheet of thinner translucent paper having a mat surface. When the lamp is lighted, the end of the slit nearest the opal glass is seen to be very bright, and this brightness fades away gradually toward the other end of the slit. When an exterior illumination falls on this photometric device, the outer portions are illuminated almost wholly by this exterior illumination, while the slit is illuminated chiefly by the light from the inside of the box. At the point where the brightness of the exterior portion is the same as the brightness of the slit, the saw teeth fade away and are hard to distinguish. This point which can be recognized without difficulty provided the papers are properly selected for the purpose, indicates the photometric value on the scale.

The completed apparatus in its experimental form is shown in Fig. 17. In this figure the photometric box is seen mounted on a larger box which contains a single dry cell serving as the source of current. The box also carries a small precision voltmeter and a rheostat. The photometric box is so arranged that it can be removed from the rest of the apparatus, the flexible cord conductor with which it is connected being stowed away in the larger box when the two are used as a unit. Photometric readings are taken in a direction at right angles to the axis of the box. The exact angle to the vertical at which these readings are made seems to be of relatively little importance in the general case.

By slight structural modifications the instrument can be adapted to the measurement of the brightness of surfaces as well as the illumination incident upon them.

Mr. R. ff. Pierce15 has also produced an instrument of this same eneral class.

ILLUMINATION MEASUREMENTS

Practice in illumination measurements is so varied according to conditions governing any particular test, that to go into anything approaching a complete discussion would be beyond the scope of

15 American Gas Light Journal, page 67, August 14, 1916.

this lecture. Certain general precautions, however, need to be taken in practically every case. Among the most important ones are the following:

To be sure that the comparison lamp in the photometer is giving its correct candle-power. This involves a comparatively recent standardization of the same taken in connection with its electrical measuring instrument, and an assurance that the electrical measuring instrument is sufficiently reliable and accurate for its purpose. To use the photometer in such a way that there is no undue loss of light on the test-plate due to the presence of the operator or other person.

To select the test stations properly according to the design of the test.

To see that the test-plate is level and in the proper position. To see that the scale readings are properly recorded and any condition such as the introduction of a neutral glass screen is noted.

The subject of precautions to be taken in illumination measurements has been quite fully treated by Little16 in an Illuminating Engineering Society paper.

MEASUREMENT OF BRIGHTNESS

For many purposes it is desirable to know the brightness values of objects or of walls and ceiling in a room or of a shade or reflector of a lamp. The standard forms of portable photometers designed to measure illumination enable these measurements to be made simply by removing the test-plate, if there be one, and sighting the photometer directly on the object in question. Photometric balance is then secured between the object and the diffusing plate in the photometer. The reading of the scale needs to be multiplied by a constant to give the brightness value either in candle-power per square inch or in millilamberts.17 The determination of this constant is a matter for the standardizing laboratory and is not particularly easy, inasmuch as it involves illuminating to a known degree a surface of known area which is then photometered as a source of light. It should be noted that the brightness constant of a photometer is a function only of the test-plate which is used with it, and that with changes in the calibration of a photometer this constant is unaffected, provided the test-plate is unchanged. The relation between the brightness constant expressed in apparent

14 Little, Transactions I. E. S., Vol. 10. page 766, 1915.

17 The millilambert is a unit of brightness, and is equal to the brightness of a perfectly reflecting and perfectly diffusing surface on which one millilumen per square cenfimeter falls.

lumens emitted per square foot, and the illumination which is the lumens incident per square foot, is evidently the transmitting or reflecting power of the test-plate according as the test-plate is of the transmitting or reflecting type. In practice in the standardizing laboratory it is convenient to have carefully preserved a standard test-plate of known brightness constant which can serve as a reference plate for the calibration of other plates.

DAYLIGHT MEASUREMENTS

For measuring daylight the use of a light filter to secure an approximate color match is indispensable. Inasmuch as the quality of daylight varies greatly, dependent upon the character of the sky, no one filter will enable a match to be made, but a single filter may in practice be used because the outstanding differences are not so excessive as to prevent fairly good measurements being made. On account of the very high values of illumination usually given by daylight, it is more convenient to put the filter on the daylight side rather than on the side of the comparison lamp. In the practice of the Electrical Testing Laboratories a sheet of suitably colored gelatin is sometimes utilized such as is employed in spot-lighting in theatrical work. Daylight foot-candle values alone are frequently, perhaps usually, of subsidiary importance because illumination values vary so much from time to time with the outdoor or sky conditions. Rather the photometrist must give a value at the place which is studied, coupled up in some was with a value representing outdoor conditions. The condition most commonly chosen is the brightness of some portion or of all of the visible sky, or the illumination produced by some portion of all of the visible sky. For this purpose various types of apparatus have been produced by which the brightness of the sky can be compared with the brightness of a test-plate in a room, for instance. As illustrative of this class of problems, the methods employed by the Electrical Testing Laboratories in studying the obstruction of daylight to buildings caused by alterations in a structure in the street will be instructive. In this work one photometer with a vertical test-plate was placed close to the window-line of the building. Alongside of it was another photometer having a fan-shaped arrangement placed over the test-plate whereby the test-plate received only the light from the unobstructed portion of the sky. (See Fig. 18.) Readings were made simultaneously with these photometers and thereby a relation was obtained between the illumination produced