LECTURE XII. ILLUMINATION AND ILLUMINATING ENGINEERING. 110. Artificial light is used for the purpose of seeing and distinguishing objects clearly and comfortably when the day- light fails. The problem of artificial lighting thus comprises con- sideration of the source of light or the illuminant; the flux of light issuing from it; the distribution of the light flux in space, that is, the light flux density in space and more particularly at the illuminated objects; the illumination, that is, the light flux density reflected from the illuminated objects, and the effect produced thereby on the human eye. In the latter, we have left the field of physics and entered the realm of physiology, which is not as amenable to exact experimental determination, and where our knowledge thus is far more limited than in physical science. This then constitutes one of the main difficulties of the art of illuminating engineering: that it embraces the field of two dif- ferent sciences — physics and physiology. The light flux entering the eye is varied in its physical quantity by the reaction of the eye on light flux density in contracting or expanding the pupil. The effect of the light flux which enters the eye is varied by the fatigue, which depends on intensity and also on color. Distinction is due to differences in the light flux density from the illuminated objects, that is, differences of illumination, which may be differences in quality, that is, in color, or differences in intensity, that is, in brightness, and as such includes the effect of shadows as causing differences in intensity at the edge of objects. The physical quantities with which we have to deal in illumi- nating engineering thus are : The intensity of the light source or the illuminant, and its brilliancy, that is, the flux density at the surface of the illuminant; The flux of light, that is, the total visible radiation issuing from the illuminant; 256 ILLUMINATION AND ILLUMINATING ENGINEERING. 257 The light flux density, that is, the distribution of the light flux in space, and The illumination, that is, the light flux density issuing from the illuminated objects. The intensity of a light source is measured in candles. The unit of light intensity, or the candle, is a quantity not directly related to the absolute system of units, but reproduced from specifica- tions or by comparison with maintained standards, and for white light is probably between 0.04 and 0.02 watt. Intensity has a meaning only for a point source of light; that is, an illumi- nant in which the flux of light issues from a point or such a small area that, at the distance considered, it can be considered as a point. " Intensity of light " thus is a physical quantity of the same nature as " intensity of magnet pole," which latter also presupposes that the total magnetic flux issues from a point, and thus is applicable only when dealing with such distances from the source of the light flux or magnetic flux, that the flux can be assumed as issuing from a point. Frequently the inten- sity of a light source is different in different directions, and then either the distribution curve of the light intensity is required for characterizing the illuminant, or the average of the intensities in all directions in space is used, and is called the "mean spherical intensity." The unit of light intensity, or the candle, is the intensity which produces unit flux density at unit distance from the light source, and thus produces a total flux of light equal to 4 n units (the surface of the sphere at unit distance from the light source). The unit of light flux is called the lumen, and one candle of light intensity thus produces 4 n lumens of light flux (just as a magnet pole of unit intensity produces 4 x lines of magnetic force). The light flux is the essential quantity which characterizes the usefulness of an illuminant, and it is the raw material from which all illuminating engineering starts. Any source of light can be measured in units of light flux or lumens — the diffused daylight entering the windows of a room, or the visible radia- tion of the mercury lamp or a Moore tube as well as that of a point source — by adding all the flux densities intercepted by any surface enclosing the source of light. In a point source of light, the intensity, in candles, is the total 258 RADIATION, LIGHT, AND ILLUMINATION. flux of light, in lumens, divided by 4 x. In any illuminant which is not a point source, we cannot speak of an intensity, except at such distances at which the source of light can be assumed as a point; and in interior illumination this is rarely the case. Since, however, the candle power, as measure of the intensity of light, has become the most familiar quantity in characterizing illuminants, very commonly even sources of light which are not point sources — as a Moore tube or the diffused daylight — are expressed in " equivalent candle power" and when thus speaking of the candle power of a mercury lamp, or of the diffused daylight from the windows, we mean the candle power of a point source of light, which would give the same total flux of light as the mercury lamp, or the daylight from the windows, etc. The " equivalent candle power," or frequently merely called " mean spherical candle power/' thus is the total light flux divided by 4 TT, hence in reality is not a unit of intensity, but a unit of light flux. This explains the apparent contradiction between the claims that sources of light, as the mercury lamp or the Moore tube, can- not be expressed in candle powers, while at the same time their specific consumptions are given in candle power per watt : mean- ing equivalent candle power, which refers to the total flux of light, and thus is a definite and measurable physical quantity. While it is not probable that the custom of rating illuminants in candles, regardless of their shape, will quickly disappear, and no objection exists against it, provided that it is understood to mean the equivalent candle power, it is preferable to use the cor- rect unit of light flux, and express the output of a source of light in lumens, adding where necessary the equivalent candle power in parenthesis. Obviously, the use of the candle power in any particular direction — horizontal, or terminal, or maximum can- dle power — has a meaning only in characterizing the distribution of the light flux, as applicable for a particular purpose, as street lighting, but, when used for rating the illuminant by its light flux output, is an intentional or unintentional deception. Incandes- cent lamps have been rated, and to some extent still are, in hori- zontal candle power, but in this case the horizontal candle power has ceased to mean the actual horizontal candle power, but is the horizontal candle power which with a certain standard dis- tribution of light flux would correspond to the light flux of the ILLUMINATION AND ILLUMINATING ENGINEERING. 259 lamp, and thus also is merely a practical measure of the light flux, retained by convenience : one horizontal candle power rep- resents 0.78 mean spherical or equivalent candle power of the standard distribution curve, and thus 4 TT X 0.78 lumen. In general, intensity, or candle power, thus is an angular measure, useful in characterizing the distribution of the light flux, but not the total light flux. 111. Light- flux density is the light flux per unit area traversed by it, thus is measured in lumens per square meter (or square foot), just as the magnetic density is measured in lines of mag- netic force per square centimeter. In illumination, as unit of length, usually the meter is employed, and not the centimeter, as in the absolute system of units, and 102 thus is the reduction factor to absolute units. Frequently also the foot is used as practical unit of length. For a point source of light the light flux density is the inten- sity of the light source, in candles (in the direction towards the point of observation, if the distribution is not uniform in all directions), divided by the square of the distance, in meters, or feet, and the light flux density thus is frequently expressed in meter-candles, or foot-candles. Thus at 10 feet distance from a 16 candle power lamp, the light flux density is 0.16 foot-candle, or 0.16 lumen per square foot. Very commonly, therefore, the light flux density produced by sources of light which are not points, is also expressed in meter-candles or foot-candles — which numerically is the same value, that is, the same quantity, as lumens per square meter or square foot, but physically would refer to the equivalent candle power of the light source. Illumination is the light flux density reflected from the illu- minated object, and as flux density thus is measured also in lumens per square meter or square foot, or in meter-candles or foot-candles. Brilliancy is the light flux density at the surface of the illumi- nant, and as flux density thus could also be measured in lumens per square meter or square foot, but, as this would usually give enormous values, brilliancy of the light source generally is meas- ured in lumens per square centimeter, or per square millimeter. It is a quantity which is of high importance mainly in its physio- logical effect. Light intensity, brilliancy and light flux thus are character- 260 RADIATION, LIGHT, AND ILLUMINATION. istics of the illuminant, while flux density is a function of the space traversed by the light flux, but not of the source of light : with the same source of light, in the space from the surface of the illuminant to infinite distance, all light flux densities exist between the maximum at the surface of the illuminant (its brilliancy) and zero. Brilliancy thus is the maximum of the light-flux density. While intensity and brilliancy depend upon the shape of the illuminant, light flux is independent thereof. Illumination is a quantity which depends not only on the source of light, that is, light flux and flux density, but also on the illumi- nated objects and their nature, and thus is the light flux density as modified by the illuminated objects. Very commonly, how- ever, the term " illumination " is used to denote " light flux density," irrespective of the illuminated objects. 112. The light flux thus is the raw material with which illuminating engineering starts, and the first problem then is to distribute the light flux through space so as to give at all points the light flux density required for satisfactory illumi- nation. Some problems, as the lighting of a meeting place, school- room, etc., require a uniform or general, and fairly high intensity of illumination, while in street lighting a uniform but fairly low intensity of illumination is desirable. In other cases, mainly a local or concentrated illumination is needed. Usually, however, a combination of a local or concentrated illumination, of fairly high intensity, with a general illumination of lower intensity, is required: the former at those places where we desire to distinguish details, as where work is being done, at the reading- table, work bench, dining-table etc., while the general illumina- tion is merely for orientation in the space, and thus may be of lower intensity, and for reasons of economy, and also physio- logical reasons, should be of lower intensity. We thus have to distinguish between local or concentrated, and general or uniform, illumination, and a combination of both, and have to distribute the light flux in accordance there- with, that is, produce a high flux density at the points or areas requiring high concentrated illumination, a low and uniform flux density throughout the remaining space. This can be done by choosing a light source of the proper distribution curve, as, for instance, in street illumination a lamp ILLUMINATION AND ILLUMINATING ENGINEERING. 261 giving most of the light flux between the horizontal and 20 deg. below the horizontal; in many cases of indoor illumination a light source giving most of the light between the vertical and an angle of from 30 to 60 deg. from the vertical — depending on the diameter of the area of concentrated illumination and the height of the illuminant above it. It can also be done by modifying or directing the light flux of the illuminant by reflec- tion or diffraction and diffusion, either from walls and ceilings of the illuminated area, or by attachments to the illuminant, as reflectors, diffusing globes, diffracting shades, etc. Further- more, the required flux distribution can be secured by the use of a number of illuminants, and with a larger area this usually is necessary. Frequently the desired flux distribution is pro- duced by using an illuminant giving more light flux than neces- sary, and destroying the excess of flux in those directions where it is not wanted, by absorption. Obviously this arrangement is uneconomical and thus bad illuminating engineering; the desired flux distribution should be secured economically, that is, without unnecessary waste of light flux by absorption, and this usually can be done by a combination of a number of light sources of suitable distribution curves. The most economical method of securing the desired distribution curve obviously is to choose a light source coming as near to it as possible, and then modifying it by reflection or diffraction. 113. Thus far, the problem is one of physics, and the result, that is, the objective illumination, can be measured by photometer or luminometer, and thus checked. The duty of the illuminat- ing engineer, however, does not end here, but with the same objective illumination, that is, the same distribution of light flux throughout the entire illuminated area, as measured by photometer, the illumination may be very satisfactory, or it may be entirely unsatisfactory, depending on whether the physio- logical requirements are satisfied or are violated ; and very often we find illuminations which seem entirely unsatisfactory, tiring, or uncomfortable, but when judged by the density and the distribution of the light flux, should be satisfactory. Even numerous commercial illuminants, designed to give suitable distribution curves, fail to do justice to their light flux and its distribution, by violating fundamental physiological require- ments. 262 RADIATION, LIGHT, AND ILLUMINATION. The physiological problems of illumination, that is, the effects entering between the objective distribution of light flux in space, and the subjective effects produced on the human eye, thus are the most important with which the illuminating engineer has to deal, and the first feature which must be recognized is that the objective illumination, as measured by the photometer, is no criterion of the subjective illumination, that is, the physiological effect produced by it, as regard to clearness, comfort and satis- faction, and it is the subjective illumination by which the success of an illuminating engineering problem is judged. The most important physiological effects are : (a) The contraction of the pupil. The pupil of the eye auto- matically reacts, by contraction, on high brilliancy at or near the sensitive spot, that is, the point of the retina, on which we focus the image of the object at which we look, and to a some- what lesser extent on high brilliancy anywhere else in the field of vision. If, therefore, points or areas of high brilliancy are in the field of vision, especially if near to objects at which we look, the pupil contracts the more the higher the brilliancy, and thereby reduces the amount of light flux which enters the eye, that is, produces the same result as if the objective illumination had been correspondingly reduced, intensified by the uncomfortable effect of seeing high brilliancy. The exist- ence of points of high brilliancy in the field of vision thus results in a great waste of light flux, and additional discomfort, and, for satisfactory illumination, points of high brilliancy thus must be kept out of the field of vision. Light sources of high brilliancy must be arranged so that they cannot directly be seen, but the illumination accomplished by the light reflected from ceilings, etc., or from reflectors attached to the illuminant: indirect light- ing; or at least the light sources should be located where we are rarely liable to look at them, that is, with moderate-sized rooms, at or near the ceilings. Or light sources of moderate intrinsic brilliancy should be used, as the Moore tube, the mercury lamp, the Welsbach mantel. Or, with illuminants of high brilliancy, as the electric arc, the incandescent lamp (especially the tungsten filament), etc., the brilliancy of the illuminant must be reduced by enclosing it with a diffusing or diffracting globe or shade, as an opal or frosted or holophane globe, etc. ILLUMINATION AND ILLUMINATING ENGINEERING. 263 No illumination, however, can be satisfactory in which the eye at any time can be exposed to the direct rays from a tungsten filament or an arc. While the methods of removing the high brilliancy of the illuminant usually involve a considerable loss of light flux, by absorption at the refracting surface, in the frosted or opal globe, etc., and the objective illumination thus is de- creased, if the methods of reducing the brilliancy are anywhere reasonably arranged, the light flux entering the eye, and thus the subjective illumination, is increased, and often very greatly. Thus while frosting an incandescent lamp decreases its light flux by about 15 per cent, in spite thereof usually more light flux enters the eye from the frosted lamp than from a clear glass lamp at the same distance. It is, therefore, inefficient to use illuminants of high brilliancy in the field of vision, and in addition makes the illumination uncomfortable and thereby unsatisfactory. Physiologically the brilliancy of the light source thus is one of the most important quantities. 114. (6) Fatigue. When exposed to fairly high light flux den- sity, that is, high illumination, the nerves of the eye decrease in sensitivity, by fatigue, and inversely, in lower illumination or in darkness, increase in sensitivity. This reaction, or adjust- ment of the sensitivity of the nerves of vision to different intensi- ties of illumination, enables us to see equally well in illuminations varying in intensity by more than 10,000 to 1 (as daylight and artificial light). Thus, when entering a well-illuminated room from the darkness, it first appears glaring, until gradually the impression fades down to normal. Inversely, coming from a well-lighted room into a space of much lower illumination, it first appears practically dark, until gradually the eye adjusts itself, that is, the nerves of vision increase in sensitivity by their rest, and then we again see fairly well. Fatigue and contraction of the pupil thus are similar in their action, in that they reduce the physiological effect for high intensities. The contraction of the pupil, however, is almost instantaneous, and is a protective action against excessive bril- liancies in the field of vision, while the fatigue is a gradual adjustment to the average intensity of illumination, within the operating range of the human eye. By exposure for a considerable period to the fairly high illumi- 264 RADIATION, LIGHT, AND ILLUMINATION. nation required when working by artificial light, the sensitivity of the eye decreases, the illumination appears less bright, and thus a higher illumination is required than would be sufficient in the absence of fatigue, and the continuous use and absence of rest cause the sensation of strain, that is, irritation or an uncomfortable feeling, as especially noticeable when working or reading for a considerable length of time in rooms having a high uniform intensity of illumination, as meeting-rooms, some libraries, etc. If, however, the eye can rest even momentarily, by a change to lower intensity of illumination, fatigue is decreased, never becomes as complete and uncomfortable, and the concen- trated illumination of the working-table appears brighter than it would without the possibility of rest. A room having a uniform intensity of illumination thus appears glaring and uncomfortable, and for satisfactory illumination it is necessary not only to provide a sufficiently high intensity at the place where needed, but it is just as necessary to keep the intensity of illumination as low as permissible, wherever it is not needed, so as to afford to the eye rest from the fatigue. In some cases, as meeting-halls, schoolrooms, this may not be possible, but a uniform high intensity required, to be able to work or read anywhere in the room. Where, however, it is not necessary, it is not merely uneconomical to provide a uniform high intensity of illumination, but it is an illuminating engineer- ing defect, and a high intensity should be provided, as concen- trated illumination, only at those places where required, as at the reading-tables of the library, but the general illumination should be of lower intensity. While we rarely realize the cause, we feel the superiority of the combination of high concentrated and lower general illumination, by speaking of such illumination as home-like, restful, etc. Especially in places where considerable work has to be done by artificial illumination, as in libraries, factories, etc., to get satisfactory results, it is important to consider this effect of fatigue, and to properly combine a moder- ately low general illumination with a local higher intensity of illumination at the places of work. The latter can usually be given by a light source having a downward distribution, located sufficiently high above the place of work. The average standing or reading lamp, however, generally is not sufficiently high to accomplish the result. Obviously, in such local illumination, ILLUMINATION AND ILLUMINATING ENGINEERING. 265 the brilliancy of the illumination must be kept low, as discussed above. Of considerable importance regarding fatigue is the quality, that is, the color, of the light : fatigue at high intensities occurs far more with yellow and orange rays than with white light, and very little with green and bluish-green light. Thus, in arti- ficial illumination, in which practically always the yellow and orange rays greatly preponderate, the question of fatigue is far more important than with the bluish-white diffused daylight, and the irritating effects of fatigue thus are mostly felt with artificial illumination. 115. (c) Differences. Objects are seen and distinguished by differences in quality, that is, color, and in intensity, that is, brightness, of the light reflected by them. If there were no differences in color or in intensity throughout the field of vision, we would see light but would not distinguish objects. Therefore, in good illumination, the differences in color and in intensity should be sufficiently high to see clearly by them, but still limited so as not to preponderate to such extent as to distract the atten- tion from smaller differences. The differences in intensity, to give distinction, should be high, but at the same time are limited by the phenomena of fatigue and of the contraction of the pupil: the minimum intensity must still be sufficiently high to see clearly, and the maximum intensity not so high as to cause fatigue and contraction of the pupil, much beyond that corresponding to the average intensity, otherwise the vision becomes indistinct and unsatisfactory, and uncomfortable by too much contrast; that is, the intensity differences must give a sufficient, but not an excessive, contrast, if the illumination is to be satisfactory. Differences in quality, that is, in color, are to a limited extent only under the control of the illuminating engineer. In some cases the illuminating engineer can control or advise regarding the color of objects, as the walls, ceilings, etc. In most cases, however, the absolute color of the illuminated objects is not within the control of the illuminating engineer: for instance, in street lighting, the color of the street surface, its surroundings, as vegetation, houses, etc., are fixed and cannot be changed for effects of illumination. So also in most cases of indoor illumina- tion. To some extent, however, the subjective color can be con- 266 RADIATION, LIGHT, AND ILLUMINATION. trolled by the choice of the proper shade of light, and thereby slight color differences increased and made more distinct, or decreased and thus obliterated. For instance, the color resulting from age and dirt is usually the color of carbon and of iron, yellowish brown or reddish brown, that is, colors at the long wave end of the spectrum. Spots and blemishes due to dirt or age, thus are made more distinct by using an illuminant defi- cient in the long waves of light, as the mercury lamp, while in- versely they are decreased by using a reddish-yellow illuminant, as the incandescent lamp or the candle. Thus the white arc lamp and still more so the bluish-green mercury lamp shows blemishes and slight color differences of age and dirt harsh and exaggerated, while the yellow light softens them and makes them disappear; and while, for a ballroom, the yellow light is thus preferred, and the mercury arc or even the ordinary white carbon arc would give a harsh and disagreeable effect, inversely the yellow light would be unsuitable where such slight differences should be distinguished. It is therefore essential for the illuminating engineer to choose as far as it is feasible the proper color of light, and an otherwise good illumination may be spoiled by using too white or too yellow a light. The main distinction of objects, however, is due to differences in intensity or brightness, and, for producing these, the shadows are of foremost assistance, and indeed the differences of inten- sity, by which we see objects, are to a large extent those due shadows. The study of the shadow thus is one of the most important subjects of illuminating engineering. If we have no shadows, but a perfectly diffused illumination, even if the intensity of illumination is sufficient, the illumination is unsatis- factory, as we lose the assistance of the shadows in distinguishing objects, and therefore find seeing more difficult, the illumination restless and uncomfortable. The use of shadows for illumination requires that we must have directed light, that is, light coming from one or a number of sources, and thus causing shadows, and not merely diffused illumination, that is, light coming from all directions and thus causing no shadows. While, however, in general perfectly dif- fused illumination is unsatisfactory, an illumination having only directed light is also unsatisfactory. If the light is all directed, as from a single arc, the shadows are absolutely black, we can- ILLUMINATION AND ILLUMINATING ENGINEERING. 267 not see anything in them, and, in attempting to see the objects in the shadows, the illumination becomes tiring to the eyes, irritating and restless. For satisfactory illumination, it therefore is necessary to have sufficient directed light to mark the edge of the objects by their shadow, and thereby improve distinction, but at the same time sufficient diffused light to see clearly in the shad- ows; that is, a proper proportion of directed and diffused light is necessary. In cases in which all the objects assume practically the same color, as in flour mills or foundries, a diffused illumination without shadows would make the illumination so bad as to be practically useless. In other cases, as a drafting-room, where all the objects requiring distinction are in one plane, as the drafting board, and the distinction is exclusively by differences of color and intensity, but not by shadows, a perfectly diffused illumination is required, and shadows would be objectionable and misleading, and this is one of the cases where directed light is objectionable. While with a single light source all the light issuing from it is directed light, by using a number of illuminants, the overlap of their light fluxes causes more or less light to reach objects from all directions, and thereby gives the effect of diffused light, except at those places where the shadows cast by the different light sources coincide, and by proper positions of sufficient numbers of light sources this can be avoided. The use of a number of light sources thus offers a means of increasing the proportion of diffused to directed light. 116. It is not sufficient, however, to have merely a combination of diffused and directed light in the proper proportion, but the direction of the latter also is of importance. In some simple cases this is obvious, as, in writing, the directed light should be from in front on the left side above the table, so as not to cast the shadow on the work. The purpose of the shadow in illumination is to mark the edge of the object, and its height by the length of the shadow. The shadow, therefore, should not extend too far from the object to which it is related, other- wise it loses its close relation to it and becomes misleading and thereby interferes with good illumination. Thus the directed light should come from above, that is, in a direction making a considerable angle with the horizontal, so as to limit the length 268 RADIATION, LIGHT, AND ILLUMINATION. of the shadow without, however, being vertical, as the latter would largely obliterate shadows. Perhaps an angle of 45 to 60 degrees with the horizontal would be most satisfactory. The practically horizontal shadows cast in the usual form of street lighting therefore are not satisfactory for best illumi- nation. The number of shadows is of less importance. While in nature objects have one shadow only, cast by the sun, indoors we are familiar with seeing several shadows due to the diffused day- light from several windows. Of high importance, however, is the shape of the illuminant, in so far as it determines the outer edge of the shadow. The purpose of the shadow is to give an intensity difference at the edge of the object, and thereby make it easier to see the object. The shadow, however, has another edge, its outer end, and that we should not see, as no object ends there, or at least it must be such that it cannot be mistaken for the edge of an object. The problem thus is not merely to provide sufficient directed light to cast a shadow, but the shadow should be such that only one side, at the edge of the object, is sharply defined, while the other edge of the shadow, which ter- minates on the flat surrounding surface, should gradually fade or blur. If we have to look closely to determine that the outer edge of the shadow is not the edge of another object, the strain of distinguishing between the edge of an object and the edge of a shadow makes the illumination uncomfortable and thus unsatisfactory. In the shadows cast by a single arc in a clear glass globe, this difficulty of distinguishing between the edge of a shadow and the edge of an object is especially marked, and, combined with the invisibility of objects in the shadow, makes such shadows appear on first sight like ditches or obstructions. In the use of shadows in illuminating engineering it thus is necessary to have the outer edge of the shadows blur or gradually fade, and this requires that the source of directed light be not a point, but a sufficiently large area to scatter the light at the outer edge of the shadow, preferably even more than is the case with the shadows cast by the sun. This requires enclosing the illuminant by a fairly large opal globe or other similar device; that is, have the light issue from a fairly large luminous area. It must be recognized that the proper treatment of the shadows ILLUMINATION AND ILLUMINATING ENGINEERING. 269 is one of the most important problems determining the success or failure of an illumination. 117. Color sensitivity. The maximum of sensitivity of the eye shifts with decreasing illumination from yellow to bluish green, and where a low intensity of illumination is used, as in street lighting, a source of light which is rich in the shorter waves, that is, a white light, is superior in its physiological illuminating value to a yellow light of the same or even higher light flux, while inversely at high values of illumina- tion, as for decorative purposes, the yellow light is more effective. Therefore it is a mistake to choose a yellow light source for illumination of very low intensity, or a white or bluish-green light for illumination attempting high intensity effects. Thus, for the average street lighting of American cities, the white arc is superior to the yellow flame arc, but, to produce a glare of light, the latter would be superior. While there are further physiological effects which are of im- portance in illuminating engineering, the above four may illus- trate the long step which exists between the distribution of the light flux as measurable by the photometer, and the success or failure of the illumination represented by it. The requirements of satisfactory illumination can thus be grouped in two main classes, referring respectively to economy and to comfort, and the characteristics are: (1) General or uniform, and local or concentrated illumination, and combination of both. This is of importance for economy : to avoid the production of unnecessary light flux; and comfort: to reduce the effect of fatigue. (2) Diffused and directed illumination, and combinations of both, and the theory of the shadow. This is of importance for the comfort of illumination, in securing clearest distinction. (3) Quality or color of light, of importance in economy, to suit the color to the intensity of illumination, and to comfort, in increasing or softening differences in color shades. (4) Massed and distributed illumination, as controlling the distribution of the light flux, and thereby the economy and also the diffusion. (5) Direct illumination and indirect illumination, shaded, diffracted, diffused, or reflected light, in its relation to the bril- 270 RADIATION, LIGHT, AND ILLUMINATION. liancy of the light source, and thereby the effect of the contraction of the pupil, on economy and comfort. Some of the common mistakes made in illumination are : (1) Unsatisfactory proportion of general and of concentrated light. (2) Exposure of high brilliancies in the field of vision, as naked filaments. (3) Unsuitable proportion of diffused and directed light. (4) Improper direction of directed light and thereby improper length of shadows. (5) Sharp edges of shadows. In order to illustrate the preceding principles, some typical cases may be considered : (a) Domestic lighting. 118. Domestic lighting usually requires a combination of a concentrated illumination of fairly high intensity locally at the work-table, dining-table, etc., and a general illumination of low intensity, to secure comfort and economy. Occasionally, as in halls, etc., the local lighting is absent and only general illumina- tion required, while for instance in a sick room the general illumi- nation is absent and only local illumination required. In this illumination the proportion between directed and dif- fused light should be such as to give the proper effect of shadows. The problem of domestic illumination thus is to produce a defi- nite distribution of light flux density, with a definite proportion between diffused and directed light. If we deviate from the proper proportion on one side, the room appears cold and uncom- fortable; if we deviate in the other direction, it appears dark and gloomy. The light issuing directly from a single illuminant is directed light; the light issuing from a number of illuminants is diffused in proportion to the number of sources by the overlap of the light fluxes of the illuminants. The light reflected from walls and ceilings is diffused light. The proportion between the light reflected from walls and ceilings, or the indirect light, and the direct light from the illuminants, varies with the reflecting power of walls and ceilings, that is, their brightness or darkness. The proportion between directed and diffused light thus can be changed, and the diffused light increased by increasing the num- ber of illuminants, and also by increasing the brightness of walls ILLUMINATION AND ILLUMINATING ENGINEERING. 271 and ceilings. With a given brightness of walls and ceilings, the desired distribution of the light flux — a local high and general low intensity — can be produced by a single illuminant having the proper distribution curve of light flux. In this case, however, usually we get too much directed, and not enough diffused, light. The same distribution of light flux can be produced by a number of illuminants properly located : nearer together for the local than for the general illumination. In the latter case we get more diffused and less directed light, and thus by choosing the number of light sources it is possible, with any given brightness of walls and ceilings, to get the desired distribution of light flux and at the same time the proper proportion of directed and diffused light. With a different brightness of walls and ceilings, the dis- tribution curve of a single light source, required to give the desired light flux distribution, is correspondingly changed, and, the lighter the walls and ceilings, the more light is reflected, giving a diffused general illumination, and thus less direct light from the illuminant is required for the general illumination. With in- creasing reflecting power of walls and ceilings, the proportion of diffused light increases, and the number of light sources which are required to give the proper proportion between directed and diffused light is decreased, and inversely it is increased with increasing darkness of walls and ceiling. Therefore, in a room with light walls, a smaller number of light sources is required for good illumination than in a room with dark walls, assuming the same intensity of local and of general illumination. 119. The problem of domestic illumination: to get a certain distribution of illumination, with a definite proportion between directed and diffused light, thus leaves one independent variable — the brightness of walls and ceilings. This is necessary, as the problem of domestic illumination is twofold: to get the proper illumination by means of the daylight, and also to get it for artificial illumination. During daytime, the windows are the source of light, the directed light issues from the windows, the diffused light from the walls and ceilings and by the overlap of the light from several windows. The proper distribution between local and general illumination during daytime, and at the same time the proportion of directed and diffused light, thus deter- mines the number of windows and the brightness of walls and ceilings, in the manner as discussed before. 272 RADIATION, LIGHT, AND ILLUMINATION. As the reflecting power of walls and ceilings is fixed by day- light considerations, it cannot be chosen, or at least only to a limited extent, by considerations of artificial illumination, but, as found above, this is not necessary, since by a combination of a suitable number of light sources of proper distribution curves the problem of artificial illumination may be solved. To some extent, due to the quality of artificial light and daylight, the walls can give a different reflecting power for the one than for the other. As artificial light is deficient in blue and green, a bluish or greenish shade of walls and ceilings gives them a greater reflect- ing power for daylight than for artificial light — which usually is desirable — and inversely with a reddish-yellow shade. (b) Street Lighting. 120. The problem of street illumination is to produce a uni- form low intensity. For reasons of economy, the intensity must be low, at least in American cities, in which the mileage of streets, for the same population, usually is many times greater than in European cities, and, at the same time, the same type of illumi- nant is usually required for the entire area of the city. The low intensity of illumination requires the quality of light which has the highest physiological effect at low densities, that is, white light, and excludes the yellow light as physiologically inefficient for low intensities. Still better would be the bluish green of the mercury lamp, but is not much liked, due to its color. Quite satisfactory also is the greenish yellow of the Welsbach mantel for these low intensities. The American practice of preferring the white light of the carbon or magnetite arc thus is correct and in agreement with the principles of illumination, and the yellow-flame arc can come into consideration — even if it were not handicapped by the necessity of frequent trimming — only in those specific cases where a high intensity of illumination is used, as would be only in the centers of some large cities. Uniformity of illumination is specially important in street light- ing, where the observer moves along the street, and, due to the low intensity, the decrease of subjective illumination by fatigue is especially objectionable. For a street illuminant, a distribu- tion curve is required which gives a maximum intensity some- what below the horizontal, no light in the upper hemisphere, and very little downward light. Street lamps therefore should be judged and compared by the illumination given midways be- ILLUMINATION AND ILLUMINATING ENGINEERING. 273 tween adjacent lamps, or at the point of minimum intensity, or, in other words, by the intensity in a direction approximately 10 deg. below the horizontal. This also is in agreement with American practice. However, it is very important that the downward intensity be very low, and in this respect it is not always realized that the light thrown downward is not merely a waste of light flux, but is harmful in producing a glaring spot at or near the lamp and, by the fatigue caused by it, reducing the effective illumination at the minimum point between the lamps. Most objectionable in this respect is the open direct current car- bon arc and those types of lamps giving a downward distribution, but even with the enclosed arc lamp the distribution of light on the street surface is still far from uniform, and the intensity too high near the lamp, and in this respect improvements are desirable. 121. The greatest defects of the present street illumination, which frequently makes it inferior in subjective illumination even to the far lower illumination given by the full moon, are the absence of diffused light, and especially the improper direction and termination of the shadows, and also the high brilliancy of the illuminant. The light of the usual street lamp is practically all directed light, issuing in a nearly horizontal direction from a point source. Thus the shadows are far longer than permissible, and terminate sharply and without blur; objects in the shadows are practically invisible, and the end of the shadow looks like the edge of an object, thus producing a misleading effect, which results in unsatisfactory illumination. To give a somewhat better direction to the light requires considerable increase of the height of the lamp above the street surface. This also would essentially decrease the intensity of illumination below and near the lamp, without appreciably affecting the intensity at the minimum point, and thus would give a more uniform and thereby better illumination. No valid reason usually exists against greatly increasing the height of the lamps, except that of the greater cheapness of short lamp posts, which is hardly justifiable. It is, however, more difficult to give a proper blur to the ends of shadows, so as to distinguish them from edges of objects. This would require an increase of the surface of the illuminant, by opal or frosted globe, etc. Enclosing the arc by an opal globe, however, scatters the light more uniformly in all directions, and 274 RADIATION, LIGHT, AND ILLUMINATION. thereby spoils the distribution curve, and interferes with the required uniformity of illumination : with an opal globe, the intensity in the downward direction does not differ very much from that in the horizontal, while with lamps 20 feet above the street level, and at distances of 200 feet from each other, the downward intensity for uniform illumination should be not much more than one-twenty-fifth of that under an angle of sin $ = 20 — = 0.2; or 12 deg. below the horizontal. Very much better is -L \J{J the effect of a frosted or sand-blasted globe. The best way of maintaining a proper distribution curve and at the same time diffusing the light, so as to reduce its brilliancy and blur the shadows, appears the use of prismatic diffraction, on the principle of the Fresnel lenses of lighthouses (holophane). Obviously, where the lamps are close together, as in the center of large cities, their light fluxes overlap and thereby give a better diffusion, and, at the same time, the midway point between lamps is under a greater angle against the horizontal ; thus a more downward dis- tribution of the light flux permissible. For the largest part of American street lighting, however, this does not apply. 122. In the early days of using arc lamps for American city lighting, lighting towers were frequently used, and such tower lighting has still survived in some cities. One or a number of arc lamps are installed on a high tower and were supposed from there, like artificial suns, to spread their light over an entire city district. This method of city lighting was found unsatisfactory, as it did not give enough light. It is unsatisfactory, however, not in principle, but because it was too ambitious a scheme. If, in street illumination, we double the distance between the lamps, each unit must have four times the light flux to get the same minimum flux density, as the distance is doubled, and the flux density decreases with the square of the distance. At twice the distance between the lamps, each lamp thus must have four times the light flux, and each mile of street thus requires twice the power. Reducing the distance between lamps to one- half reduces the power to one-half with the same minimum illumination. In street lighting it is therefore of advantage to use as many units of illuminants as possible, and bring them together as close as possible, and correspondingly lower their ILLUMINATION AND ILLUMINATING ENGINEERING. 275 intensity, up to the point where the increasing cost of taking care of the larger number of units and increasing cost of poles and connections compensates for the decreasing cost of energy. There is a minimum which probably is fairly near our present practice. When, however, you come to square and exposition lighting, you find that the distance between the illuminants has no effect on the efficiency. Let us assume that we double the distances between the lamps which light up a large area. Then each lamp requires four times the light flux to get the same minimum flux density between the lamps, but at twice the distance between the lamps each lamp illuminates four times the area, and the total power per square mile of lighting a large area, like an exposition, thus is independent of the number of lamps used, and, whether you place them close together or far apart, you require the same total flux of light, and if you keep the same proportions of height from the ground and distance between lamps, you also get the same variation between maximum and minimum intensity. But, supposing the lamps to be placed further apart, the maximum or minimum points also are further apart, and you get a more satisfactory illumination by having a less rapid intensity variation. That points to the conclusion that, for exposition lighting, the most efficient way would be to use a relatively moderate number of high-power sources of light on high towers at distances from each other of the same magnitude as the height of the towers. We would get a greater uniformity and better physiological effect by having the illumi- nants further apart, and they would require the same total light flux, and therefore the same power, as if you bring the lamps close to the ground, and place them very close to each other. The tower lighting therefore is the ideal form for lighting a large area. When the arc was first introduced, it was so much superior to any other illuminant known before, that people vastly over- rated it. They thought that they could light the whole city by it, and in trying to do so these towers would have been the proper way, but very soon it was found that even with the effi- ciency of the arc, to light not only the streets, but the whole area of the city, would require an entirely impracticable amount of light flux. It thus was too ambitious a scheme for city lighting, but it should be done in exposition work. City illumi- 276 RADIATION, LIGHT, AND ILLUMINATION. nation thus has come down from this first ambition to light the whole city to an attempt to light only the streets. For the latter purpose, however, lighting towers are inefficient, since much of the light flux is wasted on those places which we no longer attempt to light. In exposition lighting, however, the most effective general illumination would be given by white arcs on high towers, leaving the concentrated or decorative illumination to the incandescent lamp and flame arc, of yellow color.