Aerial Haze and its Effect 
 on Photography from the Air 
 
 Monographs on the Theory 
 of Photography, from the 
 Research Laboratory of the 
 Eastman Kodak Company 
 
 No. 4 
 
Monographs on the Theory of Photography from the 
 Research Laboratory of the Eastman Kodak Co. 
 
 No. 4 
 
 Copyright 1923 
 Eastman Kodak Company 
 
Aerial Haze and its Effect on 
 Photography from the Air 
 
 ILLUSTRATED 
 
 D. VAN NOSTRAND COMPANY 
 NEW YORK 
 
 EASTMAN KODAK COMPANY 
 ROCHESTER, N. Y. 
 
 l 9 2 3 
 
Monographs on the Theory oe Photography 
 
 Edited by 
 
 C. E. Kenneth Mees 
 Mildred Spargo Schramm 
 and 
 
 Elsie L. Garvin 
 
 THfi GETTY RESEARCH 
 INSTITUTE LIBRARY 
 
Monographs on the Theory 
 of Photography 
 
 No. 1 The Silver Bromide Grain of Photographic 
 Emulsions. By A. P. H. Trivelli and S. E. Shep¬ 
 pard, D. Sc. 
 
 No. 2 The Theory of Development. By A. H. Nietz. 
 
 No. 3 Gelatin in Photography, Volume I. By S. E. 
 Sheppard, D. Sc. 
 
 No. 4 Aerial Haze and its Effect on Photography 
 from the Air. 
 
 Price Each $2.50 
 
 . Other volumes soon to appear: 
 
 The Physics of the Developed Photographic Image 
 
 By F. E. Ross, Ph.D. 
 
 Gelatin in Photography, Volume II. By S. E. Sheppard, 
 D. Sc. 
 
Preface to the Series 
 
 The Research Laboratory of the Eastman Kodak Company 
 was founded in 1913 to carry out research on photography 
 and on the processes of photographic manufacture. 
 
 The scientific results obtained in the laboratory are pub¬ 
 lished in various scientific and technical journals, but the work 
 on the theory of photography is of so general a nature and 
 occupies so large a part of the field that it has been thought 
 wise to prepare a series of monographs, of which this volume 
 is the fourth. In the course of the series it is hoped to cover 
 the entire field of scientific photography, and thus to make 
 available to the general public material which at the present 
 time is distributed throughout a wide range of journals. Each 
 monograph is intended to be complete in itself and to cover 
 not only the work done in the laboratory, but also that 
 available in the literature of the subject. 
 
 A very large portion of the material in these monographs 
 will naturally be original work which has not been published 
 previously, and it does not necessarily follow that all the 
 views expressed by each author of a monograph are shared 
 by other scientific workers in the laboratory. The monographs 
 are written by specialists qualified for the task, and they are 
 given a wide discretion as to the expression of their own 
 opinions, each monograph, however, being edited also by the 
 director of the laboratory, by Mrs. Schramm, and by Miss 
 Garvin, who is now the active editor of the series. 
 
 Rochester, New York 
 August, 1923 
 
Preface 
 
 In 1918, the Research Laboratory of the Eastman Kodak 
 Company undertook, in collaboration with the Department 
 of Military Aeronautics of the United States Army, a study 
 of photography from the air with regard especially to the 
 problems presented by aerial haze, the immediate object of 
 the study being, of course, its application to military aerial 
 photography. 
 
 The work in the laboratory was under the immediate 
 direction of Mr. Kenneth Huse, who was assisted by Drs. 
 Walter Colby and William Sleator, of the University of 
 Michigan, who spent a year in the Research Laboratory for 
 the purposes of this investigation. Many other members 
 of the staff of the laboratory collaborated in the work, which 
 was under the general direction of Dr. C. E. K. Mees, the 
 director of the Research Laboratory. 
 
 Captain A. K. Chapman, stationed in Rochester for the 
 Science and Research Division, collaborated in all the aerial 
 work of the Eastman Kodak Company and especially in the 
 arrangements necessary for the carrying out of this investiga¬ 
 tion. Captain H. E. Ives, who was in general charge of the 
 aeroplane investigations of the Science and Research Division, 
 also assisted greatly in the work. Thanks are also due 
 especially to Lieutenant A. H. Nietz who, while stationed at 
 the Langley Flying Field at Ltampton, Va., did a great deal of 
 flying in connection with the work, and placed many of his 
 results at our disposal. 
 
 The work had not been completed when the Armistice was 
 signed, in November, 1918. By consent of the military 
 authorities, it was brought to a conclusion sufficient to enable 
 general results of practical value to be deduced from the 
 measurements, and the work was then closed, though in 
 only a partially completed state. Nevertheless, it has been 
 thought desirable to publish the results obtained and especially 
 the methods employed, so that any future work along these 
 and similar lines may have the advantage of the experience 
 gained. We feel that in spite of the incomplete state of the 
 work, the results obtained are of considerable value as indi¬ 
 cating the best conditions for aerial photography, the types 
 of materials which should be used, and the conditions under 
 which they should be employed. 
 
 We are indebted to the Department of Military Aeronautics 
 for its kind permission to publish the work in this monograph. 
 
 Rochester, New York 
 August, 1923 
 
Aerial Haze and its Effect on Photography 
 from the Air 
 
 CONTENTS 
 
 Page 
 
 Preface . 7 
 
 Chapter I. Introduction.11 
 
 Chapter II. The methods of photographic 
 
 photometry.19 
 
 Chapter III. The measurement of aerial haze . . 36 
 
 Chapter IV. The duplication of aerial conditions in 
 
 the laboratory.54 
 
 Chapter V. The material and conditions best adapted 
 
 for aerial photography.64 
 
 Appendix The haze effect produced by pure dry air 72 
 
 Bibliography . 79 
 
 Index of Authors. 80 
 
 Index of Subjects . 82 
 

 
Aerial Haze and its Effect on Photography 
 from the Air 
 
 CHAPTER I 
 Introduction 
 
 ' Aerial photography differs from most branches of photogra¬ 
 phy in that the distance between the camera and the object 
 photographed is very great, so that between the lens and the 
 subject there exists a depth of air which reflects light back into 
 the lens. This light, reflected from particles of water or dust 
 suspended in the air and, to a less extent, from the molecules 
 of air themselves, superimposes a uniform illumination over 
 all parts of the subject photographed, causing a veiling haze 
 which diminishes the contrast available for reproduction in 
 the photograph. The term “haze,” though indefinite, specifies 
 a phenomenon with which every one is familiar. Distant 
 landscapes in regions of ordinary humidity almost always 
 show haze, which appears as a thin veil of bluish cast. 
 
 In photographing mountains haze frequently causes trouble, 
 diminishing the contrast to a serious extent. In no branch of 
 photography, however, is haze so important as in photography 
 from the air. Owing to the large proportion of actinic light 
 reflected in haze, its effect is much greater photographically 
 than visually, often being serious in photographs taken under 
 visually clear conditions. At altitudes of 10,000 feet or more, 
 the haze effect is frequently so marked that it becomes difficult 
 to distinguish ground detail in the photograph, while the 
 taking of long distance obliques is almost impossible without 
 precautions for the elimination of haze. 
 
 The following series of photographs illustrates quite force¬ 
 fully the very undesirable results arising from the presence of 
 haze in the air. Fig. 1 shows two successive photographs of 
 the same subject taken from a height of 10,000 feet. Fig. 1A 
 was made on an ordinary plate with no special means of dealing 
 with haze. Fig. IB was made on a special plate under condi¬ 
 tions such that the haze effect was largely removed. The 
 prints from these negatives were made without any special 
 treatment. 
 
 Fig. 2 shows two photographs made on the same emulsion 
 at exactly the same time by means of a camera with multiple 
 
 11 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 lenses. Fig. 2A was made in the ordinary way; Fig. 2B shows 
 the elimination of the haze effect. It is quite surprising how 
 
 Fig. 2A 
 
 Taken on ordinary plate 
 
 much more detail, especially in the woods and fields (the dark 
 areas), is shown by the picture less affected by haze. These 
 photographs were taken at an altitude of 11,000 feet, and it is 
 interesting to note that haze was not apparent to the eye, the 
 day being considered a very good one for air work. 
 
 Fig. 3 shows a case of much denser haze, although still not 
 of haze at its maximum. In this case the altitude was 10,500 
 
 12 
 
AERIAL HAZE 
 
 feet. Clouds were rather uniformly distributed in a layer at 
 about 6,500 feet. Looking downward from a height of 10,500 
 
 Fig. 2B 
 
 Taken with filter 
 
 feet, the earth was almost invisible, as the spaces between the 
 clouds were apparently occupied by haze. These spaces 
 appeared very blue. 
 
 There are times when haze (in the absence of any cloud 
 formations) becomes so dense that the earth can scarcely be 
 seen, even from low altitudes. This is more likely to occur in 
 autumn than in spring, the haze consisting to a greater 
 
 13 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 extent of dust and smoke particles than of water vapor. 
 Certain regions would, of course, never experience haze of 
 this sort, and, furthermore, some districts would be on the 
 average quite free from haze in troublesome quantities. The 
 southwestern part of the United States is quite free from 
 haze, and there no difficulty is experienced, as a rule, in 
 securing clear aerial photographs by ordinary methods. It 
 should not be understood that haze is always present in suffi¬ 
 cient amount to annoy the photographer. From lower alti- 
 
 Fig. 3A 
 
 tudes little difficulty is encountered from this cause, but at 
 high altitudes better photographs can always be secured by 
 considering the haze effect. 
 
 There is no likelihood of our being able to photograph 
 through extremely dense layers of haze, since radiation (of 
 whatever wave-length) reflected from the haze body will often 
 be greater than that reflected from the earth and transmitted 
 by the haze. Therefore, when anyone speaks of photo¬ 
 graphing through several miles of “mist” or “fog” he is either 
 misapplying the terms mist and fog to what we usually con¬ 
 sider haze, or he is working on the credulity of the general 
 
 14 
 
AERIAL HAZE 
 
 public, which has not yet become familiar with the process. 
 It is impossible by any method known at present (such, for 
 instance, as photography by the infra-red) to photograph 
 through mist which is any more dense than that through 
 which one can just distinguish the highlights of the subject 
 to be photographed. 
 
 Our visual perception of space depends upon contrast—that 
 is, upon differences of brightness. While it is often stated 
 
 Fig. 313 
 
 Showing elimination of haze 
 
 that objects are visible by virtue of the light which they 
 reflect, it is more precise to say that such visibility is due to 
 the differences in the light which they reflect. This is evident 
 if we consider an object of exactly the same color and bright¬ 
 ness as the background against which it is viewed. Under 
 such conditions the object is indistinguishable from the back¬ 
 ground and hence is invisible. Difference in either the 
 intensity or the quality of the light is sufficient for the produc¬ 
 tion of visibility. Intensity differences cause objects to 
 
 15 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 appear of various brightnesses, while variations in the quality 
 of the light are perceived by the eye as differences of color. 
 
 Since the photographic process can not reproduce color, 
 variations of quality are translated by it into brightness 
 differences, the translation depending upon the color sensi¬ 
 tiveness of the photographic materials. Variations of bright¬ 
 ness are produced in nature either by differences in the 
 reflecting power of surfaces subject to the same illumination 
 or by differences in the illumination of various parts of a surface 
 having a uniform reflecting power. Black ink reflects less 
 light than white paper, and so with the same light falling on 
 each there is a difference in brightness by which ink marks 
 are distinguishable. On the other hand, if a plaster cast be 
 lighted from one side, parts of the surface will be in shadow, 
 and although the whole cast has the same reflecting power, 
 there will be differences in brightness by virtue of which the 
 details of the cast will be visible. 
 
 Therefore any diminution of the contrast between the 
 different parts of a subject will diminish the ability of the 
 observer to discriminate detail; and, since the object of aerial 
 photography is the discrimination of detail in the subject 
 photographed, the effect of haze is to produce a most serious 
 loss in the efficiency of the process. The study of haze and 
 of the best methods of diminishing its photographic effects 
 thus becomes of primary importance for aerial photography. 
 
 The factors in aerial photography which can be controlled 
 are: 
 
 1. The color of the light by which the photograph 
 
 is taken; 
 
 2. The properties of the sensitive material used 
 
 for making the negative; 
 
 3. The exposure given; 
 
 4. The developer and the time of development. 
 
 Of these factors, the most important is 1, but the choice of 
 the color to be used is necessarily conditioned by the color 
 sensitiveness of the sensitive materials available. 
 
 As will be seen later, light of short wave-lengths is much 
 more strongly reflected by the small particles of the air than 
 is light of longer wave-lengths, so that the longer the wave¬ 
 length—that is, the redder the light used—the less will the 
 haze effect be. Unfortunately, photographic materials are 
 much more sensitive to light of short wave-lengths than to the 
 longer waves, and the red and yellow light can, therefore, be 
 employed only on materials sensitized especially for the pur- 
 
 16 
 
AERIAL HAZE 
 
 pose and even with such materials only with an increased 
 exposure. 
 
 Since the exposure which can be given in aerial photography 
 is limited by the conditions of the work, the use of red or yellow 
 light for the elimination of haze is also limited, and it becomes 
 of importance to ascertain how far the improvement in contrast 
 can be effected by the correct selection of factors 2, 3 and 4, 
 in spite of the greater importance of the factor of color. 
 
 The work dealt with in this monograph was undertaken in 
 order to determine primarily the types of photographic 
 materials most suitable for photography through haze and the 
 conditions under which those materials would be used, and, 
 secondarily, to study the occurrence, distribution and value 
 of haze. 
 
 It was decided at the outset that the problem could best be 
 attacked in three definite stages. 
 
 (1) The extent and distribution of the haze and its spectral 
 quality can be determined only by observations from aero¬ 
 planes under different weather conditions and at different 
 altitudes;- and since the quality must be considered from a 
 photographic point of view, it was decided to use photographic 
 methods for this investigation. 
 
 (2) The behavior of photographic materials when used for 
 photographing a subject having a general veiling haze super¬ 
 imposed was studied in the laboratory by means of a so-called 
 "haze cabinet”, in which a camera was arranged to photpgraph 
 any set of diagrams required, while by means of a semi¬ 
 transparent mirror before the lens a uniform intensity from a 
 separate light source could be superimposed upon the image. 
 With this apparatus it was possible to investigate the advan¬ 
 tages and disadvantages of different photographic materials 
 under various haze conditions (amounts of veiling haze), 
 and to determine the conditions of exposure and development 
 which would give the best results with a given aimunt of 
 haze. The calibration of this apparatus on the basis of the 
 data obtained from the air as to the amount of veiling haze 
 present under varying conditions, enables decisions to be 
 made in the laboratory concerning the materials which should 
 be used for field work, and the treatment which those materials 
 should receive. 
 
 (3) The utility of various materials from the point of view 
 of photographing through haze having been investigated by 
 means of the haze cabinet, these materials were subjected to 
 the standard sensitometric methods of investigation. The 
 
 17 
 
MONOGRAPHS ON THE THEORY OE PHOTOGRAPHY 
 
 properties shown by the sensitometric data were correlated 
 with the behavior of the materials in the haze cabinet, thus 
 making it possible to deduce from the ordinary sensitometric 
 tests the suitability which a given material has for aerial work. 
 
 Only a part of the work planned was accomplished. Methods 
 for the study of haze were devised, and some preliminary 
 measurements from the air were made. A much more com¬ 
 plete study, correlated with meteorological data, is, however, 
 necessary, and should be undertaken in the interests both of 
 meteorology and of aerial photography. 
 
 The types of sensitive material suitable for aerial photog¬ 
 raphy were determined and criteria for their suitability 
 established, as were also the conditions of exposure and 
 development which give the best results. 
 
 18 
 
CHAPTER II 
 
 The Methods of Photographic Photometry 
 
 When a photographic material such as a film or plate is 
 exposed to an increasing series of intensities, the deposit pro¬ 
 duced upon development depends upon the intensity acting 
 during exposure, so that if a strip of the material is exposed 
 in steps, these steps are represented by a series of densities in 
 the developed plate. The property of the developed density 
 which is of interest from the photographic point of view is 
 its light-stopping power. This can be measured by means of 
 a photometer, in which the light transmitted through the clear 
 portion of the image is equalized, by means of some optical 
 device, with that transmitted by the density to be measured. 
 
 When the densities thus 
 measured are plotted against 
 the exposures, a logarithmic *' 5 
 scale of exposures being chosen, 
 it has been found that a curve 
 of the general type shown in Fig. 
 
 4 is obtained. Throughout a 
 considerable portion of its ° 
 range, the rise in density is seen 10 
 to be proportional to the loga¬ 
 rithmic increase in exposure, - 5 
 and through this portion a 
 straight line can be drawn cor¬ 
 responding to the equation 
 
 D = y (log E - log «'), 
 
 where y represents the slope of the straight line—that is r 
 the tangent of the angle which it makes with the exposure 
 axis — and log i is the intercept on the exposure axis. 
 
 This characteristic curve of a photographic material was 
 first described by Hurter and Driffield, who found that for 
 normal development the point of the intercept of the straight 
 line on the exposure axis remains invariable, and that a change 
 in the time of development corresponds only to a change in 
 the value of y; that is, of the slope of the straight line. For 
 this reason, the exposure i corresponding to the intercept log i 
 was called by them the inertia of the photographic material, 
 while y was termed the development factor. 
 
 19 
 
 LOQ E 
 
 Fig. 4 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 According to Hurter and Driffield the sensitiveness of a 
 material is defined by dividing the value of the inertia into 
 some constant. In the lower portion of the characteristic 
 curve before the straight line is reached, the curve is convex 
 to the exposure axis, increasing in slope from a zero value to 
 that of y. This region was termed by Hurter and Driffield 
 the region of under-exposure, while the corresponding region, 
 where the curve becomes concave to the exposure axis and 
 the slope diminishes from y until it finally becomes zero, 
 was called the over-exposure region. These names arise from 
 the fact that it is only through the straight-line portion of the 
 curve that the opacities of a negative correspond inversely 
 to the brightnesses of the original subject, and, consequently, 
 it is throughout this region that the reproduction of the tones 
 of the subject is correct in a negative. Any portion of the 
 scale of brightnesses which falls upon the under- or over¬ 
 exposure portions of the curve produces an incorrect and dis¬ 
 torted rendering of the scale of tones . 1 
 
 The whole characteristic curve can, of course, be expressed 
 by means of a mathematical equation, and several such 
 equations have been proposed by photographic workers . 2 
 Such equations are integrations of differential equations 
 derived from assumptions as to the absorption of light by the 
 film and as to the distribution of sensitiveness among the 
 grains composing the sensitive emulsion. 
 
 The characteristic curve can obviously be used for the 
 translation of the density of any photographic deposit into 
 the corresponding intensity by which it was produced. Thus, 
 if we photograph from the air two areas such, for instance, as a 
 wooded area and an open field, and impress upon the plate 
 before development a known scale of intensities, then, after 
 development, we can measure the densities of the known 
 scale, draw the characteristic curve, measure the densities 
 corresponding to the areas in which we are interested, and by 
 interpolating these densities upon the curve deduce the light 
 intensities corresponding to the areas. 
 
 Photographic photometry thus consists of the determination 
 of the intensity of the exposures to be measured by interpola¬ 
 tion of the densities produced by them on a scale of densities 
 produced by known intensities. It is obvious that the accu¬ 
 racy of photometric measurements made in this way will 
 depend upon the scale of densities from known exposures 
 
 1 Jones, L. A., On the theory of tone reproduction. J. Frankl. Inst. 190 : 39. 1920. 
 
 2 Ross, F. E., On the relation between photographic density, light intensity, and exposure 
 
 time. J. Opt. Soc. Amer. 4: 255. 1920. 
 
 20 
 
AERIAL HAZE 
 
 being produced under conditions which are exactly the same 
 as those under which the densities to be measured are produced. 
 We must eliminate (1) variations due to the material—i. e., 
 irregularities in sensitiveness, thickness of coating, etc.; 
 
 (2) variations due to the treatment—i. e., differences in 
 developing the intensity scale and the densities to be measured; 
 
 (3) variations in the intensity or time of exposure of the two 
 scales (it is not justifiable to assume that time and intensity, 
 the two components of exposure, are reciprocally equivalent); 
 
 (4) variations due to the quality of the light. The scale must 
 be made by light of the same wave-length as that which 
 produced the exposures to be measured. Let us consider 
 these factors in turn: 
 
 1. In order that variations in the material may be reduced 
 to a minimum, it is advisable to impress the intensity scale 
 and the densities to be measured upon the same plate or film. 
 This automatically eliminates variations due to different times 
 of development. Where this is not possible; the exposures 
 should be made on material from the same package, which 
 receives identical treatment before and after exposure. When 
 exposures are made upon the same material, the only variations 
 likely to occur are due (1) to differences of sensitiveness in 
 different parts of the material, generally produced during the 
 drying process. Thus there is usually an area of somewhat 
 greater sensitiveness around the edges of the plates than in 
 the middle; and (2) variations in the thickness of the coating 
 such as may occur, especially with plates, owing to waviness 
 of the glass on which the emulsion is coated. 
 
 In the course of a study of the uniformity which may be 
 expected in different parts of the same photographic plate, 
 the average deviation for white light was found to be of the 
 order of four per cent, and the maximum deviation about 
 eight per cent. When panchromatic plates were used with 
 color filters, and the plates exposed to red, green, and blue 
 light, greater deviations were observed for red than for green, 
 and for green than for blue light. The maximum deviation 
 in this series of experiments was ten per cent for blue light 
 and fifteen per cent for red light. It is clear that the average 
 precision of a single measurement upon plates by the methods 
 of photographic photometry cannot be expected to be greater 
 than five per cent, owing to variations in the sensitiveness or 
 coating of the material. With film, the precision is somewhat 
 greater, as the emulsion is coated upon a wide band of support, 
 and differences in sensitiveness due to drying conditions are 
 much less likely to occur than in the case of plates. The 
 
 21 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 waviness characteristic of the glass on which ordinary plates 
 are coated also does not occur with films. The average devi¬ 
 ation for portrait film is found to be three per cent, the max¬ 
 imum deviation approximately six per cent, both these 
 measurements being made to white light. 
 
 2. The Effect of Treatment. During development, the slope 
 of the straight-line portion of the characteristic curve (y) 
 increases, the straight line rotating around the inertia point. 
 The slope can not, however, be increased indefinitely by 
 increase of development, as y increases from a zero value to a 
 limiting value, y m , which depends upon the nature of the 
 material. See Fig. 5. The increase of y with time is ex¬ 
 ponential, corresponding to the equation y = | or. in 
 
 its logarithmic form, 
 
 v 1 , Too 
 
 K = — log- 
 
 t y oo 
 
 T 
 
 Fig. 5 
 
 The value of y oo may be as 
 lowas 1.5 for fast materials, such 
 as portrait plates or film; for the 
 average plate used in general 
 photography it will vary from 
 1.8 to 2.3; while for process 
 plates, having the greatest con¬ 
 trast, it will generally exceed 
 3.0. The constant K which, 
 together with y oo, determines the velocity of development and 
 represents the velocity constant of the chemical reaction, is 
 dependent upon the nature and concentration of the devel¬ 
 oper, upon the temperature, and upon the material. 
 
 It is convenient to adjust development so that a y of 
 unity is obtained in a time of about five minutes. When the 
 developer contains a soluble bromide, the intercept of the 
 straight-line portion of the characteristic curve upon the 
 exposure axis is no longer fixed, but decreases during develop¬ 
 ment to a limiting position corresponding at y oo to the normal 
 position obtained in the absence of bromide. (Fig. 6.) It 
 can be shown that all the straight-line portions of the curves 
 obtained with different times of development in a bromided 
 developer meet at a point, just as with a non-bromided devel¬ 
 oper; but this point is no longer situated upon the exposure 
 axis, but is below it. The depression is a measure of the 
 
 22 
 
AERIAL HAZE 
 
 effect of the bromide upon the reduction potential of the 
 developer. 1 (Figs. 7A and B.) 
 
 From this discussion of the variation of the characteristic 
 curve with development it will be seen that in photographic 
 
 Fig. 6 
 
 photometry it is essential to obtain uniform development 
 for all densities which are to be compared with each other. 
 It is especially necessary to avoid accidental development 
 effects such as may be produced by eddy currents in the 
 developer, by stationary waves, or by local exhaustion. In 
 spite of all precautions there are certain defects produced by 
 the effect of the surrounding densities upon the concentration 
 of the developer, which may affect the value of density deter¬ 
 minations by photographic methods. These have been 
 studied recently by a number of workers, to whose papers 
 the reader is referred. In the work on aerial photography, 
 
 1 Nietz, A. H., The theory of development. Monograph No. 2. 
 
 23 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 it was believed that these effects were not of sufficient import¬ 
 ance to require special investigation, and they were ignored. 1 
 
 Fig. 7A 
 
 1 Ross., F. E., The mutual action of adjacent photographic images. Astrophys. J. 53: 349. 
 1921. Eberhard, G., Ueber die gegenseitige Beeinflussung benachbarter Felder auf 
 einer Bromsilberplatte. Physik. Zeits. 13: 288. 1912. 
 
 24 
 
AERIAL HAZE 
 
 3. Up to the present, the exposure has been treated as if it 
 were a definite quantity, but an exposure necessarily involves 
 two factors, the intensity of the light acting and the time for 
 which it acts. Thus, 
 
 E = f-It. 
 
 The simplest assumption with regard to these two factors is 
 that the variation of one of them can be exactly compensated 
 by a reciprocal increase in the other, so that we can write 
 
 E = I t. 
 
 On investigation it is found, however, that this simple relation 
 is not true, especially when t is large, and for values of t exceed¬ 
 ing 100 seconds the relation certainly can not be held to be 
 valid. Schwarzschild suggested that a closer approximation 
 to the relation would be given by the relation 
 
 E = I t p , 
 
 but it is doubtful whether this is any more than a further 
 approximation. As Ross has pointed out, if this equation can 
 be assumed to hold it is easy to determine p, since if we impress 
 on a plate two scales, one varying the time and the other 
 varying the intensity, then 
 
 y for the time scale 
 ^ y for the intensity scale 
 
 In photographic photometry, however, the difficulty can be 
 avoided completely by using an intensity scale when measuring 
 intensities, since in this way the exposures corresponding to 
 given densities are strictly comparable as regards intensity. 
 It must be remembered, however, that the function is probably 
 dependent upon the absolute time of exposure, so that not 
 only must the relative intensity scale be correct, but the 
 absolute intensity must be of the right order, since if the 
 densities to be measured were impressed, for instance, in 
 1 /100 of a second by exposure from an aeroplane and the 
 density scale was impressed by an exposure of several minutes 
 in the laboratory, the two intensity scales would not neces¬ 
 sarily be comparable. 
 
 4. Vaviations in the Quality of the Light. The characteristic 
 curve varies in shape and in slope with the wave-length of 
 the light producing it. This is shown in the range of wave¬ 
 lengths used with ordinary plates in Fig. 8, which is taken 
 
 25 
 
MONOGRAPHS ON THE THEORY OE PHOTOGRAPHY 
 
 from a paper by F. E. Ross. 1 It is seen that both the slope 
 and the shape of the curve vary considerably with the wave¬ 
 length of the light. The general relation between y and 
 wave-length’is shown in Fig. 9 which is also from Ross’s paper. 
 It is evident that the variation of y with wave-length is so 
 considerable that photometric measurements can be relied 
 upon only when the exposure scale is made with light of exactly 
 the same quality as that by which the densities to be inter¬ 
 polated were produced. This is well shown in the following 
 experiment: 
 
 A plate was exposed through three color filters, one trans¬ 
 mitting light from 460^ upwards, the second from 510^ 
 
 Fig. 8 
 
 Variation of the characteristic curve with wave-length of light 
 
 upwards, and the third from 555;j.p. upwards. The exposures 
 were made to a tungsten glow lamp with an intensity scale of 
 neutral densities, the lamp being run at different currents, 
 thus changing its spectral distribution: 
 
 .r 
 
 Filter Filter Filter 
 
 Current No. 1 No. 2 No. 3 
 
 1.5 amps. 1.67 1.79 1.87 
 
 1.8 amps. 1.43 1.61 1.80 
 
 2.1 amps. 1.35 1.54 1.75 
 
 1 Ross, F. E. t Photographic photometry and the Purkinje effect. Astrophys. J. 52: 86. 
 1920. 
 
 26 
 
AERIAL HAZE 
 
 An increase of current in the glow lamp shifts the maximum 
 of the spectral distribution curve to shorter wave-lengths, 
 and it will be noticed that y decreases as the current is 
 
 increased and increases as the 
 wave-length of the filter absorp¬ 
 tion limit increases, so that the 
 highest y is obtained with the 
 reddest filter at the lowest cur¬ 
 rent, and the lowest y with the 
 lightest filter at the highest cur¬ 
 rent. This experiment shows 
 the great errors which may be 
 introduced by changes in the 
 color of the light by which the 
 scale is impressed upon the 
 materials,and, as will be seen in 
 Chapter III, that very special 
 precautions are necessary to 
 ensure that the comparison 
 scale is impressed by light of exactly the same quality as that 
 reflected from the objects to be photographed. 
 
 METHODS OF MEASUREMENT 
 
 Exposure scales are produced by instruments known as 
 sensitometers, and they are expressed usually in c.m.s., the 
 unit representing the light of a one-candle-power standard 
 lamp at a distance of one meter for one second. No definition 
 of the photographic candle is accepted generally at the present 
 time, and since light sources which are visually identical have 
 quite different photographic effects, according to the quality 
 of the light emitted, there can be said to be no definite unit of 
 exposure. Each laboratory adopts its own unit expressed in 
 candle-power of some standard lamp such as the pentane 
 lamp, an electric lamp burning at a given temperature, or an 
 acetylene burner of a specific type. Fortunately, in the work 
 under discussion, the absolute magnitude of the intensities 
 was of no importance, and the exposures could be expressed in 
 arbitrary units without difficulty. 
 
 Sensitometers can be of the forms which impress either a 
 time scale or an intensity scale, and these scales can be im¬ 
 pressed either continuously or intermittently. 1 The inter¬ 
 mittent sensitometer varies the exposure on the plate by 
 means of a rotating sector wheel in which angular openings 
 
 1 Jones, L. A., A new non-intermittent sensitometer. J. Frankl. Inst. 189: 303. 1920. 
 
 27 
 
 Fig. 9 
 
 1 Ord. plate 3 Pan. plate 
 
 2 Ord. plate 4 Ortho, plate 
 
MONOGRAPHS ON THE THEORY OP PHOTOGRAPHY 
 
 are cut, permitting the differential transmission of light in 
 some known ratio. This wheel, which is mounted at one end 
 of the sensitometer box, is rotated at a constant speed in 
 front of the plate during exposure. Immediately behind it 
 is an opening in which the plate holder carrying the plate is 
 inserted. The light source is mounted at the other end of the 
 box and at a definite distance from the plate holder. The 
 instrument, comprising rotating sector wheel, light source, and 
 plate holder, is enclosed in a light-tight housing so that the 
 only light incident on the plate during exposure comes from 
 the standardized light source and passes through the opening 
 in the sector wheel as it rotates in front of the plate. Expo¬ 
 sures in this type of sensitometer are incident intermittently, 
 and investigation has shown that the resulting densities are 
 lower in value than those from an equal, continuous exposure, 
 especially for low exposure values. 
 
 In non-intermittent sensitometers, the sensitive material 
 is supported in a frame equipped with a device for screening 
 it in successive steps. The movement of the screen is so 
 timed that consecutive steps receive exposures whose times 
 vary as powers of some convenient number. As in the inter¬ 
 mittent type, this causes the logarithms of the exposures to 
 differ by constant amounts, and they can, therefore, be plotted 
 as abscissae equally spaced. The intensity of the light falling 
 on the plate can be varied by changing the distance between 
 the source and the plate or by altering the source itself. 
 
 A method of intensity-scale sensitometry which varies the 
 exposure by changing the light intensity rather than the time 
 of exposure consists of placing in contact with the emulsion 
 of the photographic plate a tablet divided into several sections 
 of consecutively increasing density. By exposing through 
 this tablet for times comparable with those used in the case 
 to be investigated, the danger of exceeding the reciprocity 
 limits of the exposure formula may be avoided. The tablet 
 also forms a means of obtaining definitely graded exposures 
 in the field, or under circumstances where the elaborate 
 apparatus for the other method can be neither obtained nor 
 manipulated. The chief objection to the tablet method is a 
 possible lack of neutrality of the superposed material. 
 
 If the light intensity is not merely cut down but selectively 
 absorbed, the photographic plate will receive light of a different 
 color from that given by the source, and this difference will 
 change as we proceed from one density to another in the tablet. 
 It is quite possible, however, to form tablets which are neutral 
 within the limits of photographic detection. The tablets 
 
 28 
 
AERIAL HAZE 
 
 used in the field work described in this monograph were made 
 by exposure in the laboratory with the non-intermittent 
 type of sensitometer and developed in monomethyl par- 
 aminophenol (Elon), a developer particularly suitable for 
 producing neutral deposits. The resulting tablets were 
 tested for selectivity of transmission by comparing exposures 
 through them with exposures in the laboratory sensitometer. 
 No effects attributable to color greater than the experimental 
 error could be detected. The transmission of these tablets 
 was then measured both by means of a photometer and photo¬ 
 graphically, and used as a factor to define the amount of light 
 transmitted by each step. 
 
 The density of a photographic plate was defined by Hurter 
 and Driffield as the logarithm of the opacity, the opacity being 
 the reciprocal of the transparency of the plate. Thus, if / 
 be the intensity of the incident light, and /, the intensity of 
 the light transmitted, then 
 
 I\ 
 
 Transparency 
 
 = - 
 
 = T- 
 
 
 / 
 
 
 
 1 
 
 I 
 
 Opacity 
 
 = - 
 
 = -= O: 
 
 
 T 
 
 h 
 
 Density 
 
 = log ,io 
 
 0 : 
 
 
 
 1 
 
 
 - log 10 - 
 
 | 
 
 
 
 T 
 
 
 
 I 
 
 
 - logio - 
 
 - = D. 
 
 /, 
 
 Thus, when D = .301, 0=2, and T = .5 ; D = 1, 0 =10, 
 and T = .1; D = 2,0 = 100, and T = .01; D = 3, 0 = 1000, 
 7’ = .001, and so on. 
 
 It will be seen that if we assume the silver deposit to follow 
 Beers’ law, D will be proportional to the mass of silver in unit 
 area; and it has been shown experimentally that this relation 
 holds with great accuracy for photographic deposits, the 
 relation between D and the mass of silver per square centimeter 
 being termed the photometric constant. For normal emul¬ 
 sions and a density of unity, there corresponds a mass of .1 
 milligram of silver per square centimeter. The measurement 
 
 29 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 of the density can be made with any of the usual devices for 
 comparing light intensities. 1 
 
 Many different types of photometers, microphotometers, 
 and spectrophotometers have been designed and used for the 
 measurement of photographic densities. In these instru¬ 
 ments, numerous means have been used for varying the intens¬ 
 ity of the light, such as adjustable slits, polarizing and analyz¬ 
 ing nicol prisms, and calibrated neutral gray wedges. In most 
 of the instruments one light source is used to illuminate both 
 halves of the photometric field, so that any variation in the 
 intensity of the source during the course of a set of readings 
 does not affect the determinations. 
 
 The essential requirements in any density-measuring 
 instrument are: (1) a photometric field consisting of two 
 parts so arranged that when they are of the same brightness 
 the dividing line almost or entirely disappears; (2) a means of 
 illuminating the two parts of the field uniformly and equally; 
 (3) a means of placing the density to be measured in the path 
 of the light illuminating one part of the field; and (4) a means 
 of cutting down the intensity of the light illuminating the 
 other part of the field to a variable and precisely known 
 extent. 
 
 Certain forms of polarization photometers are quite sensitive, 
 free from systematic errors, and extremely convenient for 
 measuring the coefficients of transmission for white light. _ One 
 of the best and simplest of these and the one used in this inves¬ 
 tigation was the Martens (Fig. 10). Light from the source, a 
 tungsten glow lamp, enters at the two apertures, A and A i, is 
 rendered convergentbyO,and is polarized by a Wollaston prism 
 W. The resulting circular field is divided along the diameter by 
 a line which practically disappears when a balance is obtained. 
 A nicol prism N is interposed in the path of the double beam. 
 When the prism is placed at an angle of 45° to the planes of 
 polarization, equal amounts of light are transmitted. Rotat¬ 
 ing the prism reduces the transmission of one beam and 
 
 1 Sheppard, S. E., and Mees, C. E. K., Investigations on the theory of the photographic 
 process. 
 
 30 
 
AERIAL HAZE 
 
 increases that of the other, the ratio of transmitted intensities 
 varying as the square of the tangent of the angle which the 
 prism makes with the position of complete extinction. The 
 two halves of the field are illuminated uniformly and equally 
 by light passing through a disk of opal glass placed in front 
 of the light source. The photographic plate is placed on the 
 opal glass in such a manner that the density to be measured 
 covers one half of the field, while the other half remains 
 clear. The instrument as a whole is extremely sensitive 
 and reliable. 
 
 The deposit in a photographic emulsion consists of a large 
 number of small particles of silver, and this deposit diffuses or 
 scatters some of the light incident upon it, the amount scat¬ 
 tered depending upon the size of the grains and their distribu¬ 
 tion in the film. The transmission of such a density for a 
 specular beam will therefore consist partly of specular light 
 and partly of scattered light, and since photometers necessarily 
 measure only the light leaving the density over a small angle, 
 the light scattered by the density will be counted as if it were 
 absorbed. As a result of this, a density measured by light 
 which is approximately parallel will give a higher reading in 
 the photometer than the same density measured by light 
 which is completely diffused before it enters it, since in the 
 first case light will be lost by scattering, whereas a completely 
 diffuse light cannot further be fully scattered. This subject 
 was investigated by Callier, 1 who termed the relation between 
 the density measured by parallel light and that measured by 
 scattered light Q, and showed that the value of Q varies from 
 slightly above unity for the finest grained and most transparent 
 deposit to a value as high as 1.6 for the coarsest and most 
 scattering deposits. Callier showed that the most practical 
 and consistent method of measuring densities is to use com¬ 
 pletely diffuse illumination. This is accomplished by making 
 the measurements with the emulsion side of the photographic 
 plate in contact with a uniformly illuminated piece of pot 
 opal glass. By doing this, the results obtained with different 
 photometers are comparable, and the conditions closely 
 approximate those which exist when contact prints are made 
 from negatives. 
 
 If the area of the density to be measured is too small to 
 cover the photometer aperture, it is necessary to project a 
 magnified image of it on to a screen placed in direct contact 
 with the aperture. A microphotometer of this type was used 
 in part of the work. 
 
 33-'2bt) Iier i9oV Absorption an ^ scatter °f fight by photographic negatives. Phot. J. 
 
 31 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 COLOR SENSITIVE MATERIALS 
 
 Since the objects of this investigation included a study of 
 the color of atmospheric haze and the possibility of its elim¬ 
 ination by the use of suitable light filters, the materials used 
 were necessarily color sensitive; that is, sensitive to light of 
 longer wave-length than 500/^, which is the limiting wave¬ 
 length of sensitiveness for normal bromo-iodide emulsions. 
 Such color sensitiveness is produced by the use of sensitizing 
 dyes which, when added to the emulsion, stain the siher 
 bromide grains and render them sensitive to light of the wave¬ 
 lengths which are absorbed by the dye used. 
 
 Two groups of dyes are used in color sensitizing, and conse¬ 
 quently color sensitive materials fall in one of two general 
 classes: (1) Dyes of the fluoresceine series, notably the 
 sodium salt of tetra-iodo-fluoresceine, generally known as 
 erythrosine, sensitize silver bromide very strongly for the 
 yellow-green rays which they absorb, and are generally used 
 for the orthochromatic plates of commerce. These dyes are 
 acid, being used in the form of the soluble sodium salts, and 
 their action upon the emulsion is enhanced very much by the 
 presence of ammonia. (2) On the other hand, there is a 
 series of basic dyes produced by the condensation of quinoline 
 with quinaldine derivatives which are also powerful sensi¬ 
 tizers for photographic emulsions and which confer a wide 
 band of sensitiveness stretching through the whole of the green 
 and far into the red regions of the spectrum. Photographic 
 materials, sensitive to practically the whole of the visible 
 spectrum, are known as panchromatic. They may be made 
 either by the addition of the dyes to the emulsion or by treat¬ 
 ment of the dry plates with a solution of the dye. "1 he cyanine 
 dyes have been known for many years to sensitize photo¬ 
 graphic plates, but they have the disadvantage that the 
 materials sensitized with them fog easily and have very poor 
 keeping qualities. In 1897 Miethe discovered a dye which 
 shows a strong sensitizing action on silver bromide without 
 the disadvantages displayed by the cyanine dyes. This dye, 
 Ethyl Red, was the first representative of a new series of 
 dyes known as isocyanines, which are isomeric with the cya¬ 
 nines, and are produced by condensing together ethiodides of 
 quinoline and quinaldine derivatives. 
 
 The investigation of these dyes was extended by E. Koenig, 
 who supplemented Miethe s first dye, known as Ethyl Red, 
 by other dyes made from substituted quinaldines and quin¬ 
 olines, which were placed on the market under the trade names 
 
 32 
 
AERIAL HAZE 
 
 of Orthochrome T, Pinaverdol, and Pinachrome. Later, 
 Koenig discovered that by condensing the dye components 
 in the presence of formaldehyde, a new series of dyes was 
 obtained, the first of which was placed on the market under 
 the name of Pinacyanol, this dye displaying extraordinary red 
 sensitizing power and showing a strong maximum of sensitive¬ 
 ness at 645 m/*. This new series of dyes has recently been 
 termed the carbocyanines. Still other dyes of a similar type 
 but belonging to a different series, have sensitizing power in 
 the extreme red and even in the infra-red. 1 For practical 
 purposes the most useful dyes are Pinacyanol, Kryptocyanine 
 and one or two of the isocyanines such as Orthochrome T, 
 Pinaverdol, or Pinachrome, the particular isocyanine to be 
 selected depending upon the characteristics of the emulsion. 
 These dyes, which were made only in Germany before 1914, 
 were duplicated in the allied countries during the war. 
 
 Since exposures from the air are necessarily very brief, and 
 since, as will be shown in the course of this monograph, the 
 use of strongly colored filters is advantageous, every effort 
 was made to increase the color sensitiveness of the materials 
 available, and the sensitizing of materials and methods of 
 measuring their color sensitiveness were extensively studied. 
 Numerous improvements were made in the preparation of the 
 materials, notably in panchromatic plates and films sensitized 
 with pinacyanol and the isocyanine dyes. Small quantities 
 of highly sensitized materials were made by bathing finished 
 plates in dye baths containing a large amount of alcohol and 
 some ammonia, but equally good results were found to be 
 obtainable by the use of materials sensitized in the emulsion 
 and hypersensitized by treatment with a dilute bath of 
 ammonia before use. 2 At the present time, plates and films 
 of sensitiveness sufficient for aerial photography when used 
 with strong yellow filters are commercially obtainable. If 
 red filters are to be used, hypersensitizing with ammonia gives 
 excellent results. 
 
 The measurement of the color sensitiveness of the materials 
 can be accomplished either sensitometrically or by means of a 
 spectrograph. A simple grating spectrograph using tungsten 
 
 1 Adams, E. Q., and Haller, H. L.. Kryptocyanines. A new series of photosensitizing 
 
 dyes. J. Amer. Chem. Soc. 42: 2661. 1920. Mees, C. E. K., and Gutekunst, G. O., 
 
 Some new sensitizers for the deep red. J. Ind. Eng. Chem. 14: 1060. 1922. 
 
 The Kryptocyanines were not discovered until after the work described here was com¬ 
 pleted. Kryptocyanine sensitizes strongly with a maximum at 770p,pt. and should be very 
 suitable for aerial photography, but owing to the falling off in the energy of the solar 
 spectrum beyond 700 j(,[j, and to the strong absorption bands between 760JJ,^JL and 780|j,JJl it 
 is very difficult to get sufficient exposure. 
 
 2 Burka, S. M., Hypersensitizing commercial panchromatic plates. J. Frankl. Inst. 
 
 189: 25. 1920. 
 
 33 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 light as the source, and a wave-length scale held in front of the 
 sensitive plate so that it is impressed upon the plate at the 
 time of photographing a spectrum are all that are required, 
 but it is convenient to have a neutral-tinted wedge in front of 
 the slit by means of which a curve of the sensitiveness of the 
 material is drawn automatically so that the position of the 
 sensitive bands can be seen at a glance (Fig. 11). The spectro¬ 
 graph used for this purpose is known as a wedge spectrograph A 
 
 Fig. 11 
 
 For quantitative measurements, however, a more satis¬ 
 factory method of determining sensitiveness is to use a time 
 or intensity scale sensitometer and to make exposures through 
 color filters transmitting known regions of the spectrum. A 
 convenient set of filters is that known as the standard tricolor 
 set of filters used for color photography, the transmission 
 limits of these filters being as follows: 
 
 A. 580 to infra-red; 
 
 B. 475 /a/ 4 to 635 
 
 C. 350 to 510 /jl/x. 
 
 With such filters in the sensitometer, the material is exposed 
 in four strips, one without a filter and one each with the three 
 color filters, the increase of exposure necessary to get the 
 same density through each filter being measured. 
 
 Since the characteristic curves are not parallel to each other 
 (Fig. 12), it is necessary to decide on the density and develop¬ 
 ment at which the comparison is to be made, and for the sake 
 of precision the plate is developed so that the curve of the 
 plate without a filter shows a y of unity, and the exposure 
 necessary to produce a density of unity behind each filter is 
 then used for computing the increase of exposure required by 
 the filter. These increases are known as the filter factors. 
 As a material is made more color sensitive, the factor for the 
 blue filter naturally increases, and the factors for the red and 
 green filters diminish. The filter factors for the tricolor 
 filters form what is generally known as the filter ratio. For 
 
 1 Wratten, S. H., and Mees, C. E. K., The wedge spectrograph. Brit. J. Phot. 54: 384. 
 1907. 
 
 34 
 
AERIAL HAZE 
 
 the standard Wratten tricolor filters A, B and C, the ratios 
 of plates made before the war were generally 
 Blue, 6; Green, 12; Red, 16. 
 
 These ratios were improved by intensive work until for an 
 ordinary plate a ratio of 
 
 10 ; 10 ; 10 
 
 may be considered normal, and by hypersensitizing a ratio of 
 
 12 ; 8 ; 6 
 
 may easily be obtained. However, if this resulted in a general 
 loss of sensitiveness, it would present no advantage for aerial 
 photography. A more useful record of color sensitiveness can 
 
 Fig. 12. Characteristic curves for different filters 
 
 be obtained by stating the effective speed through the filters, 
 this being found by taking the total H. and D. speed to white 
 light and dividing it by the factor of the filter concerned. 
 Thus, if we have a plate of speed 240, and it is to be used with 
 a No. 12 Wratten filter having a factor of 6, the effective speed 
 will be 40, while with the red filter having a factor of 10, the 
 effective speed will be 24. It may be quite possible to make a 
 similar plate with a factor for the red filter of only 6, but with 
 a speed reduced to 120 by the excessive amount of dye used. 
 It will be seen that the effective speed of this through a red 
 filter will be only 20 as compared with 24 for the first plate, 
 so that the additional dye would be a disadvantage for aerial 
 photography, although the ratio might at first sight suggest 
 that it would be advantageous. 
 
 35 
 
CHAPTER III 
 
 The Measurement of Aerial Haze 
 
 Haze is the optical turbidity of the atmosphere. It may 
 be caused by dust, smoke, water vapor, irregular temperature 
 distribution, or even by dry air. In fact, any atmospheric 
 material which tends to diminish the transparency of the space 
 it occupies may be said to be a source of haze. There are 
 many types of haze. At times it is stratified near the ground 
 or at high altitudes, and at other times it extends uniformly 
 to a great height. Its quality or color also varies, not only 
 with the varying sources of haze but also with the size of the 
 suspended particles it may contain. Haze, therefore, is due 
 to a light-scattering medium, which creates a veiling glare 
 between the camera and the subject to be photographed. It 
 superposes a uniform light intensity over all parts of the 
 subject, causing not only a brightening of it but also a great 
 reduction in its contrast. 
 
 The determination of the distribution, the quantity, and 
 the quality of haze is possible only by observations from aero¬ 
 planes at various altitudes and under different weather 
 conditions, and, since these important factors must be consid¬ 
 ered from a photographic viewpoint, photographic methods 
 are employed. Therefore, it is necessary to frame a definition 
 of haze in terms of its measurable effects upon the developed 
 photographic material. This definition is based upon the two 
 obvious effects. The suspended materials in the atmosphere 
 scatter sunlight and hence send back light to the camera, 
 which adds to the exposure creating an image of the subject: 
 these materials also subtract somewhat from the light reflected 
 upward from the ground. Although these tendencies work 
 in opposite directions in their effects upon the exposure which 
 the photographic plate receives, there is no reason for assuming 
 that they exactly counteract one another. The absorption 
 or subtractive effect of the haze is concerned with much less 
 light and is no doubt of less importance. 
 
 Assume that the subject to be photographed contains objects 
 which are white, gray, and black. Let E w , E g and E< 0 , 
 respectively, represent exposure values, in candle-meter- 
 seconds if convenient, due to these objects. When the camera 
 is near the ground, no haze effect except,of course, the general 
 decrease in ground illumination, is involved. At any altitude 
 A let the exposure due to the light from the haze be e and let h' 
 
 36 
 
AERIAL HAZE 
 
 be the ratio of e to E w so that 
 
 e = 
 
 ti 
 
 Therefore, the total exposure giving rise to the image of the 
 white portion of the subject will be 
 
 E w (1 —a) + h' E w 
 
 where (1 - a) represents the absorption or subtractive effect 
 of the haze. Similarly, for the black portion of the subject 
 the total exposure is Eb(l — a) + h'E w and for the gray 
 Eg -a) + h'E w . This is true since the relative amounts 
 subtracted and the absolute exposure added are identical 
 for the white, gray, and black portions of the subject. There¬ 
 fore, if on a photographic material, exposed at any altitude, 
 the ratio of exposures on the white and black objects respec¬ 
 tively is K, then 
 
 E w (1 a) -j- h' E w _ „ 
 
 E b (1-a) + h r lE w ~ ' 
 
 or, dividing both numerator and denominator by Ew( 1 -a), 
 
 h' 
 
 t + (l-o ) K 
 3e_\ _ J L 
 
 E w (1—a) 
 
 h' ^7 
 
 Now let ~t~ -; = h, and further let = C, which is 
 
 ( — a) E b 
 
 measured from near the ground, say at an altitude of 500 feet 
 
 Solving for h 
 
 Then 
 
 1 + h 
 T + h 
 
 = K. 
 
 C - K 
 C(K- 1) 
 
 = h. 
 
 h is the haze effect expressed in such terms as are readily 
 obtainable by photographic methods. The evaluation of C 
 and K from plates exposed from various altitudes will be con- 
 
 37 
 
MONOGRAPHS ON THE THEORY OE PHOTOGRAPHY 
 
 sidered later. However, it may be remarked here that the 
 
 Fig.13 
 
 Test objects are shown at lower right of picture 
 
 h' E 
 
 fraction -— 7 --is the ratio of the exposure due to the 
 
 E w (1 — a) 
 
 38 
 
AERIAL HAZE 
 
 haze to the exposure given through the haze by the white 
 portion of the subject; and this ratio is obviously equal to h, 
 the haze effect. It is also evident that this evaluation of haze 
 is independent of the actual brightness values—a fact which 
 will be discussed later. 
 
 The three factors which it was desired to determine are the 
 distribution, the quantity, and the quality of the haze. The 
 method employed permitted the determination of all these 
 from the same negative. The observations upon which the 
 value of h and its distribution and color are based were made 
 at Langley Field, Hampton, Va., during December, 1918, and 
 January, 1919. Preliminary measurements, which aided 
 materially in the establishment of the method finally adopted, 
 were made from negatives obtained during the summer and 
 fall of 1918 at Baker Field, Rochester, N. Y. 
 
 EXPERIMENTAL METHODS 
 
 The method consisted primarily in photographing three 
 test objects a black, a white, and a gray canvas, each 60 
 
 X 60 feet, of known reflecting 
 power, spread in order upon 
 level ground. (Fig. 13.) These 
 canvases were photographically 
 non-selective in reflection—a 
 necessary condition since the 
 quality of haze was one of the 
 factors to be determined. The 
 contrast between the white and 
 the black canvas was approxi¬ 
 mately 1 to 8. The reflecting 
 power of the three canvases was 
 measured both visually and pho¬ 
 tographically, with the results 
 given in the following table: 
 
 TABLE I 
 
 Canvas Reflecting Power 
 
 Photographic Visual 
 
 White. 56.0 64 8 
 
 Gray. 17.5 20.1 
 
 Black. 6.7 75 
 
 The cameras used were a four-lens type especially designed 
 for this work. (Fig. 14.) Four IC Tessar lenses, each of 
 ten-inch focal length, are placed at equidistant intervals in 
 
 39 
 
MONOGRAPHS ON THE THEORY OE PHOTOGRAPHY 
 
 the lens board. (Fig. 15.) The plate holder (Fig.fjjl6), 
 carries four 4X5 inch plates and is fitted with a notched 
 template to make it possible to determine the position of a 
 plate in the camera. Provision is made in each lens barrel 
 for the insertion of color filters. _ The camera thus served 
 as a photographic spectrometer, since the filters chosen were 
 of such cut as to divide the spectrum into sharp intervals of 
 known area. 
 
 Fig. 15 FlG - 16 
 
 Lens board Plate holder 
 
 In order to make the exposures received by the four plates 
 as nearly equal as possible, the camera was calibrated. The 
 stop or diaphragm values of the lenses, the speed of the shut¬ 
 ter, and the relative total transmissions of the filters used 
 were considered. The times of exposure for the four different 
 slit widths and the six tension values of the focal plane shutter 
 were determined by means of a shutter tester, which depended 
 upon the operation of a tuning fork of known frequency. The 
 values of exposure times obtained were plotted as functions 
 of the distances moved by the slit, and the time value which 
 corresponded to the value at the center of the plates was 
 selected for use. The filter factors, which differed foi the 
 various types of photographic materials used, were defined as 
 the ratio of exposures through the color filter and through a 
 plain gelatine or dummy filter when both exposures produced 
 a density of unity, the plates having been developed together 
 for a time that gave a y value of unity on the plate exposed 
 through the dummy filter. These factors were obtained by 
 means of a non-intermittent sensitometer, using identical 
 
 40 
 
AERIAL HAZE 
 
 exposures on the two plates by adjustment of the distance 
 from the source of light to the plate and the aperture over the 
 source; so that when these factors of distance and aperture 
 were considered, the two characteristic curves intersected at a 
 density of unity when the y of the plate exposed through the 
 dummy filter was also unity. In this way it was determined, 
 for example, that the filter factor for the Wratten and Wain- 
 wright No. 12 filter is 6.4 for a particular panchromatic plate, 
 while for an orthochromatic plate the factor for the same 
 filter is 8.5. Suppose, now, that the motion of the focal plane 
 shutter is from lens A to lens B, and that A has a blue filter, 
 B a dummy filter; and, further, that it has been found that 
 for the desired slit and tension, the ratio of exposure times of 
 A to that of B is as 3 to 2. The plate to be used is ortho¬ 
 chromatic. Because of the motion of the aeroplane during 
 the taking of the pictures, the shortest adequate time is to 
 be used, and the stop of lens A is at //4.5. The exposure at B 
 must be 1/8.5 of that at A. Thus, 2/3 X 1/A = 1/8.5, and X 
 will represent the fractional reduction to be produced by the 
 stop. Therefore, the stop of lens B is set at//10.7. With 
 the stop values thus set for each filter to be used, the adjust¬ 
 ment was checked by photographing a non-selective white 
 diffusing surface in sunlight. Equal densities under the condi¬ 
 tions of calibration as regards development were thus produced. 
 The latitude of the material ordinarily covers small departures 
 from precise equality in the resulting exposures. 
 
 The establishment of a numerical relation between density 
 and exposure by interpolation on a scale of densities produced 
 by known exposures was provided for by an intensity scale 
 adaptable to field use. In this case, the intensity scale method 
 consisted of printing a series of areas of known transmission 
 and selectivity on the material used to photograph the test 
 objects. One of the lenses was replaced by a glass plate 
 supplied by the manufacturer of the lens and intended to 
 duplicate it in absorption and reflection. In the corresponding 
 section of the plate holder a 4 X 5 inch tablet was placed, 
 having five sensitometric strips of ten sections each, of which 
 the transmissions were known. Except for these strips, the 
 tablet was opaque. Color screens were placed between each 
 strip and the glass cover plate, the heaviest filter corresponding 
 in position to the slowest shutter speed. The photographic 
 material placed over the tablet could receive light only through 
 the filters and accompanying sections. The exposure was 
 made over the white canvas. The filters in the tablet were 
 the same as those used in the three other lenses of the camera. 
 
 41 
 
MONOGRAPHS ON THE THEORY OE PHOTOGRAPHY 
 
 In the construction of this tablet it was desirable to have 
 the log transmissions of the successive steps in any strip in 
 arithmetical progression. These log transmissions differ by a 
 constant from the log exposure given through them to the 
 photographic plate, and when the characteristic curves for 
 the observations are to be plotted, it is undesirable that the 
 log exposure values should be unevenly grouped along the 
 axis. In order that these log transmissions of successive 
 steps on the strip should be uniformly separated, the corre¬ 
 sponding exposures were calculated, as follows: In the coordin¬ 
 ate system given in Fig. 17 the negative X axis represents the 
 values of log transmissions of the tablet uniformly spaced. 
 Since density equals log opacity 
 
 (log 
 
 -:— 7 — or — log transmission), 
 
 transmission 
 
 the same values and spaces transferred to the positive 
 Y axis represent densities of the areas of the tablet. 
 In the first quadrant is given the characteristic curve of the 
 
 plate and the time of development used, which plate is to be 
 used to make the tablet. Now, if the equally spaced points 
 along the positive Y axis are transferred to the positive X 
 
 42 
 
AERIAL HAZE 
 
 axis, the new points will represent the various values of log 
 exposure to be given in making the tablet. These points are 
 again transferred to the negative Y axis by means of a logarith¬ 
 mic curve and the ordinates so determined are the exposure 
 values which should be given the plate of which the char¬ 
 acteristic curve is shown in the first quadrant, in order that 
 the tablet have the transmissions desired. The determination 
 of these values gives the exposure times to be used in the 
 preparation of the tablets in the non-intermittent sensitometer. 
 It appears, for example, that the first, fourth, seventh, and 
 tenth steps are to be produced by exposures proportional to 
 0.16, 0.31, 0.58, and 1.2 respectively, or equal to these num¬ 
 bers in the units represented along the positive X axis. Many 
 tablets were made in this way, with various highest density 
 values. This was to compensate for the low total transmission 
 values of the filters to be placed over the tablets. 
 
 For example, in order to produce a good H. and D. curve by 
 the same time of exposure through the tablets, over which 
 different color screens were used, it was necessary to place 
 under the heaviest filter a strip of low densities, and under the 
 lighter filters strips of higher densities. The tablets were 
 developed in an Elon developer so as to produce a deposit of 
 as little selectivity as possible. These tablets formed the 
 intensity scale sensitometers used to interpret the air data. 
 Exposures were made through the tablets (over which suitable 
 color screens were placed as described), by light of the same 
 quality as that by which the photograph was taken. Thus 
 the density of any image in the aerial photograph is readily 
 referable to the characteristic curve of the plate on which it 
 is taken, and the ratio of exposure values derived. From 
 these the value of the haze effect, h, is determined. 
 
 In this method of field sensitometry, then, it is possible to 
 make exposures from a plane in the air at the time when the 
 pictures of the test objects are taken. This could be done in 
 the fourth section of the four-lens camera, the three remaining 
 sections giving photographs of the test objects. This, how¬ 
 ever, makes the time of exposure for the sensitometric strip 
 and for the image plates equal. As the altitude increases, the 
 light through the plates or tablets comes from a wider and 
 wider extent of territory. The region included at Langley Field 
 would be at first only grassy fields, then some buildings, some 
 woods, and finally areas of water would be included. Because 
 of this fact the quality of the light would vary with altitude, 
 and inasmuch as such variation directly influences the relation 
 between density and exposure, it is to be avoided. Even 
 
 43 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 in the method finally adopted, where the exposures were made 
 directly over the white canvas test object, the camera being 
 held by an observer on the ground, some discrepancy is pos¬ 
 sible. The light by which the picture is taken from an altitude 
 of 10,000 feet, for instance, might not be the same as that near 
 the ground even when the picture and the sensitometric strip 
 are made at the same time. To determine the effect of this 
 difference upon the sensibility of the plate—that is, in technical 
 terms, the effect of altitude on y — flights over a uniform 
 background, e. g., water, were made. 
 
 Laboratory methods usually endeavor to duplicate in the 
 sensitometer the illumination out-of-doors on an average day, 
 i. e., a combination of sky and sunlight such as one would 
 receive by reflection from a perfectly white subject in natural 
 surroundings. It is readily seen that the aerial photographer 
 must discriminate between such sensitometric data and those 
 which would be obtained from this illumination after it had 
 been reflected from the various subjects encountered in the 
 average terrain. Both contrast and color factors may differ 
 widely for light reflected from water and from a plowed field. 
 The following data bring out clearly the inadvisability of 
 obtaining sensitometric data by exposing the strip in the four- 
 lens camera simultaneously with the taking of the photographs 
 of the test objects. The plates, the data from which are given 
 in the following table, were exposed between a sensitometric 
 tablet whose five strips were masked with five different color 
 filters in a camera in which provision was made for distributing 
 the light homogenously over the plate. Exposures were 
 made at low altitudes in order to restrict the camera angle to 
 a given type of subject. 
 
 TABLE II 
 
 VALUES OF GAMMA 
 
 Emulsion No. 2270 (Special Red Sensitive) 
 
 Filter 
 
 Subject Dummy Aero No. 1 No. 12 No. 21 No. 25 
 
 Water. 1.36 1.76 2.13 1.92 2.02 
 
 Woods. 1.67 1.82 2.13 2.00 2.14 
 
 Hampton City. 1.72 2.02 2.16 2.20 2.14 
 
 Light fields. 1.80 1.91 2.13 2.14 2.24 
 
 Emulsion No. 1445 (Special Red Sensitive) 
 
 Filter 
 
 Subject Dummy Aero No. 1 No. 12 No. 21 
 
 Marsh. 2.24 2.14 2.14 1.80 
 
 Light fields. 2.22 2.10 1.94 1.74 
 
 Hampton City. 2.18 2.38 2.23 1.77 
 
 44 
 
AERIAL HAZE 
 
 It will be noticed on examining the tables horizontally that 
 these two emulsions behave very differently with change of 
 color. No. 2270 shows an increase of y with shift of color 
 to red, whereas No. 1445 acts in the opposite manner. Water 
 (Back River) was evidently the bluest of the subjects. For 
 emulsion No. 2270, water produced the lowest y and bare 
 light fields the highest. An equally interesting table of results 
 is formed from the filter factors of the above filters in the same 
 cases. The filter factor is defined (as previously), as the ratio 
 of the exposure necessary to produce a given density through 
 the filter in question as compared with that through the 
 dummy when developed to the same standard value of y, 
 usually unity. In the present case, in which the strips exposed 
 through the different filters were on the same plate, they 
 were necessarily given the same treatment, with consequent 
 variation in y. The ratio of exposures was, therefore, taken 
 by extending the straight-line portions of the H. and D. curves 
 until they intersected the log E axis, and reading the differences 
 between these intercepts. We have noted in Chapter II that 
 this intercept is constant for various treatments except for 
 materials showing regression of inertia. Moreover, this effect 
 is slight for the changes involved. Within this effect TableIII 
 gives true filter factors. 
 
 TABLE III 
 
 FILTER FACTORS 
 
 Emulsion No. 2270 
 
 Filter 
 
 Subject Aero No. 1 No. 12 No. 21 No. 25 
 
 Water. 3.98 7.59 9.55 24.55 
 
 Woods. 2.95 6.61 9.33 18.63 
 
 Hampton City. 2.98 5.68 8.23 11.76 
 
 Light fields. 2.41 4.46 6.47 12.60 
 
 Emulsion No. 1445 
 
 Marsh. 2.51 4.57 6.17 
 
 Light fields. 2.95 4.62 6.98 
 
 Hampton City. 4.02 6.77 9.55 
 
 The interest in this table is in the change of factor with 
 subject. The change with filter is largely a total transmission 
 effect and is important in determining exposure times in 
 practice with the particular filter and emulsion. It should 
 be noted that the factors vary with the subjects in the same 
 order as the values of y. Water gives strikingly high values. 
 The second emulsion confirms the order of light fields and the 
 city of Hampton, and places the maximum of light from 
 marshes farther toward the red. The marsh was covered 
 
 45 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 with dead red-brown foliage. The vegetation of woods and 
 city was largely evergreen. 
 
 It is seen, therefore, that it would be unwise to apply to 
 the whole problem the factors and y values which were 
 characteristic of the light from the white test object on the 
 ground without first determining whether or not the H. and 
 D. curve on the plates to be used suffered a measurable change 
 as the altitude increased. To test this, it was necessary to 
 expose through the tablet over a uniform subject large enough 
 to insure no change in character of the light at higher altitudes. 
 Two such subjects were accessible, the water of Chesapeake 
 Bay and that of Hampton Roads. Many flights were made, 
 and although a difference in the color of the two subjects was 
 easily detected, no change with altitude was measurable. 
 A typical flight over Hampton Roads gave the following data: 
 
 TABLE IV 
 
 Altitude 3,500 4,500 5,500 7,000 8,000 9,000 10,000 11,000 12,000 feet 
 y 89 .92 .88 .98 .86 .84 .88 .98 .93 
 
 On another flight exposures were made over both Hampton 
 Roads and Chesapeake Bay. The values of y show the same 
 slight but irregular variation with altitude as those in the 
 above table. They averaged for Hampton Roads 1.40, for 
 Chesapeake Bay, 1.11. The difference in color of the two 
 subjects was easily visible. These results made it possible to 
 apply to all the images of the test objects made in a single 
 flight, the curves obtained simultaneously on the ground by 
 exposing through the tablet to light, reflected from the same 
 test object. Care was always taken in development. One or 
 two plates with strips were treated in the same tray with 
 three or four of the corresponding plates from the air, and 
 marked accordingly, so that the interpretation of the negatives 
 made in the aeroplane depended upon sensitometric strips of 
 which the treatment had in all respects been the same. When¬ 
 ever possible, plates exposed through a single filter were 
 developed together, as this made it unnecessary to measure 
 more than one strip of the five on the tablet. 
 
 METHODS OF MEASUREMENT 
 
 It is obvious, since the altitude of exposures varied from 
 less than 500 to more than 12,000 feet, that the images on the 
 plates must be of various sizes; indeed, they are squares 
 measuring from 1 to 25 millimeters on each side. In the 
 measurement of density, the large squares required only the 
 
 46 
 
AERIAL HAZE 
 
 ordinary arrangement, with the plate to be measured placed 
 under the opening of a Martens photometer. For the measure¬ 
 ment of the smaller images, a projection apparatus was arranged 
 as shown in Fig. 18. The end of the Martens photometer is 
 covered by a paper diffusing screen, S 2 , in front of which a 
 black hood H is placed. The plate to be measured, P, is 
 illuminated by light from the bulb B, through a watercell W, 
 
 Fig. 18 
 
 Diagram of micro-photometer ' 
 
 and a diffusing screen S u and L is the projection lens. The 
 emulsion is scraped from the plate up to a sharp line defining 
 the image of the canvas, and an enlarged image of that part of 
 the plate near the boundary is projected upon S 2 so that the 
 clear glass covers one of the two openings of the photometer 
 head, and the image covers the other. 
 
 The sensitometric strip corresponding to the negative was 
 always measured in the same way as the images on the nega¬ 
 tive.. Therefore, although the two methods did not measure 
 precisely the same quantity, as the projection method gave a 
 value for the density corresponding more nearly to that 
 obtained by measurement by parallel light, there is little or 
 no difference in the tabulated results of exposure ratios. In 
 certain cases where it was possible to measure a plate in both 
 ways, the two values of K obtained were in satisfactory agree¬ 
 ment. Because of inequalities in the test objects, which pro¬ 
 duced inequalities in the density in some of the larger images, 
 it was necessary to make several settings in different parts of 
 the same square. This made the results comparable with 
 those from the negatives made at higher altitudes, where the 
 method naturally exerted an integrating effect. To determine 
 the exposure ratio which in the formula for h was denoted 
 as K, it was necessary to determine the photometer angle 
 readings corresponding to the several sections of the sensi¬ 
 tometric strip and to plot these angles against the log exposures 
 ■—that is, a constant plus the log transmission of the tablet 
 through which the strip was exposed. The constant involved 
 
 47 
 
MONOGRAPHS ON THE THEORY OP PHOTOGRAPHY 
 
 need not be known. The angles corresponding to the images 
 of the canvas are then ascertained by the same method. By 
 applying their values to the curve given by the strip, the 
 appropriate difference in log exposure (which is the same as 
 that between two values of log transmission) may at once be 
 read. The antilog of this number is K, the ratio of exposures. 
 It is usual in sensitometry to plot density rather than photo¬ 
 meter angle as a function of log exposure, but here where the 
 curves serve only as an intermediate means of interpretation, 
 converting the angles to densities has no advantage. 
 
 REPRESENTATION OF RESULTS 
 
 The calculation of the photographic measurement of haze 
 called h requires first a value for K at altitude zero, which has 
 been denoted by C, the ground contrast. This represents the 
 contrast presented by the test objects taken at 500 feet or less, 
 where it was assumed that no haze was present. The value 
 of C was found by plotting the values of K obtained from the 
 data of any one flight as a function of h, thus getting the con¬ 
 trast-haze curve, and the most appropriate regular curve was 
 extended to meet the axis of ordinates, where the intercept 
 was the value of ground contrast C. It was observed that all 
 the flights made could be divided into three distinct groups so 
 arranged with respect to their dates and in relation to the 
 weather conditions that the ground contrast C, which is the 
 ratio of the brightness of the two canvases, remains unchanged 
 for any one group of flights. The several values thus obtained 
 for the flights of this group were weighted on the basis of the 
 regularity or reliability of the corresponding curves and the 
 “weighted mean” value used for C. In some cases, of course, 
 the exposure of the plates was too great, and the density of 
 the images of the white canvas was not represented by a 
 corresponding density on the sensitometric curve of that 
 plate. Under such circumstances, the curve gave a value 
 for the ratio due to the black and the gray test objects, and 
 this could be converted into K if multiplied by the ground 
 contrast existing between the white and the gray test objects, 
 which in turn could be obtained from another set of plates. 
 
 The calculation of h by means of the equation 
 
 C - K 
 C(K- 1) 
 
 gave the values which appear on the haze-altitude curves. 
 The arrangement of the points allowed considerable latitude 
 
 48 
 
AERIAL HAZE 
 
 as to the shape of curve best suited to represent the facts. 
 Three general forms of curves, however, appear more or less 
 adequate, and a straight-line, a logarithmic, and a parabolic 
 curve were adjusted to the observations by a least-square 
 solution. The parabolic curve represents a more uniform 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 FIELD 
 
 DATA 
 
 
 
 
 
 
 
 
 
 
 ^<JAN 14 
 
 1919 AM 
 
 
 
 
 
 
 
 
 JA 
 
 N 21 1919 
 
 PM^^ 
 
 /JAN 27 
 
 1919 PM 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ALTITUDE 
 
 Fig.19 
 
 Relation between haze and altitude 
 
 distribution of haze-forming material with altitude than the 
 other type. The logarithmic type, being convex upward, 
 represents a distribution in which the density of the haze 
 diminishes with increasing altitude. The straight-line relation 
 presents a case where haze is more or less proportional to 
 altitude. These types of curves are shown in Fig. 19. 
 
 The variation of haze values with altitude under various 
 weather conditions having been determined, the data with 
 respect to the spectral quality of the haze remain to be inter¬ 
 preted. According to the terms of the definition of A, 
 E w (1 -a) + h'E w is the exposure at any altitude. Of this 
 amount, the first part E w (1 -a) is the exposure due to the 
 canvas as influenced by haze. The second part h r E w is the 
 exposure due to the haze itself. The ratio of these two com¬ 
 ponents is obviously h '/(1 -a), or h. It may be considered, 
 then, that at a certain altitude the haze sends through a given 
 
 49 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 filter an amount of light which is a fraction h of the amount 
 sent by the canvas through the haze and that filter. The 
 canvas is assumed to be non-selective or neutral and, therefore, 
 the light reflected from it has approximately the same spectral 
 distribution as sunlight. The absorbing effect of the haze, 
 however, is undoubtedly selective; and, therefore, E w (1 -ti) 
 departs from the quality of sunlight. 
 
 It would be of interest to know at least the probable value of 
 a, the absorption effect of haze. The ratio of E w (1 -ti) + 
 ti E w to E w is the ratio of the two exposures at high and at 
 zero altitudes respectively. That ratio is slightly greater than 
 
 Fig. 20 
 
 A—Distribution of sunlight C\ Possible distribution of 
 
 B—Distribution of light from Dj light by haze 
 
 neutral canvas 
 
 unity, and under conditions where the value of h equals 0.3, 
 it is possible with a fair degree of accuracy to set (1 -a) + ti 
 equal to 1.1. This value is undoubtedly less than 1.2. Taking 
 (1 _ a ) + h' as 1.1 and ti/{\ -ti) as 0.3, we get a = approxi¬ 
 mately 0.15 and ti approximately 0.25. If more data on 
 the ratio of high and low exposures were at hand, it would 
 probably be easy to separate these two values more definitely; 
 nevertheless, it appears that E w ( 1 ti) has a value of ap¬ 
 proximately 85 per cent of E w . In the diagram in Fig.. 20, 
 A represents the distribution of sunlight, B that of the light 
 
 50 
 
AERIAL HAZE 
 
 from the neutral canvas, and the curves C and D possible 
 representations of E w (l —a), the areas under C and D being 
 15 per cent less than that under B. The data lead to the belief 
 that the distribution indicated by curve D is more probable 
 than that given by curve C. 
 
 As a first approximation and in order to exhaust the informa¬ 
 tion available in the data, it is assumed that the light sent by 
 the canvas through the haze is of the quality of sunlight. 
 Then in any interval of the spectrum the illumination due to 
 the canvas is proportional to the areas included between 
 corresponding ordinates under the distribution curve of white 
 light. The proportionality factor depends upon the reflecting 
 power of the canvas. If this area be multiplied by a value of 
 h derived from data obtained through a filter transmitting the 
 spectral region concerned, the product represents, by the same 
 proportion, the light from the haze. For two filters whose 
 transmission bands are widely different, these two products 
 
 will, in general, differ, since neither h nor the area multiplied 
 are the same. Such a difference might be attributed to a 
 haze effect in the region not common to the two filters. There¬ 
 fore the difference of the products represents an area whose 
 base on the wave-length axis is the spectral region transmitted 
 by one filter and not by the other. The altitude is obviously 
 the area divided by the base. These rectangles form the 
 
 51 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 basis of the /z-wave-length curves (Fig. 21) which are so drawn 
 that the areas under the curve and between the sides of any 
 one of the rectangles are the same as that of the rectangle 
 itself. In calculating these areas the values of h were taken 
 from the haze-altitude curves for 10,000 feet. The trans¬ 
 mission bands of the several filters extended variously from 
 2800 to 7000 A. U. Evidently, the narrower and more 
 numerous the rectangles in the spectral region involved, the 
 more definitely the haze-wave-length curve may be determined. 
 
 Fig. 22 
 
 It remains now to consider the nature of the change which 
 would be produced in the color curves, were it possible to use 
 the true distribution curves of the canvas exposures rather 
 than those of daylight. In the first place, the correction to 
 the curves probably falls where h is greatest, even though the 
 haze value has been erroneously obtained. To take an 
 extreme case, the curves given in Fig. 22 are to represent E w 
 and E w { 1 —a). In computing the first rectangle for the 
 haze-wave-length curve, of which the base is EF, the expres¬ 
 sion used was area = h 2 a 2 -hia u whereas what should have 
 
 52 
 
AERIAL HAZE 
 
 been used was area = h 2 {a 2 — b) —h\a x , which equals 
 
 hia 2 -hidi -h 2 b. Therefore, the area of the rectangle and 
 its altitude should have been smaller. For the second rec¬ 
 tangle, whose base is DE, the area was taken to be h 3 a 3 -h 2 a 2 
 instead of, as would have been correct, h 3 (a 3 -2b) -h 2 (a 2 -b), 
 which is h 3 a 3 —h 2 a 2 — 2h 3 b +h 2 b. Both the rectangle and its 
 altitudes were too large and should have been reduced. For 
 the third rectangle, with base CD, the area used was /z 4 a 4 -h 3 a 3 , 
 instead of the true value /? 4 (a 4 -2b) -h 3 {a 3 -2b) ; that is, 
 hidi —h 3 a 3 —2 b(h 4 —h 3 ). This rectangle was a little too large, 
 since in general Ju is greater than h 3 . The general effect is 
 that, although all the ordinates of the haze-wave-length curves 
 are too great, the maximum values should be reduced, but 
 at the same wave-length. The conclusion is, therefore, that 
 the light from the haze is more nearly white than it is repre¬ 
 sented to be by the haze-wave-length curves. Another 
 determination which was not made for lack of time was the 
 relative exposures at high and low altitudes from a single test 
 object. With such information, it would be possible to 
 separate the two components of h, —that is, ti and 1 -a. 
 Also, if this ratio could be obtained by light through various 
 filters, the spectral quality of the exposure E w ( 1 -a), and 
 hence the quality of light from the haze, could be obtained 
 more definitely than has been possible. The observations 
 necessary in order to secure these results would present no 
 particular difficulty, although the two exposures could not be 
 made upon the same plate. 
 
 53 
 
CHAPTER IV 
 
 The Duplication of an Aerial Condition 
 in the Laboratory 
 
 As already mentioned, haze so diminishes the contrast of 
 the subject to be photographed that it is necessary to determine 
 the photographic materials which should be used to produce 
 the best results from subjects of extremely low contrast, as 
 well as the best conditions for their use. For this purpose it 
 was desirable that a constant and controllable haze con¬ 
 dition should be established with which the material could be 
 studied. There was therefore erected in the laboratory an 
 apparatus by which an aerial condition could be simulated. 
 This apparatus, termed the “haze cabinet,” was a chamber 
 approximately eight feet square, with walls and ceiling 
 painted white. In it was placed a series of uniformly illumi¬ 
 nated patches of different reflecting powers, similar to those 
 used in the field work. The wall opposite the test objects 
 held two batteries, each containing five 750-watt daylight 
 tungsten bulbs which uniformly illuminated the test objects, 
 and which were enclosed in a parabolic reflector. The 
 aperture to the camera, which was also on this wall, was so 
 protected as not to receive any direct light from the lamps. 
 When the test objects were photographed a semi-transparent 
 and reflecting mirror was placed between the lens of the 
 camera and the objects. By means of this mirror a uniform 
 veil of light could be superposed on the images of the patches 
 to be photographed. The intensity of this veiling light could 
 be controlled by means of a neutral-tinted gelatine wedge. In 
 order that the cabinet might represent actual aerial conditions, 
 it was calibrated in terms of haze as measured under con¬ 
 ditions described above. The data obtained at Langley Field 
 which established the quality and the quantity of haze present 
 at various altitudes under various conditions, were utilized 
 for this purpose. A gelatine screen was made for converting 
 the light from the tungsten haze lamp to the color of the haze 
 as determined for any particular flight. Then, from observa¬ 
 tions of the distribution and quantity of haze at various 
 altitudes, the neutral tinted wedge was so calibrated that its 
 settings duplicated the quantity of haze at any desired altitude. 
 The exposure ratio at 500 feet (the ground contrast) was 
 also established by means of the test objects in the cabinet, 
 so that a given aerial condition was quite accurately reproduced. 
 
 54 
 
r 
 
 AERIAL HAZE 
 
 In order to know the actual contrast of the subject in the 
 cabinet and to determine how its range is disturbed by 
 superimposing known amounts of haze, an intensity scale 
 sensitometer similar to that used in the field work was printed 
 in contact with the plate by light reflected from a large white 
 screen of known selectivity in the cabinet. This screen was 
 then raised, the image of the strip just exposed covered, and 
 the test objects photographed. 
 
 Fig. 23 
 
 Exterior of haze cabinet 
 
 Fig. 23 shows the exterior of the cabinet with the camera, 
 the method of introducing the haze, and the control for both 
 the interior cabinet lights and the haze lights. Control for 
 both sets of lamps was effected by maintaining constant 
 amperage. 
 
 Fig. 24 illustrates in detail the means of producing the haze. 
 The light source is a 250-watt nitrogen-filled tungsten stere- 
 opticon lamp, above which is placed a parabolic reflector 
 focusing the image of the filament on a piece of ground glass 
 enclosed in a circular metal collar. An oblong case contains 
 a neutral tinted wedge of dyed gelatine by which the intensity 
 
 55 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 of the light falling through the second piece of ground glass 
 and thence on to the mirror is controlled. An electric blower 
 
 Fig. 24 
 
 Haze cabinet showing method of introducing haze 
 
 was used to keep the gelatine wedge cool. In order to obtain 
 perfect and uniform diffusion over an area sufficiently large 
 
 56 
 
AERIAL HAZE 
 
 for the purpose, the two diffusing media shown were neces¬ 
 sary. The diffused light is reflected from a semi-transparent 
 reflecting mirror through the lens and into the camera. The 
 mirror was prepared by the cathodic deposition in vacuum 
 of platinum iridium on glass, the coefficients of transmission 
 and reflection being controlled by the depth of the metallic 
 deposit. The light reflected from the test objects inside the 
 cabinet passed through the aperture in the wall, then through 
 the mirror placed at 45°, and thence into the camera. The 
 test objects could therefore be photographed “unhazed,” 
 or any desired amount of haze could be introduced, the amount 
 depending upon the setting of the wedge. The color of the 
 haze was controlled by introducing a color filter over the 
 second or lower opal glass. 
 
 The camera was fitted with a multiple back, in order to 
 make nine exposures of the subject on the same plate. Thus 
 photographs with various amounts of haze could be made on 
 the same plate, which assured a minimum of photographic 
 errors. The lens used was an Anastigmat (//7.7), of a type 
 similar to those used on the aerial cameras. 
 
 The illumination at the focal plane of the camera was first 
 determined visually by an illuminometer. A white screen at 
 the end of the cabinet reflected light through the platinized 
 mirror and lens system and illuminated a magnesium carbonate 
 block at the focal plane. A series of readings were taken 
 with the illuminometer directed toward this block. A similar 
 measurement was made using the haze light, again observing 
 at the block in the focal plane, the camera being focused on 
 the test objects in the cabinet. The wedge was set at zero, 
 in which position it transmits completely. However, the 
 illumination was too faint for settings of great accuracy, and 
 the data thus acquired are probably not so reliable as those 
 obtained by the photographic method. 
 
 The necessary calibration of the wedge was obtained visu¬ 
 ally from readings of the light incident on the platinized 
 mirror with the camera lens removed. The factor to be 
 applied to reduce this to its value at the focal plane was found 
 from readings taken for the zero position of the wedge. The 
 wedge proved to have densities very nearly proportional to the 
 distance from one end—that is, the logarithms of the trans¬ 
 mitted illumination varied directly as the wedge settings. At 
 the focal plane, the haze illumination varied between .4356 
 and .00213 meter-candles. At the position of the plate used 
 in exposing through the tablets, however, this was reduced, for 
 the zero wedge settings, from .4356 to .2240 meter candles. 
 
 57 
 
MONOGRAPHS ON THE THEORY OE PHOTOGRAPHY 
 
 The illumination from the white screen in the cabinet was 
 almost exactly one-third of that received from the haze light 
 through the wedge in zero position at the same plane. The 
 haze light could therefore be changed continuously from zero 
 to a value three times that of the brightest part of the subject. 
 
 As stated above, the test objects used in the cabinet were 
 miniature duplications of the canvases used in the field. For 
 these, six sheets of smooth-finished photographic paper were 
 exposed for increasing lengths of time and developed in a 
 developer which insured neutral results. The reflecting powers 
 of these test objects were measured visually with the Nutting 
 absolute reflectometer. Though this instrument does not 
 duplicate exactly the lighting conditions in the cabinet, the 
 results obtained give with sufficient accuracy the visual con¬ 
 trast of the subjects, as follows: 
 
 Subject No. 1 2 3 4 5 6 
 
 Reflecting Power 78.06 67.34 42.62 18.61 13.74 10.40 per cent. 
 
 Thus the extreme contrast between the subjects is about eight 
 to one, with sufficient gradations between to allow a study of 
 intermediate exposures. 
 
 The haze lamp of the cabinet was screened to a bluish color 
 to duplicate the color of haze on one particular flight. The 
 wedge was calibrated in terms of altitude or of haze values at 
 various altitudes on the same flight. The neutrality of the 
 wedge was then determined by noting any change in y on 
 plates exposed at intervals throughout the total range of the 
 wedge. This was done on ordinary, ortho, and panchromatic 
 plates, and no change in y, indicating a change of quality of 
 the light transmitted by the various sections of the wedge, 
 could be noted. Thus the cabinet duplicated a given flight 
 under conditions which could be maintained and controlled. 
 The photographic results were then related to the H. and D. 
 curves in such a way that the advantages or disadvantages of 
 various materials and treatments could be analyzed. 
 
 The importance of obtaining the sensitometric data under 
 precisely the conditions of the exposure studied has already 
 been emphasized. To ensure this, tablets were made which 
 would allow a strip to be exposed on each end of the plate used 
 for photographing the test objects. The strips used for the 
 intensity scale sensitometry, which were prepared in the same 
 way as those used in the field work, were found to be photo¬ 
 graphically neutral. One of the strips was exposed directly 
 to the haze glare when the cabinet lights were extinguished. 
 
 In making the other strip, for which the cabinet lights only 
 were used, the light was reflected from the white curtain 
 
 58 
 
AERIAL HAZE 
 
 which could be lowered in front of the test objects, and which 
 completely filled the field. To ensure uniform lighting of the 
 tablets, the plate was placed about ten centimeters behind the 
 focal plane. The strips were made on the same plate as the 
 images of the test objects, and the use of two strips made it 
 possible to compare the quality of haze and the cabinet 
 lights. By means of the multiple back, it was possible to 
 make nine exposures on each plate, thus obtaining a complete 
 range of haze intensities. 
 
 After development and subsequent fixation and drying of 
 the plates exposed as described above, the densities obtained 
 were read and referred to the curve plotted from the relation 
 between the densities obtained from the tablet and the ex¬ 
 posures which produced them, thus enabling a determination 
 of the range of the subject photographed to be made. This 
 was done with various amounts of haze superimposed. 
 
 Thus it was possible to study the way in which haze reduces 
 contrast and how various photographic materials behave under 
 identical haze conditions. The data obtained were plotted in 
 curves showing the relation between the contrast of subject 
 and the haze over the range obtainable under cabinet con¬ 
 ditions, and the relative behavior of the various types of 
 photographic materials. At the same time, data were ob¬ 
 tained as to the exposure and development for securing the 
 best results under haze conditions. 
 
 The sensitometric characteristics of the plates best suited 
 for air work were studied and methods of interpretation 
 devised by which suitable sensitometric methods and data 
 would give all possible information concerning the usefulness 
 of a plate for photography in the air. 
 
 The effect to be studied may be called “contrast,” but the 
 word must not be understood to be used in the sense in which 
 it was used in the section on sensitometry. There contrast 
 signified the rate of change of density with the logarithm of 
 exposure, and depended solely on the plate used and its 
 development. Here it is desirable to classify the plates as 
 records rather than as related to the exposures to which they 
 were subjected. Thus we can measure the contrast as the 
 difference in the densities produced for a given pair of subjects 
 or, more conveniently, as the ratio of the opacities, the value 
 of which the density difference is the logarithm. The most 
 suitable plate for aerial photography will give the greatest 
 opacity ratio for a given pair of subjects under the most 
 unfavorable conditions of haze and under exposure. 
 
 59 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 When both the subjects lie on the straight-line portion of 
 the H. and D. curve it is quite easy to predict the effect that 
 haze glare will have on the opacity ratio, especially when the 
 haze and the subject are of the same color. Let Ei represent 
 the exposure due to the first subject, E 2 that due to the second, 
 and e the exposure due to haze, to be added to both E i and E 2 . 
 The equation for the straight-line portion of the H. and D. 
 curve may be written. 
 
 D log 0 = y (log E - log i ); 
 
 and for the two subjects, therefore 
 
 log 0i= f (log [Ei + e] - log i) 
 
 log 0 2 = Y (log [E 2 + e]— log i). 
 
 Subtracting, 
 
 log (0i/0 2 ) = y 
 denoting Oi/0 2 by C, 
 
 which may be written (C — — 1) (E 2 + e) = E x — E 2 
 
 Y 
 
 In the case where y is unity, this is the equation of a 
 rectangular hyperbola between C and e with the asymptotes 
 C = 1, e = — E 2 - It intersects the axis e = 0 at C = Ei/Es, 
 the true exposure ratio. As y increases, all values of C 
 increase, and the curve does not cling to its asymptote C = 1 
 so soon. Lower y’ s decrease all values of C. 
 
 It is clear that if one could be sure of obtaining all exposures 
 on the straight-line portion of the curve, the problem would 
 consist only in obtaining the highest possible value of y. 
 But exposures from the air are frequently so short that the 
 densities are apt to lie on the toe of the curve. Aside from 
 under-exposure, the question of latitude is a vital one, especi¬ 
 ally for high values of y, where it is not only reduced but also 
 complicated by fog at the low densities. 
 
 The effect of y on opacity ratio in practice is shown in the 
 curves in Fig. 25. The different curves correspond to settings 
 of the neutral wedge, that is, to different amounts of haze 
 glare. According to the equation above, these curves should 
 be logarithmic. They have this general shape, although 
 somewhat distorted by toe and shoulder effects and unequally 
 
 60 
 
OPACITY RATIO 
 
 AERIAL HAZE 
 
 Fig. 25 
 
 Variations of contrast with y for 
 various amounts of haze 
 
 so because corresponding ordi¬ 
 nates are not necessarily given 
 by equivalent exposure times. 
 
 Fig. 26 shows another set of 
 curves obtained with a Seed 
 23 plate. The abscissae here 
 are haze brightnesses. We see 
 the hyperbolic type predicted 
 from the equation for these 
 coordinates. Each curve cor¬ 
 responds to a given y and a 
 fixed exposure time. The 
 group of similar flat-lying 
 curves, marked A, which are 
 displaced upward with increas¬ 
 ing y, were obtained with an 
 exposure which kept the sub¬ 
 ject as near as possible in an 
 equivalent position on the cor¬ 
 responding H. and D. curve. 
 The curves marked B were 
 given by longer exposures, 
 
 Fig. 26. Relation between contrast and haze for various values of y. 
 B curves for longer exposures than A curves 
 
 61 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 which reduce the toe effect on low haze values and increase the 
 shoulder effect on high haze values. This illustrates how im¬ 
 portant the shape of the toe may be when latitude has been 
 exceeded, or when the exposure is shorter than that usually 
 required by the plate. One may see also in these curves a fact 
 which was still more obvious in Fig. 25, viz., that as y increases, 
 the opacity ratio becomes increasingly sensitive to slight varia¬ 
 tion of y and consequently of the time of development. 
 
 The question of exposure 
 
 EXPOSURE TIME 
 
 Fig. 27 
 
 Relation between contrast and 
 exposure time for differing 
 amounts of haze 
 
 time is further dealt with in Fig. 
 27. Here the abscissae are ex¬ 
 posure times and the various 
 curves are drawn for different 
 amounts of haze. Themaximum 
 of the curve in each case indi¬ 
 cates the exposure bestsuitedto 
 give high contrast for the pair 
 of subjects in question. The 
 increasing intensity of light 
 (subject + increasing haze), 
 naturally shifts this maximum 
 to the right for decreasing haze. 
 These curves were obtained 
 from Seed 23 plates and for high 
 y’s. It is interesting to note 
 how steep these curves may be 
 for slight haze,and how marked 
 their maximum. 
 
 esti 
 
 Fig. 28 shows a group of curves similar to those in Fig. 
 26, but selected so that the exposure falls as nearly as possible 
 at its optimum. Such a series of curves represents an emul¬ 
 sion at its best. In practice, exposures could not be 
 mated so accurately. The pho¬ 
 tographer is therefore inter¬ 
 ested in knowing how rapidly 
 the contrast diminishes with a 
 change of exposure. 
 
 (In Fig. 29 an endeavor has 
 been made to assemble opti¬ 
 mum curves of the type of Fig. 
 
 28 for a group of emulsions. 
 
 These curves unfortunately 
 contain no information as to 
 speed, which is doubtless very 
 necessary. They give, however, 
 
 62 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 6A 
 
 NM/ 
 
 
 NS 
 
 TAh 
 
 T 
 
 
 
 
 
 
 
 
 
 E) 
 
 t K5 
 
 SUI 
 
 E 
 
 ‘IM 
 
 E C 
 
 ON 
 
 1A 
 
 Ml 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 r 
 
 r 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 V 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 v 
 
 
 s 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 X 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 — 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Fig. 28 
 
 Relation between contrast and 
 haze for various gammas 
 
AERIAL HAZE 
 
 HAZE CMS 
 
 Fig. 29 
 
 Relation between contrast and 
 haze for various emulsions 
 
 the conditions under which the 
 
 the relative value of the emul¬ 
 sions for aerial work at each 
 haze value, assuming that the 
 maximum exposure is suffi¬ 
 ciently high to allow each 
 plate its optimum exposure. 
 If the plate were too slow for 
 that it would be necessary to 
 substitute for its curve one 
 obtained at the best exposure 
 possible.) 
 
 Thus it is seen that, from 
 measurements made with the 
 haze cabinet, the relation be¬ 
 tween haze and the contrast 
 obtained in the photograph 
 can be determined, as well as 
 effect of haze is minimized and 
 
 the best contrast is obtained. In the next chapter, the 
 practical applications of the results obtained, as well as the 
 use of color filters to increase the contrast in aerial photography, 
 will be discussed. 
 
CHAPTER V 
 
 The Materials and Conditions Best adapted for 
 Aerial Photography 
 
 Photographic emulsions are classified by means of the con¬ 
 stants which control their characteristic curves. In Fig. 30 
 are shown the curves of two materials differing in speed and 
 in contrast, development having been carried to the limit, so 
 that the contrast is represented by the maximum slope of the 
 line obtained. One of the emulsions shown would be classified 
 as a fast emulsion of moderate contrast, y oo being 1.6, and the 
 other as a slow emulsion of great contrast, y oo being approxi¬ 
 mately 2.6. 
 
 Since in aerial photography 
 the materials are very often 
 under-exposed, their relative 
 speeds can not be compared in 
 the correct exposure region, as 
 is done according to the con¬ 
 ditions laid down by Hurter 
 and Driffield. Instead, the 
 relative exposures required to 
 produce sufficient printing 
 density throughout the region 
 of under-exposure are com¬ 
 pared , since, as has been stated, 
 in practical aerial photography 
 the region of under-exposure 
 is generally of greater im¬ 
 portance than that of correct 
 exposure. The characteristics of the under-exposure portion of 
 the curve are best shown not by plotting the density against 
 the logarithmic exposure, but by the derivative of this curve; 
 that is to say, the slope of the characteristic curve is plotted 
 against the logarithmic exposure, dD/d log E against log E. 
 This is shown in Fig. 31. The gamma of the characteristic 
 curve is now shown as the maximum height of the derivative 
 curve, the straight-line portion of the characteristic curve being 
 represented by the portion of the derivative curve which is 
 horizontal and parallel to the X axis, and the sloping portions 
 of the derivative curve corresponding respectively to the under¬ 
 and over-exposure portions of the characteristic curve. 
 
 64 
 
 A—Fast emulsion of moderate con¬ 
 trast 
 
 B—Slow emulsion of great contrast 
 
AERIAL HAZE 
 
 When the derivative curves are plotted for the same emul¬ 
 sion for different times of development, the series of horizontal 
 lines obtained correspond to the different y values of the 
 characteristic curves, and if no regression of the inertia occurs, 
 the point at which the line becomes horizontal remains fixed 
 
 Fig. 31 
 
 Relation between slope of characteristic curve and exposure 
 
 with regard to the exposure axis as development is continued; 
 that is, a line passing through the junction points of the under 
 and correct exposure regions for the different kinds of develop¬ 
 ment will be vertical, as is shown in Fig. 32. In practice, how- 
 
 LOG E 
 
 Fig. 32 
 
 Derivative curves for various gammas on same emulsion without 
 regression of inertia 
 
 65 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 ever, there is usually regression of the inertia, and the line 
 passing through the junction point is no longer vertical, the 
 influence of the regression being shown by a shift of the point 
 as development is continued (Fig. 33). Fig. 34 shows a 
 typical case of this, the plate being the Stanley Commercial. 
 Increased development has caused a considerable regression 
 of the inertia, the horizontal portion of the curve (correspond¬ 
 ing to the straight-line portion of the H. and D. curve beginning 
 at lower exposure for the upper curve. Fig. 35 shows the 
 behavior of a group of emulsions with regard to the under¬ 
 exposed region, these being drawn for the best values of y. 
 
 From the work in the haze cabinet it is clear that the plate 
 which produces the greatest opacity ratio for a given contrast 
 of subject under the most unfavorable haze conditions is most 
 suitable for aerial photography. The materials selected, 
 therefore, will be of high contrast and consequently of some¬ 
 what low sensitiveness, and the regression of the inertia will 
 make it possible to use materials of higher contrast than would 
 
 Fig. 33 
 
 Derivative curves for various gammas showing regression of inertia 
 
 otherwise be practicable, owing to the limitation imposed by 
 considerations of exposure. No material is suitable for use in 
 aerial photography the y oo or limiting contrast of which is 
 lower than 2, while the practical conditions of exposure show 
 very clearly that the speed under the conditions of maximum 
 development must not be lower than 100 H. and D. in order 
 to allow for the use of the necessary color filter. The question 
 of exposure requires special consideration. 
 
 The longest exposure which can conveniently be given in an 
 aeroplane is about 1/50 of a second, but under ordinary work¬ 
 ing conditions it is more practical to assume that the longest 
 
 66 
 
AERIAL HAZE 
 
 exposure should not be more than 1/100 of a second because 
 of the effect of vibration. Therefore, 1/100 of a second may 
 be selected as the standard convenient exposure for airplane 
 photography. This assumes, of course, the use of an airplane 
 having the least vibration, and of the most suitable camera 
 and suspension, and the placing of the camera and suspension 
 in the best part of the aeroplane. 
 
 It can be shown 1 that the effective speed of the emulsion is 
 related to the exposure and to the intensity of the light re¬ 
 flected from the subject by the equation 
 
 LOG E 
 
 Fig. 34 
 
 Relation between slope of 
 characteristic curve and ex¬ 
 posure for two gammas 
 
 LOG E 
 
 Fig. 35 
 
 Relation between slope of 
 characteristic curve and ex¬ 
 posure for various emulsions 
 
 J00_ 
 
 LE ' 
 
 where E is given in seconds and L in foot-candles of brightness. 
 This equation applies to a lens of aperture//8. If, instead of 
 a lens of aperture //8, we use a lens of aperture //4.5, the 
 equation becomes 
 
 500 
 
 3L E ' 
 
 Filling in 1/100 for E, we have 
 
 50,000 
 F ~ 3 L ’ 
 
 1 Mees, C. E. K., Fundamentals of Photography, p. 55. (Eastman Kodak Company. 
 Rochester, New York). 
 
 67 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 For a sunlit landscape from the air, L will have a value as 
 high as 1000 foot-candles, and under these conditions 
 
 D 50,000 . , _ 
 
 r = — -qq , or approximately 17. 
 
 For aerial photography in bright sunlight, therefore, ma¬ 
 terials having a working speed (with a filter), of approximately 
 17 can be effectively used. For the most red sensitive plates 
 available, the factor of the red filter has been reduced to 6. 1 
 Under the best aerial conditions, therefore, it would be possible 
 to use such plates with a red filter, even if their H. and D. 
 speeds did not exceed 100. It is, of course, quite possible to 
 obtain materials of this speed having the requisite contrast. 
 
 It will be seen that the red filters can be used only under 
 the very best conditions, when strong sunlight is falling upon 
 the subject to be photographed. Under other conditions, 
 lighter filters are necessary. When, instead of panchromatic 
 materials, those sensitive to the yellow-green only are used, 
 the very best materials examined were found to have with the 
 No. 12 strong yellow filter an effective speed of 25. These 
 would make possible exposures from the air under conditions 
 approximately half as favorable as the best conditions which 
 occur, even with this strong yellow filter. The following table 
 of various materials and their effective speeds through different 
 filters is useful in this connection: 
 
 Speed Effective Speed through Filters 
 
 
 with 
 
 
 
 Aero 
 
 Aero 
 
 No. 
 
 No. 
 
 Material 
 
 no filter 
 
 K-l 
 
 K-2 
 
 1 
 
 2 
 
 12 
 
 25 
 
 Process Panchromatic 
 or M 
 
 60 
 
 40 
 
 20 
 
 30 
 
 25 
 
 10 
 
 5 
 
 Panchromatic Plates 
 (Ordinary) 
 
 120 
 
 80 
 
 40 
 
 60 
 
 50 
 
 20 
 
 10 
 
 Panchromatic Plates 
 (Extreme red sensitive) 
 
 200 
 
 150 
 
 70 
 
 120 
 
 80 
 
 50 
 
 30 
 
 Panchromatic Aero Film 
 
 120 
 
 60 
 
 30 
 
 45 
 
 35 
 
 15 
 
 7 
 
 Panchromatic Aero Film 
 (Hypersensitized) 
 
 200 
 
 150 
 
 70 
 
 120 
 
 80 
 
 50 
 
 30 
 
 Orthochromatic Plates 
 (Ordinary) 
 
 200 
 
 60 
 
 20 
 
 40 
 
 30 
 
 8 
 
 
 Orthochromatic Plates 
 (Extreme green sensitive) 
 
 200 
 
 100 
 
 50 
 
 70 
 
 50 
 
 25 
 
 
 Aero Ortho Film 
 
 150 
 
 50 
 
 15 
 
 30 
 
 25 
 
 8 
 
 
 Seed 30 Plates 
 
 1 See Chapter II. 
 
 400 
 
 150 
 
 68 
 
 20 
 
 50 
 
 30 
 
 5 
 
 
AERIAL HAZE 
 
 The curves shown in Figs. 21 and 22 (pp. 51 and 52), giving 
 the distribution of haze at different wave-lengths, make clear 
 the great importance of the use of color filters restricting the 
 exposure as much as possible to the longer wave-lengths. The 
 haze is predominantly blue with a decided maximum at 450 [L[k. 
 It is obvious, therefore, that a sharp cut yellow filter is of the 
 greatest advantage in removing the greater part of the haze, 
 but that the longer the average wave-length which can be used, 
 the greater will be the penetration obtained. 
 
 A considerable number of filters were used experimentally in 
 the work. The conclusion reached by the Army fliers was that 
 there is no advantage in using a filter lighter than the K-l>£, 
 but this was before the new filters made with Eastman Yellow 
 were introduced. When these were used, it was found that the 
 
 Fig. 36 
 
 Aero No. 1 filter, which is a little lighter than the K-l>^, 
 gives a very decided increase in the effective contrast. The 
 filters finally adopted were Aero No. 1, K-l>^, Aero No. 2, 
 K-2, No. 12, No. 15, No. 21, and No. 25. The absorption 
 curves of filters Aero No. 1 and Aero No. 2 are shown in Fig. 
 36. and K-2 may be eliminated, as they are certainly 
 
 inferior to Aero No. 1 and Aero No. 2. No. 15 and No. 21 may 
 also be excluded, which leaves a set of four filters best suited 
 
 69 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 for aerial photography: Aero No. 1, Aero No. 2, No. 12, 
 and No. 25. The choice between these is determined by the 
 exposure conditions available. For ordinary panchromatic 
 plates the exposure factors are as follows: 
 
 Panchro- 
 
 Extreme Red 
 Sensitive 
 
 Panchro¬ 
 
 Hypersensi¬ 
 
 Aero Ortho 
 Film and 
 
 Extreme 
 
 Sensitive 
 
 
 matic 
 
 Panchromatic 
 
 matic 
 
 tized Panchro¬ 
 
 Orthochro¬ 
 
 Orthochro¬ 
 
 Filter 
 
 Plates 
 
 Plates 
 
 Film 
 
 matic Film 
 
 matic Plates 
 
 matic Plates 
 
 Aero No. 1 
 
 2 
 
 IK 
 
 2 K 
 
 IK 
 
 5 
 
 3 
 
 Aero No. 2 
 
 3 
 
 2K 
 
 4 
 
 2 K 
 
 6 
 
 4 
 
 No. 12 
 
 6 
 
 4 
 
 8 
 
 4 
 
 20 
 
 8 
 
 No. 25 
 
 12 
 
 6 
 
 14 
 
 6 
 
 
 
 Obviously, the strongest filters, Nos. 12 and 25, can be used 
 only with the most sensitive panchromatic materials. They 
 can be used with Wratten panchromatic plates under the best 
 conditions of lighting, and No. 12 can be used with panchro¬ 
 matic film under similar conditions. When the lighting is in 
 any way inferior, and generally when it is necessary to use No. 
 25, hypersensitized materials are necessary. The filter known 
 as Aero No. 2 can be used only under good conditions of light¬ 
 ing and with strongly orthochromatic film or plates. For 
 such materials, with lower color sensitiveness, Aero No. 1 is 
 the standard filter. 
 
 The conditions of development used should be those pro¬ 
 ducing the lowest possible inertia and at the same time the 
 highest possible contrast in the resulting negative. In practice, 
 this involves the use of developers containing a large amount of 
 restrainer, which produce high contrast without fog, and at the 
 same time having high reduction potential, which favors the 
 production of negatives corresponding to the lowest possible 
 inertia in the emulsion or, in other words, gives the maximum 
 detail in the shadows. After a careful sensitometric study of a 
 number of developing formulae a formula was selected as 
 giving the best practical results; but there are two objections 
 to the use of this developer—(1) it proved very expensive, and 
 (2) it contained chlorhydroquinone, the supplies of which 
 during the war were limited, and which is not always con¬ 
 veniently available. The Aerial Service therefore substituted 
 the following formula, and though this does not give quite as 
 low an inertia as the metol-chlorhydroquinone formula origin¬ 
 ally selected, it is very similar, and is quite satisfactory: 
 
 Metol.16 grams; 
 
 Hydroquinone.16 grams; 
 
 Sodium sulphite.60 grams; 
 
 Sodium hydroxide.10 grams; 
 
 Potassium bromide.10 grams; 
 
 Water to. 1 liter. 
 
 70 
 
AERIAL HAZE 
 
 Sensitometric comparison of this developing formula with 
 the pyro or pyro-metol formula previously used, shows that the 
 former requires appreciably less exposure and that it gives a 
 somewhat higher contrast than could be obtained by the 
 latter. 
 
 71 
 
APPENDIX 
 
 The H aze Effect Produced by Pure Dry Air 
 
 Although the mass of air below a certain altitude changes 
 with the temperature and pressure, this mass is much more 
 nearly constant than are the amounts of the other constituents 
 of haze. Also, the distribution of the air, or its change of 
 density with altitude, is, for isothermal conditions, definitely 
 known. The actual transmission of light by dry air, as a 
 function of wave-length, is given in a paper by Fowle. 1 
 This makes it possible to estimate the amount of haze due 
 to air alone. The methods of the calculation apply as well to 
 the effects of water vapor, but as we do not know the mass 
 involved, its distribution, or the transparency of the material, 
 we can not get a numerical result for the amount of haze which 
 it causes. To go back to these quantities from observational 
 data may become possible, but will require the accumulation 
 of much additional material. 
 
 The magnitude of the dry air effect may be estimated as 
 follows: Assuming a uniform temperature and composition, 
 if d 0 is the density of the air at the altitude a = 0, the density 
 d, at altitude a is 
 
 — pjna 
 
 d = d 0 e RT 
 
 where e, g, m, R, and T are respectively the base of the natural 
 log system, the acceleration due to gravity, the “molecular 
 weight,” 29, of air, the gas constant, and the absolute tem¬ 
 perature. Evidently g depends upon a, but in so small a 
 measure, for the altitude concerned, that it may be regarded as 
 constant. Taking T = 283 we have 
 
 d = d 0 e ~ ba 
 
 and b = 121 • 10~ 8 in c. g. s. units. 
 
 Consider now a very thin layer of the diffusing medium with 
 parallel radiation incident upon it. This radiation is weakened 
 by scattering and absorption. The decrease in intensity di, 
 where i is the incident light (measured in ergs per cm. 2 per 
 second), is proportional to the original intensity i, to da the 
 thickness, and to d, the density. Thus, 
 
 'Fowle, The Atmospheric Scattering of Light, Smithsonian miscellaneous collections, 
 Vol. 69, No. 3. 1918. 
 
 72 
 
AERIAL HAZE 
 
 di = — c'id • da; 
 
 di represents a decrease in i, and, if da is an increase in a, the 
 minus sign should be used. Of this loss di, the part scattered 
 is, say, c'id • da. It is assumed that the light is scattered 
 equally in all directions. Now the receiving surface or the 
 open area of the lens is to be at a height S, the medium will 
 extend further up to altitude t, and the diffusing layer in 
 question is at a height a, below S. Of this whole layer, the 
 extent of which is very great, a limited area A , determined by 
 the solid angle of the lens, sends light towards the receiving 
 surface. If this solid angle is £ and d is the distance from S to 
 a, 
 
 £ = A /d 2 
 
 Now, from unit area of the layer, corresponding to a volume da, 
 the scattered radiation is: 
 
 c' i a d Q • e~ ba da, 
 
 and of this the fraction ^ where L is the effective area of 
 
 4 zd 2 
 
 the stop, will start toward the lens. Since A is small compared 
 to d 2 , the effective radiation dr from the whole layer is 
 
 dr = C A a e~ ba da. 
 
 4 x 
 
 Of this, however, not all strikes the lens, as it is subject to the 
 scattering between its own level a, and that of the lens, 5. 
 Here, for the loss in any layer, since the radiation is directed 
 upward, 
 
 di = — ci'a d 0 e ~ ba da 
 Whence, integrating from a to s, 
 
 log i, - log i = Sj* (« -bs - e - ba ) 
 or i s = i a e T (e- bs -«~ ba ) 
 
 Therefore, from the layer at altitude a there actually reaches 
 the surface of the lens the radiation 
 
 dr = 
 
 c L Me 
 
 At. 
 
 A. 
 
 —ba 
 
 e e 
 
 cd 0 (g — bs_ g — ba') 
 
 b da. 
 
 If ii is the radiation at a height t, above S and a, it will be 
 subject, in going down to a, to the same sort of diminution, 
 so that 
 
 • . cd 0 ( p —bt 
 
 «a = *t e \ e 
 73 
 
 e “ ba ). 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 Finally, 
 
 C* L Mo • -ba c i° (g~bs _|_ g-bt _ 2g-ba) 
 
 --p—- // e e b «a 
 
 c' L Mo • -- 0~ bs + e_bt ) «~ ba « b 
 
 --it ^ b 
 
 which can be written (combining symbols into J and v) 
 
 dr = J e ba e ve ba da. 
 d (e _ba ) = — &e~ ba da, 
 
 we can write 
 
 Since 
 
 J 
 
 ^ e vx dx. 
 
 —ba d (g~ ba ) 
 g-ba 
 
 dr = — e ba e ve 
 b 
 
 If x — e -ba , the total radiation from the haze is evidently 
 
 or ultimately, 
 
 -2 
 
 e 
 
 This is the light from the haze reaching the lens within its 
 whole solid angle. Let the transmission of the lens be repre¬ 
 sented by the factor M, and let / be the focal length of the 
 lens, so that l// 2 will be the solid angle corresponding to unit 
 area of the plate exposed; then, multiplying R' by M and by 
 l// 2 , and dividing by we shall have for the exposure on the 
 plate in ergs per cm. 2 per second: 
 
 This is an expression for the “luminous veil effect” of the haze. 
 
 Evidently, if s =0, we have R = 0. Also, ( d a = d 0 e~ ba 
 being the expression for the density of the medium), if b = 0, 
 we have d a = d Q , or the density is uniform. If this be the 
 case, 
 
 74 
 
AERIAL HAZE 
 
 an expression which was obtained independently, thus check¬ 
 ing the preceding formula for R. 
 
 The subtractive effect of the haze may be considered as 
 follows: As has been stated, the illumination of the test 
 objects (in ergs per cm. 2 per second) is 
 
 since a = 0. Let p represent the reflective power of the test 
 object, the fraction of normal incident radiation scattered, 
 per unit solid angle, in a vertical direction by the white can¬ 
 vas. Each square centimeter of this area sends toward the lens 
 
 ergs per second. This is subjected 
 
 to the diffusing action of the haze in the region below 5 . As 
 on page 73, where, for radiation directed upward, 
 
 the radiation incident upon the lens is, since a = 0, 
 
 This is partially transmitted by the lens, and falls upon an 
 area f 2 /s 2 of the plate. Therefore, the exposure in ergs per 
 cm. 2 per second is 
 
 If 5 = 0, 
 
 This quantity was used in the definition of haze. Also, if b =0 
 the medium is uniform and 
 
 It is possible now to obtain an expression for haze, since 
 
 E' w =(1 - d) E w , and therefore 1 - d = exp ( e bS — ^ 
 
 Also, h' E w =R. Hence, 
 
 75 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 if (i - e " s) ] - if 
 
 Therefore, since h = 
 
 h! 
 
 1 -d 
 
 h 
 
 8k pc 
 
 (■ 
 
 (1 — e _bs 
 
 -) 
 
 This is the desired expression for the haze effect of dry air, or 
 for any medium whose density changes with altitude according 
 to the same law. As a special case in which b = 0, it gives the 
 effect for a uniform distribution. To obtain actual values 
 for h , however, it is obviously necessary to have values for the 
 constants involved, namely c', cp, d Q , and b. It is possible to 
 find these for dry air, as follows: 
 
 In the first place, d refers to the scattering effect of the 
 medium, and c to the total decrease in the passing light. 
 They are not equal if there is any absorption. But, according 
 to the work of Strutt, 1 we may take the absorption to be 
 
 very small, and may be put equal to 1 in the coefficient of the 
 
 h formula. The evaluation of p, which, it will be remembered, 
 is the fraction of normal radiation scattered vertically per unit 
 solid angle, is at least approximately possible. From measure¬ 
 ments made upon the white canvas, it appears to scatter 
 throughout the hemisphere 0.6 of the radiation incident per- 
 pendiculaily, and in this nearly to follow Lambert’s law. In 
 a vertical direction the radiation is ip per unit solid angle, and 
 therefore in the small solid angle do it is ip do. In any other 
 direction determined by a, as in the figure (Fig. 37),the inten¬ 
 sity is ip d o cos 7. Since i 
 depends on a only, we can select 
 an elementary solid angle do 
 corresponding to an infinites¬ 
 imal zone contained between a 
 and a +da; that is to say, 
 
 do = 2x sin a . dz. 
 
 Therefore, at any angle, Ldo 
 becomes 
 
 2 % i v cos a sin a d a. 
 
 For the sphere the total radia¬ 
 tion is 
 
 1 Strutt, The light scattered by gases; its polarization and intensity. Proc. Roy. Soc. 
 A95: 155. 1919. 
 
 76 
 
AERIAL HAZE 
 
 _ 7C 
 
 2 x i p §cos a sin x d a 
 
 a = o 
 
 7C 
 
 i p | sin 2 x | 2 =% i p . 
 0 
 
 This is 0.6 times the incident light. Therefore ip, the vertical 
 radiation per unit solid angle, is 0.6 or 0.2 of the normal 
 incident radiation, and therefore p = 0.2. 
 
 In the paper by Fowle the fractional transmission of dry air 
 above Mt. Wilson, altitude 1730 meters, is given. As on 
 page 73, 
 
 /a = It exp e~ ht -e ~ ba ) 
 
 Here a = 173,000 cms; t is very great, since we have the 
 transmission of the whole layer, and hence l a : 1,, which is the 
 fraction given by Fowle, is 
 
 —cd 0 p —ba 
 
 e b 
 
 We know b, a, and d Q , which are, respectively, 123, 108, 173,000 
 and 0.001293 in c. g. s. units. Therefore, we can find c for 
 any wave-length for which the fraction is tabulated. The ex¬ 
 pression on page 73, then, gives h for any altitude a. In the 
 following table, / is an average value of the fractional trans¬ 
 mission of air above Mt. Wilson, and h is the haze effect, as 
 defined, at an altitude of 10,000 ft. The wave-length is given 
 in A. U. 
 
 X 
 
 / 
 
 h 
 
 3500 
 
 .630 
 
 0.086 
 
 3710 
 
 .686 
 
 0.067 
 
 3970 
 
 .752 
 
 0.050 
 
 4130 
 
 .783 
 
 0.042 
 
 4310 
 
 .808 
 
 0.036 
 
 4520 
 
 .840 
 
 0.029 
 
 4750 
 
 .863 
 
 0.024 
 
 5030 
 
 .885 
 
 0.019 
 
 5350 
 
 .898 
 
 0.017 
 
 5750 
 
 .905 
 
 0.016 
 
 In the tabte below values of h at various altitudes are given for 
 X = 4000 A. U. These values are represented graphically on 
 page 78. It will be seen, by comparison with field data, that 
 the haze effect at \ = 4000 A. U. due to dry air, is about 
 one-fifth that found. 
 
 77 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 a (ft.) h 
 
 a (ft.) 
 
 
 h 
 
 6,000 
 
 0 
 
 .0298 
 
 7,000 
 
 0 
 
 .0344 
 
 8,000 
 
 0 
 
 .0390 
 
 9,000 
 
 0 
 
 .0436 
 
 10,000 
 
 0 
 
 .0480 
 
 It may be remarked in clos¬ 
 ing that the haze altitude curve 
 for dry air, though very nearly 
 a straight line, is a little convex 
 upward. (Fig. 38.) If we should 
 plot h for a homogeneous me¬ 
 dium, by aid of the formula on 
 page 76, for the case of & = 0, 
 the curve would be concave up¬ 
 ward. The effect of a decrease 
 of density of the haze material 
 with altitude appears to be to 
 straighten this curve, or even to 
 change its curvature until it 
 The magni- 
 the relative 
 
 Fig. 38 
 
 becomes convex upward, as is the case for air. 
 tude of these effects, however, depends upon 
 values of the constants (d 0 , b, c ) involved. 
 
 78 
 
Bibliography 
 
 Adams, E. Q., and Haller, H. L., Kryptocyanines. A new series of 
 photosensitizing dyes. J. Amer. Chem. Soc. 42: 2661. 1920. 
 
 Burka, S. M., Hypersensitizing commercial panchromatic plates. T. 
 Frankl. Inst. 189 : 25. 1920. 
 
 ^-' ALI ^ ER > A., Absorption and scatter of light by photographic negatives. 
 Phot. J. 33 : 200. 1909. 8 
 
 Eberhard, G., Ueber die gegenseitige Beeinflussung benachbarter Felder 
 auf einer Bromsilberplatte. Physik. Zeits. 13 : 288. 1912. 
 
 Fowle The atmospheric scattering of light. Smithsonian miscellaneous 
 collections. 69 : No. 3. 1918. 
 
 Jones, L. A. 
 189 : 303. 
 
 A new non-intermittent sensitometer. I. Frankl Inst 
 1920. 
 
 39. 
 
 • —•, On the theory of tone reproduction. 
 1920. 
 
 Mees, C. E. K., Fundamentals of Photography, 
 pany. Rochester, New York.) 
 
 J. Frankl. Inst. 190 : 
 (Eastman Kodak Com- 
 
 , and Gutekunst, G. O., Some new sensitizers for the deep red 
 J. Ind. Eng. Chem. 14 : 1060. 1922. 
 
 Nietz, A. H., The theory of development. Monograph No. 2 (D Van 
 Nostrand Co.) 
 
 Ross, F. E., On the relation between photographic density, light intensity 
 and exposure time. J. Opt. Soc. Amer. 4: 255. 1920. 
 
 -. —• —Photographic photometry and the Purkinje effect. Astrophvs 
 
 J. 52: 86. 1920. * 
 
 > • •> The mutual action of adjacent photographic images. 
 
 Astrophys. J. 53: 349. 1921. 
 
 Sheppard, S. E., and Mees, C. E. K., Investigations on the theory of the 
 photographic process. (Longmans, 1907). 
 
 Smithsonian miscellaneous collections, Vol. 69 : No. 3., 1918. The at¬ 
 mospheric scattering of light. 
 
 Strutt, The light scattered by gases: its polarization and intensity 
 Proc. Roy. Soc. A95: 155. 1919. 
 
 Wratten, S. H., and Mees, C. E. K., The wedge spectrograph. Brit. T. 
 Phot. 54: 384. 1907. s u g u j 
 
 This is not a complete bibliography of the subject of aerial haze. A 
 complete bibliography of aerial photography may be found in the Phot. J. 
 45: 396. 1921. 
 
 79 
 
Index of Authors 
 
 Adams, E.Q., and Haller, H. L., 33 
 Beers, 29 
 Burka, S. M., 33 
 Callier, A., 31 
 
 Driffield, Hurter and-, 19, 
 
 20, 29, 64 
 Eberhard, G., 24 
 Fowle, 72, 77 
 
 Gutekunst, G. O., Mees and 
 -, 33 
 
 Haller, H. L., Adams, E. Q., 
 and-, 33 
 
 Hurter and Driffield, 19, 20, 
 
 29, 64 
 
 Jones, L. A., 20, 27 
 
 Koenig, E., 32, 33 
 
 Lambert, 76 
 
 Martens, 30, 47 
 
 Mees, C. E. K., 67 
 
 -, Wratten, S. H., and 
 
 -,34 
 
 -and Gutekunst, G. O., 33 
 
 Meithe, 32 
 
 Nietz, A. H., 23 
 
 Nutting, P. G., 58 
 
 Ross, F. E., 20, 24, 25, 26 
 
 Sheppard, S. E., 30 
 
 Strutt, 76 
 
 Schwarzchild, 25 
 
 Wratten, S. H., and Mees, C. E. 
 
 K., 34 
 
 80 
 
Index of Subjects 
 
 Absorption curve of filters, 69 
 Absorption effect of haze, 50 
 Aerial photography, 16 
 detail in, 16 
 
 filters best suited for, 68, 70 
 plates for, 59, 64, 66, 68 
 standard exposure for, 66 
 Air, fractional transmission of dry, 77 
 producing haze effect, 72 
 Ammonia, use in sensitizing plates, 33 
 Atmospheric haze, 32 (see also Haze) 
 elimination of, 32 
 
 Beers’ law, 29 
 
 Bromide, effect on developer, 22 
 
 Camera, four-lens type, 40, 44 
 Characteristic curves, 19, 23, 34, 42, 43 
 for different filters, 35, 69 
 for different wave-lengths, 26 
 
 of photographic emulsions, 19, 20, 22, 23, 25, 41, 43, 45, 58, 
 60, 64, 65, 66 
 
 Color factor, variation with source of reflected light, 44 
 Color of light, 16 
 
 effect on photographs, 17 
 Color sensitive materials, 32 
 Color sensitiveness, 32 
 
 measurement of by sensitometer, 33 
 by spectrograph, 34 
 use of dyes to produce, 32 
 Contrast, 15, 16 
 
 actual, of subject in haze cabinet, 55 
 definition of, 17 
 
 depending on difference in density, 59 
 ground (exposure ratio), 54 
 relation to haze, 61, 63, 66 
 relation to speed of photographic emulsion, 66 
 variation with source of reflected light, 44, 64, 66 
 Curves, characteristic of photographic emulsions, 19, 20, 22, 23, 25, 41, 43, 
 45, 46, 54, 60, 66 
 haze-altitude, 44, 48, 49, 78 
 haze-wave-length, 52 
 spectral distribution, 27 
 
 Density, of photographic plate, 29 
 effect of, on contrast, 59 
 formula for, 29 
 measurement of, 19, 30 
 measuring instruments, 19, 30 
 Detail in aerial photography, 16 
 Developer, characteristics of, 70 
 
 effect of bromide on, 22 
 formula for, 70 
 reduction potential of, 23 
 
 81 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 Development, 16, 21, 29, 70 
 
 uniform for all densities, 23 
 Diffusing media in haze cabinet, 57 
 Distribution of haze at different altitudes, 49, 72, 76 
 
 -wave-lengths, 67 
 
 with season, 11, 13, 39 
 Dyes to produce color sensitiveness, 32 
 carbocyanine, 32 
 cyanine, 32 
 isocyanine, 32 
 fluoresceine, 32 
 quinoline, 32 
 
 Effect, absorption, 50 
 
 of atmospheric haze, 12, 32 
 
 — bromide on developer 
 
 — colored light on photographs, 16, 17 
 
 — density on contrast, 59 
 
 — source of reflected light on contrast, 44 
 Emulsion speed, relation to intensity of light, 67 
 Exposure, 16, 17, 18, 19, 21, 25, 28, 40, 44 
 
 for panchromatic plates, 70 
 measurement of, 27, 61 
 over-exposure, 20, 64, 65 
 spectral quality of, 53 
 standard for aerial photographs, 66 
 under-exposure, 20, 64, 65 
 Exposure ratio, 45, 47, 50, 54 
 
 Exposure time, relation to preparation of sensitometric tablets, 43 
 
 Filters, absorption curve of, 69 
 
 applicable to aerial photography, 68, 70 
 characteristic curves of, 19, 23, 34, 35, 42, 43 
 speed effective with, 35 
 transmission of, 40, 51, 52 
 Filter factors, 34, 40, 45, 68 
 Filter ratio, 34, 35 
 Formula, for developer, 70 
 for haze effect, 37 
 
 Gamma, effect on opacity ratio, 60, 63, 66 
 Gelatin wedge, calibration of, 57 
 
 use of, in controlling intensity of veiling light, 54, 56 
 
 Haze, atmospheric, 12, 36 
 
 absorption effect of, 50 
 
 content variation with location, 14 
 
 definition of, 36 
 
 distribution of, at different altitudes, 49, 72, 76 
 
 -wave-lengths, 67 
 
 seasonal, 11, 13, 39 
 relation to contrast, 61, 63 
 Haze-altitude variation curves, 44, 48, 49, 78 
 Haze brightness, 61 
 
 82 
 
AERIAL HAZE 
 
 Haze cabinet, 54 
 
 construction of, 54 
 diffusing media in, 57 
 illuminometer used in, 57, 58 
 light source for, 55 
 measurements made with, 63 
 Haze effect, 11, 12, 14, 16 
 
 conditions for minimizing, 16, 63 
 formula for, 37 
 produced by air, 72 
 Haze glare, effect on opacity ratio, 60 
 Haze—wave-length variation, 5 
 
 Hurter and Driffield curves (see characteristic curves) 
 
 Illuminometer, use of in haze cabinet, 57, 58 
 Intensity scale, for field use, 41 
 
 -sensitometer, 34 
 
 of light, relation to speed of emulsion, 67 
 Intermittent sensitometers, 27 
 
 Light, effect of colored on photographs, 16, 17 
 intensity values of, 44 
 quality variations, 25 
 source of in haze cabinet, 55 
 Martens’ photometer, 30, 47 
 
 Measurement of color sensitiveness by wedge spectrograph, 33, 34 
 Measurements, accuracy of with film and plates, 21 
 photometric, 31 
 sensitivity, 18, 27 
 variation of with intensity, 23, 25 
 
 --with light source, 21, 25 
 
 -with material, 21 
 
 —--with treatment, 21, 22 
 
 with haze cabinet, 63 
 Microphotometers, use of, 31 
 Mist or fog, 14, 15 
 
 Non-intermittent sensitometers, 28, 40, 41 
 
 Opacity ratio, 59 
 
 effect of haze glare on, 60 
 
 -unity on, 60, 62, 66 
 
 Orthochromatic plates, 32 
 
 Panchromatic plates, 21, 32, 68 
 Photographic emulsions, 31 
 density of, 29 
 deviation found in, 21, 22 
 effect of ammonia on, 33 
 
 -dyes on, 32, 33 
 
 orthochromatic, 32 
 panchromatic, 21, 32, 68 
 relation to contrast, 66 
 sensitizing, 32, 33 
 sensitometric characteristics of, 59 
 speed of, 68 
 
 suitable for aerial photography, 59, 64, 66, 68 
 
 83 
 
MONOGRAPHS ON THE THEORY OF PHOTOGRAPHY 
 
 Photometers, 31 
 
 Martens’, 30, 47 
 polarization, 30 
 Polarization photometer, 30 
 
 Radiation, 14, 72, 74 
 
 equation for, 73 
 
 Reduction potential of developer, 23 
 Reflected light, color factor of, 44 
 
 effect on contrast, 44 
 Reflecting power of test objects, 58 
 
 Sensitivity of photographic emulsions, 21, 32 
 Sensitizing plates with ammonia, 32, 33 
 Sensitometers, characteristics of, 59 
 intensity scale of, 34 
 intermittent, 27 
 non-intermittent, 28 
 
 Sensitometric characteristics of photographic plates, 59 
 data, 18 
 tests, 18 
 
 Sources of haze, 36 
 Speed, effective, of emulsions, 67, 68 
 effective with filters, 35 
 Spectral distribution curve, 27 
 Spectral quality of exposure, 53 
 Spectrophotometers, 31 
 Standard sensitometric methods, 17 
 
 Tablets, sensitometric, 25, 58 
 
 exposure time used in the preparation of, 43 
 Test objects, log exposuze of, 42 
 
 log transmission of, 42, 47 
 material for, 58 
 reflecting power of, 58 
 Transmission, fractional of dry air, 77 
 of filters, 40, 52 
 of photometric tablets, 29, 42 
 
 Under-exposure, 20, 64, 65 
 
 Veiling haze, 11, 17 
 Veiling light, 54 
 
 Wedge spectrograph for measuring color sensitiveness, 34 
 
 84