Читать книгу Airplane Photography - Herbert Eugene Ives - Страница 8

Table of Contents

General Considerations.—The design and selection of lenses for aerial photography present on the whole no problems not already encountered in photography of the more familiar sort. Indeed, the lens problem in the airplane camera is in some particulars more simple than in the ground camera. For instance, there is no demand for depth of focus—all objects photographed are well beyond the usually assumed “infinity focus” of 2000 times the lens diameter. Such strictly scientific problems of design as pertain to aerial photographic lenses are ones of degree rather than of kind. Larger aperture, greater covering power, smaller distortion, more exquisite definition—these always will be in demand, and each progressive improvement will be reflected in advances in the art of aerial photography. But many lens designs perfected before the war were admirably suited, without any change at all, for aerial cameras.

Of the utmost seriousness, however, with the Allies, was the problem of securing lenses of the desired types in sufficient numbers. The manufacture of the many varieties of optical glass essential to modern photographic lenses was almost exclusively a German industry, which had to be learned and inaugurated in Allied countries since 1914. In consequence of this entirely practical problem of quantity production without the glasses for which lens formulæ were at hand, some new lens designs were produced. Whether any of these possess merits which will lead them to be preferred over pre-war designs, when the latter can again be manufactured, remains to be seen.

While the glass problem was still unsolved, aerial cameras had to be equipped with whatever lenses could be secured by requisition from pre-war importation and manufacture, and later, with lenses designed to utilize those glasses whose manufacture had been mastered in the allied countries. It is important that the historical aspect of this matter be well understood by the student of aerial photographic methods, for the use of these odd-lot lenses reacted on the whole design of aerial cameras and on the methods of aerial photography, particularly in England and the United States. Almost without exception the available lenses were of short focus, considered from the aerial photographic standpoint; that is, they lay between eight and twelve inches. This set a limit to the size of the airplane camera, quite irrespective of the demands made by the nature of the photographic problem. Lenses of these focal lengths produced images which, for the usual heights of flying, were generally considered too small, and which were, therefore, almost always subsequently enlarged. Such was the English practice, which was followed in the training of aerial photographers in America, where exactly similar conditions held at the start with respect to available lenses. French glass and lens manufacturers did succeed in supplying lenses of longer focus (50 centimeters), in numbers sufficient for their own service, although never with any certainty for their allies. The French, therefore, almost from the start, built their cameras with lenses of long focus, and made contact prints from their negatives.

Practices adopted under pressure of an emergency to meet temporary practical limitations often come to dominate the whole situation. This is particularly true of aerial photography in the British and American services. The small apparatus built around the stop-gap short focus lenses fixed the plane designer's idea of an airplane camera, and the space it should occupy. This was directly reflected in the designs of the English planes, and the American planes copied after them. Meanwhile the American photographic service in France associated itself with the French service, adopting its methods and apparatus, and using French planes whose designs were not being followed in American construction. The task of harmonizing the photographic practice as taught in America, following English lines, with French practice as followed in the theater of war, and of adapting planes built on English designs so that they could carry French apparatus, was a formidable one, not likely to be soon forgotten by any who had a part in it.

Photographic Lens Characteristics.—Whole volumes have been written on the photographic lens, and on the optical science utilized and indeed brought into being by its problems. Such works should be consulted by those who intend to make a serious study of the design of lenses for aerial use. No more can be attempted, no more indeed is relevant here, than an outline review of the chief characteristics and errors of photographic lenses, considering them with special reference to aerial needs.

The modern photographic lens is, broadly speaking, a development of the simple convex or converging lens. Its function is the same: to form a real image of objects placed before it. But the difference in performance between the simple lens and the modern photographic objective is enormous. The simple lens forms a clear image only close to its axis, for light of a single color, and as long as its aperture is kept quite small as compared to the distance at which the image is formed. The photographic lens, on the other hand, is called upon to produce a clear image with light of a wide range of spectral composition, sharply defined over a flat surface of large area, and it must do this with an aperture that is large in comparison with the focal length, whereby the amount of light falling on the image surface shall be a maximum. This ideal is approximated to a really extraordinary degree by the scientific combination and arrangement of lens elements made from special kinds of glass in the best photographic lenses of the anastigmat type. The result is of necessity a set of compromises, whereby the outstanding errors are reduced to a size judged permissible in view of the work the lens is to do. These errors or aberrations are briefly reviewed below, in order that the reader may readily grasp the terms in which the performance and tolerances in aerial lenses are described.

Fig. 12.—Diagrammatic representation of spherical aberration.

Spherical Aberration and Coma.—Suppose we focus on a screen, by means of a simple convex lens the image of a distant point of light. Suppose for simplicity that this image is located on the axis of the lens and that light of only one color is used, such as yellow. It will be found that the smallest image that can be obtained is not a point, but a small disc. This is due to the fact that the rays of light passing through the outer portions of the lens are bent more than those passing through the lens in the region near the center. This effect is shown in Fig. 12 by the usual mode of representing it graphically. Here the figures 1, 2, 3, 4, represent distances from the axis of the lens, and the letters A₁, A₂, A₃, A₄, the points of convergence of the rays from 1, 2, 3, 4, etc. These distances projected upward on to the produced lens points form a curve which shows at a glance the extent and direction of the error due to each part of the lens. This information is of value where the lens is fitted with an adjustable diafram. With some types of correction sharper definition may be obtained by reducing the aperture. With others, however, diaframing impairs definition, by destroying the balance between under and over correction which averages to make a good image. In aerial lenses it is not customary to use diaframs, as all the light possible is desired. Consequently the reduction of spherical aberration must be accomplished by proper choice of lens elements and their arrangement.

Off the axis of the lens the image of a point source takes on an irregular shape, due to oblique spherical aberration or coma.

Chromatic Aberration.—Because of the inherent properties of the glass of which it is made, a simple collective lens does not behave in the same way with respect to light of different colors. If one attempts, with such a lens, to focus upon a screen the image of a distant white light, it will be found that the blue rays will not focus at the same point as the red rays, but will come together nearer the lens. Modern photographic lenses are compounded of two or more kinds of glass in such a way as to largely eliminate this defect, the presence of which is detrimental to good definition. Such lenses are called achromatic, and the property of a lens by virtue of which this defect is eliminated is called its chromatic correction.

Chromatic correction is never perfect, but two colors of the spectrum can be brought to a focus in the same plane, and to a certain extent the departure of other colors from this plane can be controlled. Off the axis of the lens outstanding chromatic aberration results in a difference in the size of images of different colors, known as lateral chromatism.

Like spherical aberration, chromatic aberration is a contributing factor to the size of the image of a point source, which determines the defining power of a lens. It is, however, an error whose effect is to some extent dependent on the kind of sensitive plate used. Two lenses may give images of the same size (in so far as it is governed by chromatic aberration), if a plate of narrow spectral sensitiveness is used, while giving images of different size on panchromatic plates of more extended color sensibility. The choice of the region of the spectrum for which chromatic correction is to be made is thus governed by the color of the photographically effective light. While in ordinary photography the blue of the spectrum is most important, in aerial work where color filters are habitually used with isochromatic plates the green is most important, and color correction centered about this region constitutes a real difference of design peculiar to aerial lenses. Similarly the general use of deep orange or red filters with red sensitive plates, for heavy mist penetration, would call for a shift of correction to that part of the spectrum.

Astigmatism and Covering Power.—Suppose the lens forms at some point off its axis an image of a cross. Suppose one of the elements of the cross to be on a radius from the center of the field, the other element parallel to a tangent. The rays forming the images of these two elements of the cross are subject to somewhat different treatment in their passage through the lens. The curvature of the lens surfaces is on the whole greater with respect to the rays from the radial element than to those from the tangential element. They are therefore refracted more strongly and come to a focus nearer the lens. The arms of the cross are consequently not all in focus at once. This error, termed astigmatism, is rather well shown in Fig. 15, where the images of the outlying concentric circles are sharp in the radial, but blurred in the tangential direction.

Astigmatism can be largely compensated for, and its character controlled. The most usual correction brings the two images in focus together both at the axis, and on a circle at some distance out. This second locus of coincidence may or may not be in the same plane as the first, depending on which disposition produces the best average correction. The mean between the two foci determines the focal plane of the lens, which is in general somewhat curved. The covering power of a lens is given by the size of the field which is sufficiently flat and free from astigmatism for the purpose for which the lens is used. This is largely determined by the astigmatism, but the other aberrations are also important.

Illumination.—The amount of light concentrated by the lens on each elementary area of the image determines its brightness or illumination. The ideal image would, of course, be equally bright over its whole area of good definition, and for lenses of narrow angle this is approximately true. But when it is desired to cover a wide angle the question of illumination becomes serious. The relationship between angle from the axis and illumination is that illumination is proportional to the fourth power of the cosine of the angle. This relationship is shown in the following table:

Angle	Image brightness
0°	100 per cent.
10°	94.1 per cent.
20°	78.0 per cent.
30°	56.2 per cent.
40°	34.4 per cent.
50°	17.1 per cent.

If the field of view is 60°, which corresponds to an 18 × 24 centimeter plate with a lens of 25 centimeter focus, the brightness is only 56 per cent., and the necessary exposure at the edge approximately 1.8 times that at the center. This effect is shown in Fig. 15. It is very noticeable if the exposure is so short as to place the outlying areas in the under-exposure period.

Fig. 13.—Barrel and pin-cushion distortion.

Distortion.—Sometimes a lens is relatively free from all the aberrations, mentioned above, so that it gives sharp, clear images on the plate, yet these images may not be exactly similar to the objects themselves as regards their geometrical proportions; in other words, the image will show distortion. Lens distortion assumes two typical forms, illustrated in Fig. 13, which shows the result of photographing a square net-work with lenses suffering in the one case from “barrel” distortion and in the other from “pin-cushion” distortion. In the first the corners are drawn in relative to the sides; in the latter case the sides are drawn in with respect to the corners. Either sort is a serious matter in precision photography, such as aerial photographic mapping aspires to become. It must be reduced to a minimum and its amount must be accurately known if negatives are to be measured for the precise location of photographed objects. In general symmetrical lenses give less distortion than the unsymmetrical (Fig. 14).

Fig. 14.—Arrangement of elements in two lenses suitable for aerial work: a, Zeiss Tessar; two simple and one cemented components (unsymmetrical); b, Hawkeye Aerial; two positive elements of heavy barium crown, two negative of barium flint, uncemented (symmetrical).

Lens Testing and Tolerances for Aerial Work.—Simple and rapid comparative tests of lenses may be made by photographing a test chart, consisting of a large flat surface on which are drawn various combinations of geometrical figures—lines, squares, circles, etc.—calculated to show up any failures of defining power. For testing aerial lenses the chart should be as large as possible, so that it may be photographed at a distance great enough for the performance of the lens to be truly representative of its behavior on an object at infinite distance. This means in practice a chart of 4 or 5 meters side, to be photographed at a distance 20 to 30 times the focal length of the lens.

Fig. 15.—Photograph of a lens testing chart, showing failure in defining power outside area for which the lens is calculated.

A typical photograph of such a chart is shown in Fig. 15. It reveals at a glance the more conspicuous lens errors. At the sides and corners the concentric circles show the lens's astigmatism, by the clear definition of the lines radial to the center of the field and their blurring in the tangential direction. The falling off in illumination with increasing distance from the center is also exhibited; and the blurring of all detail outside the rectangle for which the lens was calculated shows that spherical, chromatic, and other aberrations have become prohibitively large.

But the only complete test of a lens is the quantitative measurement of errors made on an optical bench. A point source of light, which may at will be made of any color of the spectrum, is used as the object and its image formed by the lens in a position where it can be accurately measured for location, size, and shape by a microscope. A chart giving the results of such a test is shown in Fig. 16. In the upper left-hand corner is shown the position of the focus for the different colors of the spectrum. Below this is recorded the lateral chromatism at 21 degrees, in terms of the difference in focus for a red and a blue ray. Below this again comes the distortion, or shift of the image from its proper position, for various angles (plotted at the extreme right) from the lens axis. To the right of this is the image size, at each angle, and finally, to the right of the diagram, are plotted the distances of the two astigmatic foci from the focal plane, together with the mean of the two foci, which practically determines the shape of the field.

An important point to notice is that these data are uniformly plotted in terms of a lens of 100 millimeters focal length irrespective of the actual focal length of the lens measured. Thus this particular chart is for a 50 centimeter lens but would be plotted on the same scale for a 25 or a 100 centimeter lens. Underlying this practice is the assumption that all the characteristics of lenses of the same design and aperture are directly proportional to their focal length. If this were so, then a 50 centimeter lens would give double the size of image that a 25 centimeter does, and so on. As a matter of fact, test shows that the size of the image does not increase so rapidly as the focal length; so that while the image size for a 25 centimeter lens would be, say, .05 millimeters per 100 millimeters focal length, it will be only .03 or .04 millimeters per 100 millimeters focal length for a 50 centimeter lens. The actual size of a point image will therefore be greater, though not proportionately greater.

Fig. 16.—Chart recording measurements of lens characteristics.

The chart presents tests on a good quality lens, and so gives a good idea of the permissible magnitude of the various errors. In many ways the most important figure is that for image size, including as it does the result of all the aberrations. In the example given, this varies from .075 to .15 mm. actual size. For the same type of lens of 25 centimeters focus this range will be from .05 to .10 mm. Since these are commonly used focal lengths, a good average figure for image size, commonly used in aerial photographic calculations, is ⅒ mm. In regard to astigmatic tolerances, the two astigmatic foci should not be separated at any point by more than 6 to 7 millimeters, and the mean of these should not deviate from the true flat field by more than ½ millimeter, in each case the figures being based on the conventional 100 millimeters focal length. Distortion should not be over .08 millimeter at 18° or .20 millimeter at 24° from the axis (per 100 millimeters focal length).

Lens Aperture.—In the simple lens the aperture is merely the diameter. In compound lenses the aperture is not the linear opening but the effective opening of an internal diafram. Photographically, however, aperture has come to have a more extensive meaning. While in the telescope the actual diameter of an objective is perhaps the most important figure, and in the microscope the focal length, in photography the really important feature is the amount of light or illumination. This is determined by lens opening and focal length together; specifically, by the ratio of the lens area to the focal length. The common system of representing photographic lens aperture is by the ratio of focal length to lens diameter, the lens being assumed to be circular. Thus F/5 (often written F.5) indicates that the diameter is one-fifth the focal length.

Two points are to be constantly borne in mind in connection with this system of representation. First, all lenses of the same aperture (as so represented) give the same illumination of the plate (except for differences due to loss of light by absorption and reflection in the lens system). This follows simply from the fact that the illumination of the plate is directly proportional to the square of the lens diameter, and inversely as the square of the focal length. Secondly, the illumination of the plate is inversely as the square of the numerical part of the expression for aperture. That is, lenses of aperture F/4.5 and F/6 give images of relative brightness (6 4.5)² = 1.78.

What lens aperture, and therefore what image brightness, is feasible, is determined chiefly by the angular field that must be covered with any given excellence of definition. The largest aperture ordinarily used for work requiring good definition and flat field free from distortion is F/4.5. Anastigmatic lenses of this aperture cover an angle of 16° to 18° from the axis satisfactorily, which corresponds to an 18 × 24 centimeter plate with a lens of 50 centimeters focus. Lenses with aperture as large as F/3.5 were used to some extent in German hand cameras of 25 centimeters focal length, with plates of 9 × 12 centimeters. English and American lenses of this latter focal length were commonly of aperture F/4.5, designed to cover a 4 × 5 inch plate.

As a general rule the greater the focal length the smaller the aperture—a relationship primarily due to the difficulty of securing optical glass in large pieces. Thus while 50 centimeter lenses of aperture F/4.5 are reasonably easy to manufacture, the practicable aperture for quantity production is F/6, and for 120 centimeter lenses, F/10. This means that a very great sacrifice of illumination must be faced to secure these greater focal lengths. As is to be expected from the state of the optical glass industry, the German lenses were of generally larger aperture for the same focal lengths than were those of the Allies. Besides the F/3.5 lenses already mentioned, their 50 centimeter lenses were commonly of aperture F/4.8, their 120 centimeter lenses of aperture F/7, or of about double the illuminating power of the French lenses of the same size.

Demands for large covering power also result in smaller aperture. The 26 centimeter lenses used on French hand cameras utilizing 13 × 18 centimeter plates were commonly of aperture F/6 or F/5.6. The lens of largest covering power decided on for use in the American service was of 12 inch focus, to be used with an 18 × 24 centimeter plate (extreme angle 26°); the largest satisfactory aperture for this lens is F/5.6.

Ordinarily the question of aperture is closely connected with that of diaframs, whereby the lens aperture may be reduced at will. Diaframs have been very little used in aerial photography. All the aperture that can be obtained and more is needed to secure adequate photographic action with the short exposures required under the conditions of rapid motion and vibration peculiar to the airplane. Any excess of light, over the minimum necessary to secure proper photographic action, is far better offset by increase of shutter speed or by introduction of a color filter. For this reason American aerial lenses were made without diaframs. In the German cameras, however, adjustable diaframs are provided (Fig. 43), controlled from the top of the camera by a rack and pinion. In the camera most used in the Italian service an adjustable diafram is provided, but this is occasioned by the employment of a between-the-lens shutter of fixed speed, so that the only way exposure can be regulated is by aperture variation, a method which has little to recommend it.

The Question of Focal Length.—In aerial photography the lens is invariably used at fixed, infinity, focus. Under these conditions the simple relationship holds that the size of the image is directly proportional to the focal length and inversely proportional to the altitude. If any chosen scale is desired for the picture the choice of focal length is determined by the height at which it is necessary to fly. This at least would be the case were there no limitation to the practicable focal length—which means camera size—and were one limited to the original size of the picture as taken; that is, were the process of enlargement not available. But the possibility of using the enlarging process brings in other questions: Is the defining power of a short focus lens as good in proportion to its focal length as that of a long focus lens? If so a perfect enlargement from a negative made by a short focus lens would be identical with a contact print from a negative made with a lens of longer focus. Is defining power lost in the enlarging process with its necessary employment of a lens which has its own errors of definition and which must be accurately focussed?

Certain factors which enter into comparisons of this sort in other lines of work, such as astronomical photography, play little part here. These are, first, the optical resolving power of the lens, which is conditioned by the phenomena of diffraction, and is directly as the diameter; and, second, the size of the grain of the plate emulsion. The first of these does not enter directly, because the size of a point image on the axis of the lens, due merely to diffraction, is very much less than that given by any photographic lens which has been calculated to give definition over a large field, instead of the minute field of the telescope. Yet it may contribute toward somewhat better definition with a long focus lens because of the actually larger diameter of such lenses. The second factor is not important, because, as will be seen later, the resolving power of the plates suitable for aerial photography is considerably greater than that of the lens. The emulsion grain is in fact only a quarter or a fifth the size of the image as given by a 25 centimeter lens, and enlargements of more than two or three times are rarely wanted.

A series of experiments was made for the U. S. Air Service to test out these questions, using a number of representative lenses of all focal lengths, both at their working apertures and at identical apertures for all. With regard to lens defining power, as shown by the size of a point image, the answer has already been reported in a previous section. Lenses of long focus give a relatively smaller image than lenses of the same design of short focus. In regard to the whole process of making a small negative and enlarging it, the loss of definition is quite marked, as compared to the pictures of the same scale made by contact printing from negatives taken with longer focus lenses.

This answer is clear-cut only for lenses calculated to give the same angular field. Thus a 10 inch lens covering a 4 × 5 inch plate has about the same angle as a 50 centimeter lens for an 18 × 24 centimeter plate. When, however, it comes to the longer foci, such as 120 centimeters, the practical limitation to plate size (18 × 24 cm.) has been passed, and the angular field is less than half that of the 50 centimeter lens. The 120 centimeter lens need only be designed for this small angle, with consequent greater opportunities for reduction of spherical aberration. It is therefore an open question whether a 50 centimeter lens designed to cover a plate of linear dimensions 50

120 times that used with the regular 50 centimeter lens could not be produced of such quality that it would yield enlargements equal to contacts from a 120 centimeter lens. If so, lenses of larger aperture could be used, and a considerable saving in space requirements effected.

Focal lengths during the Great War were decided by the nature of the military detail which was to be revealed and by the altitudes to which flying was restricted in military operations. In the first three years of the war the development of defences against aircraft forced planes to mount steadily higher, so that the original three or four thousand feet were pushed to 15,000, 18,000, and even higher. Lenses of long focus were in demand, leading ultimately to the use of some of as much as 120 centimeters (Fig. 41). In the last months of the war the resumption of open fighting made minute recording of trench details of less weight, while the preponderance of allied air strength permitted lower flying. In consequence, lenses of shorter focus and wider angle came to the fore, suitable for quick reconnaissance of the main features of new country. At the close of the war the following focal lengths were standard in the U. S. Air Service, and may be considered as well-suited for military needs. Peace may develop quite different requirements.

Focal length	Aperture	Plate size
10 inch	F/4.5	4 × 5 inch
26 cm.	F/6	13 × 18 cm.
12 inch	F/5.6	18 × 24 cm.
20 inch	F/6.3 to F/4.5	18 × 24 cm.
48 inch	F/10 to F/8	18 × 24 cm.

The question of the use of telephoto lenses in place of lenses of long focus is frequently raised. Lenses of this type combine a diverging (concave) element with the normal converging system, whereby the effect of a long focus is secured without an equivalent lens-to-plate distance. This reduction in “back focus” may be from a quarter to a half. Were it possible to obtain the same definition with telephoto lenses as with lenses of the same equivalent focus, they would indeed be eminently suitable for aerial work because of their economy of length. But experience thus far has shown that the performance of telephoto lenses, as to definition and freedom from distortion, is distinctly inferior, so that it is best to hold to the long focus lens of the ordinary type.

Lenses Suitable for Aerial Photography.—Among the very large number of modern anastigmat lenses many were found suitable for airplane cameras and were used extensively in the war. A partial list follows: The Cooke Aviar, The Carl Zeiss Tessar, the Goerz Dogmar, the Hawkeye Aerial, the Bausch and Lomb Series Ic and IIb Tessars, the Aldis Triplet, the Berthiot Olor.

The Question of Plate Size and Shape.—Plate size is determined by a number of considerations, scientific and practical. If the type of lens is fixed by requirements as to definition, then the dimensions of the plate are limited by the covering power. From the standpoint of economy of flights and of ease of recognizing the locality represented in a negative, by its inclusion of known points, lenses of as wide angle as possible should be used. If the focus is long, this means large plates, which are bulky and heavy. If the finest rendering of detail is not required a smaller scale may be employed, utilizing short focus lenses and correspondingly smaller plates. Thus a six inch focus lens on a 4 × 5 inch plate would be as good from the standpoint of angular field as a 12 inch on an 8 × 10 inch plate. This is apt to be the condition with respect to most peace-time aerial photography, which may be expected to free itself quickly from the huge plates and cameras of war origin.

For work in which great freedom from distortion of any sort is imperative, small plates will be necessary, for two reasons. One is that the characteristic lens distortions are largely confined to the outlying portions of the field. The other is that a wide angle of view inevitably means that all objects of any elevation at the edge of the picture are shown partly in face as well as in plan, which prevents satisfactory joining of successive views (Fig. 128). In making a mosaic map of a city, if a wide angle lens is employed with large plates, the buildings lying along the junctions of the prints can be matched up only for one level. If this is the ground level, as it would be to keep the scale of the map correct, the roofs will have to be sacrificed. In extreme cases a house at the edge of a junction may even show merely as a front and rear, with no roof, while in any case the abrupt change at these edges from seeing one side of all objects to seeing the opposite side is not pleasing.

The table in a preceding section gives the relation of plate size to focal length found best on the whole for military needs. Deviations from these proportions in both directions are met with. In the English service the LB camera, which uses 4 × 5 inch plates, is equipped with lenses of various focal lengths, up to 20 inches. The German practice, as well as the Italian, was almost uniform use of 13 × 18 centimeter plates for all focal lengths. Toward the end of the war, however, some German cameras of 50 centimeter focal length were in use employing plates 24 × 30 centimeters.

It will be recognized that these plate sizes are chosen from those in common use before the war. A similar observation holds with even greater force on the question of plate shape. Current plate shapes have been chosen chiefly with reference to securing pleasing or artistic effects with the common types of pictures taken on the ground. These shapes are not necessarily the best for aerial photography. Indeed the whole question of plate shape should be taken up from the beginning, with direct reference to the problems of aerial photography and photographic apparatus.

A few illustrations will make this clear, taking Fig. 17 as a basis. If it is desired to do spotting (the photography of single objectives), the best plate shape would be circular, for that shape utilizes the entire covering area of the lens. If it is desired to make successive overlapping pictures, either for mapping, or for the production of stereoscopic pairs, a rectangular shape is indicated. If the process of plate changing is difficult or slow, it is advisable, in order to give maximum time for this operation, to have the long side of the rectangle parallel to the line of flight (indicated by the arrow). If economy of flights is a consideration, as in making a mosaic map of a large area, it is advantageous to have as wide a plate as the covering power of the lens will permit. Reference to Fig. 17 shows that this means a plate of small dimensions in the direction of flight. If the changing of plates or film is quick and easy, the maximum use of the lens's covering power is made by such a rectangle whose long side approximates the dimensions of the lens field diameter. This is in fact the choice made in the German film mapping camera (Figs. 61 and 63), whose picture is 6 × 24 centimeters. An objection to this from the pictorial side, lies in the many junction lines cutting up the mosaic. Another objection, if the plane does not hold a steady course, is the failure to make overlaps on a turn. (Fig. 62.) Here as everywhere the problem is to decide on the most practical compromise between all requirements.