Amateur Photographer, still going strong today, was a journal that Ernest wrote for from time to time. So far I have not done a systematic search to find all his work, so as I find more I shall add to this list.
December 31 1903 – Vertical Lines and Telephotography pp 534-37. The article itself is on pp 534-35, and the illustrations spill over onto the two following pages, embedded in the next piece. The article discusses his empirical finding that a camera back parallel to the subject does not necessarily correct the verticals. The theory of this was later summarised in Lan-Davies’s Telephotography in about 1920.
Vertical Lines and Telephotography.
By Ernest Marriage.
THERE is probably no rule of conduct which is more rigidly regarded by architectural photographers than the need for keeping the back of the camera perpendicular in order to obtain vertical lines in the photograph. Horizontal lines may be tampered with, shown convergent or divergent, but take the same liberty with upright lines, and you are at once set down for an ignoramus, and the folly of your ways is speedily pointed out by the learned. I prefer, in the majority of cases, to keep vertical and horizontal lines correct, and therefore generally endeavour to have the camera back perpendicular and parallel to the subject; yet in spite of precautions, I have sometimes been puzzled with divergent vertical lines in telephotographs, indicating that the camera back has been swung excessively, a method of procedure which all photographers would desire to avoid. Hitherto I have taken it for granted that, somehow, in spite of precautions, the camera back was not perpendicular, but recently it was forced upon my notice that this state of affairs might even be undesirable when the plane of the subject was vertical.
It was at Chartres Cathedral, the camera was pointed to telephotograph the figure high up on the face of a buttress. Fig. 1 is a photograph made with a 6-inch lens, from the same position as the two telephotographs, figs. 4 and 5. The lens was raised 2 1/2 inches from the centre of the plate, and in addition, the camera had to be tilted, but the back was swung to the perpendicular. The vertical lines are reproduced as parallel throughout, including the end of the buttress with the figure in its niche.
Diagrams showing the arrangement of the camera drawn to scale are given in figs. 2 and 3. The camera was mounted on a tilting table, T, whose base was accurately levelled. The telephotographic lens, A B, was made up of A (Dallmeyer 2B patent portrait, 8 inch), and B (high power negative, 2.4 inch), and the distance from the negative lens to the screen was 11 1/2 inches. The lens at this low magnification (about 5 3/4 times) did not quite cover the plate; this is evidenced by the darkened corners of the two telephotographs, but it was not possible to get further away from the subject, and higher magnification would have cut out details whose inclusion was desired on the 7 1/2 by 5 plate. In order to show the relative position of the camera back D, in each diagram, with regard to the perpendicular, the latter is dotted in. In fig. 2 it is less than 6 degrees out of the vertical, and the resulting telephotograph should show slight under-correction; in fig 3 the divergence. (under-correction again) is rather over 26 degrees.
The camera was focussed on the subject, in the first instance, pretty much as outlined in fig. 3, but when I proceeded to the further adjustment, swinging the back to the perpendicular, it was a surprise to see the lines of the image diverging. It will be seen that the camera back, fig. 2, was not absolutely perpendicular — for once this was a task beyond its capabilities — but still it was far more nearly in the orthodox position, and judged by ordinary practice the vertical lines should more nearly approach to the parallel condition than when it was pretty widely out. This did not, however, prove to be the case. However, though over-correction was obvious, the interest of the question was sufficient warrant for an exposure, and fig. 4 is the telephotograph taken with the camera as seen in fig. 2.
The next step was to swing the camera back until the buttress was the same width top and bottom. This was obtained with the camera in the position shown in fig. 3, and the exposure, resulting in the telephotograph, fig. 5, was made.
If the width of the buttress be measured in the two examples, it will be found that fig. 4 diverges markedly towards the top, whereas on general grounds it would be expected to converge slightly, perhaps inappreciably, whilst in fig. 5 the sides, instead of being strongly converged, are parallel.
I have no explanation to offer regarding this apparent exception to a generally recognised law, but in regard to practical telephotography it is evident that the hard and fast rule that the back must be perpendicular is to be honoured more in the breach than in the observance. Apart from the main question dealt with in this article, swinging the camera back is an evil, necessitating the use of a smaller stop and increased exposure. Fig. 4 was made with the lens stopped down to f/11, but for fig. 5 f/8 was used. The fact that the general definition is better with a larger aperture in the latter case, indicates the desirability of keeping the plane of the plate as nearly as possible at right angles to the axis of the lens. A comparison of these telephotographs with the ordinary photograph suggests that the figure is rather diminished in height in the former; more particularly is this seen in fig. 5. If there are any prominent parallel lines in the subject, it would seem better to get them parallel on the screen regardless of the position of the camera back.
The position and arrangement of the buttresses on this side of the Cathedral are also shown in the fig. 6. The subject of the telephotographs is at the top of the third buttress from the right-hand side of the print.