“Notice.  It is seen very prominently to the right, and its exact location on the house is an easy matter.  You could almost use the metric photograph for that.  The identification of the gable casting the shadow is easy.  To be exact it is 19.62 feet high.  The shadow is 14.23 feet down, 13.10 feet east, and 3.43 feet north.  You see I am exact.  I have to be.  In one minute it moved 0.080 feet upward, 0.053 feet to the right and 0.096 feet in its apparent path.  It passes the width of a weatherboard, 0.37 foot, in four minutes and thirty-seven seconds.”

Kennedy was talking rapidly of data which he had derived from his metric photograph, from plumb line, level, compass, and tape, astronomical triangle, vertices, zenith, pole and sun, declination, azimuth, solar time, parallactic angles, refraction, and a dozen bewildering terms.

“In spherical trigonometry,” he concluded, “to solve the problem three elements must be known.  I knew four.  Therefore I could take each of the known, treat it as unknown, and have four ways to check my result.  I find that the time might have been either three o’clock, twenty-one minutes and twelve seconds, in the afternoon, or 3:21:31, or 3:21:29, or 3:21:33.  The average is 3:21:26, and there can therefore be no appreciable error except for a few seconds.  For that date must have been one of two days, either May 22 or July 22.  Between these two dates we must decide on evidence other than the shadow.  It must have been in May, as the immature condition of the foliage shows.  But even if it had been in July, that is far from being September.  The matter of the year I have also settled.  Weather conditions, I find, were favourable on all these dates except that in September.  I can really answer, with an assurance and accuracy superior to that of the photographer himself—­even if he were honest—­as to the real date.  The real picture, aside from being doctored, was actually taken last May.  Science is not fallible, but exact in this matter.”

Kennedy had scored a palpable hit.  McLoughlin and Hanford were speechless.  Still Craig hurried on.

“But, you may ask, how about the automobile picture?  That also is an unblushing fake.  Of course I must prove that.  In the first place, you know that the general public has come to recognise the distortion of a photograph as denoting speed.  A picture of a car in a race that doesn’t lean is rejected—­people demand to see speed, speed, more speed even in pictures.  Distortion does indeed show speed, but that, too, can be faked.

“Hanford knows that the image is projected upside down by the lens on the plate, and that the bottom of the picture is taken before the top.  The camera mechanism admits light, which makes the picture, in the manner of a roller blind curtain.  The slit travels from the top to the bottom and the image on the plate being projected upside down, the bottom of the object appears on the top of the plate.  For instance, the wheels are taken before the head of the driver.  If the car is moving quickly the image moves on the plate and each successive part is taken a little in advance of the last.  The whole leans forward.  By widening the slit and slowing the speed of the shutter, there is more distortion.