"In a radiograph of the body," he continued, "the difference of brightness that distinguishes the heart from the lungs, bones from flesh, is due to the different densities of tissues. In these pictures the same effect is produced by the different densities of the pigments, especially of their principal and heaviest elements."

He paused and laid down a chart. "For anyone who doubts what I am about to prove, I have made a scale of oil colors arranged in accordance to their transparency to Roentgen rays by applying standard pigments to canvas in patches of equal thickness.

"I think you can see what I am driving at. For instance, a design drawn in a heavy pigment will show through a layer of a less dense pigment, under the influence of the X-ray—just as bones show through flesh. In other words, an ordinary photograph reproduces only the surface of a painting. A radiograph represents all the pigments underneath, also producing effects in proportion to their densities.

"Let me show you the practical result of all this in studying such radiographs, as worked out by a German student. I have made several very interesting and conclusive discoveries which these radiographs I have taken illustrate."

He paused a moment, for the sake of emphasis. "You will notice," he resumed carefully, "the lace frill above the bodice on the figure of Madame de Montespan, in this radiograph. In the painting the frill is sharply defined and can be clearly distinguished from the bodice. But look at this radiograph. It appears tattered. It overflows the bodice.

"That led me to suspect that the bodice was widened as an afterthought—perhaps to diminish the area of white. That is the reason why the white shows through the bodice in the radiograph. But in this other one the bodice and the frill are substantially as they must be in the original."

Again he paused, as if taking up a new point. "This radiograph,—number one, I may call it—shows a broad light band on the right hand of the figure, of which not a trace is to be found either in the other radiograph or the painting itself. It represents the first rough sketch of an arm and hand.

"Again, in this first radiograph the ring and little fingers are close together and a sixth finger appears between the index and middle fingers. From that I infer that the hand hung limp with the fingers nearly in contact in the first sketch and that the fingers were afterward separated. But in this second radiograph the arm, hand and fingers are perfect."

It was fascinating to listen to Kennedy as he delved down into the invisible beneath the very oils and dug out their hidden mystery.

"Take the head and shoulder," he continued. "Radiograph number one clearly shows flaking of the painting which has been painted over to conceal it. Ordinary light reveals no trace, either, of a long crack on the shoulder which evidently was filled with a thick mass of pigment containing too little white lead to obliterate the crack in the radiograph. White spots above the ear, in the radiograph, probably indicate an excess of white lead used in retouching. At any rate, radiograph number two contains no such defects."