466 OILS, ETC., USED IN THE PAINTING OF PICTURES. 



the painting surface. The priming had not cracked, but seemed to 

 be drawn out under the crack. 



The second picture was painted on a closely woven canvas which 

 had been lightly sized and thinly primed by the artist with a mixture 

 of joigments and linseed oil. The cracking was confined to the white 

 masses of hard dry paint in the sky. A section showed that these 

 cracks also formed broad cracks with perpendicular edges without 

 injury to the undercoating of paint. The undercoating of paint 

 did not, however, seem to be properly attached to the canvas, a layer 

 of spongy, disintegrated material apparently having formed in some 

 way. This condition of things prevailed all over the canvas and 

 suggested that, owing to some disintegration between the canvas 

 and the lower coats of paint, the coats of paint had been stretched, 

 resulting in cracking of the upper coat where it was not elastic 

 enough to yield. In order to get some light on the possible causes 

 of cracking, I determined to measure the actual movements taking 

 place in the canvas itself under different conditions. For this pur- 

 pose I attached a strip of sized canvas by one end to a glass cylin- 

 der and weighted the other end so as to keep it taut over the cylinder. 

 A platinum wire was firmly sewn to the canvas across and project- 

 ing out each side, and two little glass rods attached to the glass as 

 indicators. The length of the canvas from the attached end to the 

 platinum wire was 2 centimeters. By measuring the distance 

 between the glass rod and the platinum wire it was possible to 

 measure any expansions or contractions of the canvas itself. 



By measuring at both ends and taking a mean any twisting of 

 the canvas was eliminated. On first measuring, the average dis- 

 tance between the glass rod and the platinum wire was 0.45 milli- 

 meter. The cylinder was then inclosed for twenty hours over strong 

 sulphuric acid so as to dry the canvas thoroughl3^ The distance had 

 now increased to 0.55 millimeter, showing a contraction of the canvas 

 on drying. It was now kept for twenty hours in an atmosphere sat- 

 urated with water vapor. 



At the end of this time the readings showed a distance of 0.28 

 millimeter, showing a total expansion from dry to moist air con- 

 ditions of 0.27 millimeter. On again replacing over strong sulphuric 

 acid the distance increased to 0.57 millimeter, showing a change in 

 length of 0.29 millimeter, or, taking the mean, of 0.28 millimeter. 



The canvas was now painted thickly with yellow ocher and put 

 back in the sulphuric acid and allowed to dry. After one day the 

 distance between the points was 0.54 millimeter, and after thirteen 

 days, when the paint w^as fairly dry, it was 0.577 millimeter, showing 

 very slight changes in length during the drying of the paint. After 



