IMPROVED METHOD OF PAINTING CANVAS. 377 



experiment was made, and four hundred weight of dry 

 serviceable paint prepared from one ton of such canvas ; 

 that he had seen it when ground in oil and laid on work, 

 when it appeared to possess all the properties of good 

 paint, and had therefore been recommended by him to the 

 Navy Board. 

 SIR, 



IN answer to your letter of the 25th of April, in which Samples sent t^ 

 you informed me, that the committee were desirous that I the Society, 

 should furnish them with a sample of canvas painted in the 

 old method, and another on my improved plan, I trust that 

 I shall be able fully to comply with their requests. In the 

 first place, I have sent a small sample of the residuum of 

 the burnt canvas, fit for grinding in oil for paint, also a 

 piece of canvas painted therewith, marked No. I ; another 

 piece painted after the old method, marked No. 2 ; another 

 piece painted according to my process, marked No. 3 ; 

 and, lastly, a piece finished intirely with the new composi- 

 tion, marked No. 4 ; each sample having received three 

 coats of paint. Upon examining No. 2, you will find it 

 becoming from time to time more stubborn, in conse- 

 quence of the paint hardening; and when a small ridge is 

 formed in it, by pressing it between the finger and thumb, 

 it will soon discover, that it is subject to crack, and by this 

 means permitting the wet to enter it, will soon rot the 

 canvas. 



The space of time proper between laying on the new Time between 

 preparation and the second coat ought to be one entire 1 e coats " 

 day ; but if saving time is an object, the second coat may 

 be put on the day following the first ; for if the canvas is 

 placed in an advantageous situation for drying, the compo- 

 sition will dry or harden so as not to rub off. 



Canvas finished intirely with the composition, leaving it The canvas 

 to dry one day between each coat, will not stick together if J2J£* Stlck 

 laid in quantities, as you will find by making experiments 

 on the sample No. 4. 



Since the Navy board have given directions for ships' Requires liute 

 canvas to be painted according to my new method, I find, tline to fnuslu 

 upon calculation, that 1 have painted upwards of twenty 

 thousand yards since November last, a great part of which 



has 



