Tonight I Thought I would write a little on Sargent's working methods. Sargent's was what we call today a wet into wet painter, he painted alla prima. Unlike many artists of his time and the generation before, Sargent painted directly. He didn't transfer a finished drawing to the canvas and color it in with glazes. He sat his easel up in front of nature and painted his interpretation of its appearance.
Often when painting portraits Sargent worked with the canvas up next to the model. He was described as using this method (sometimes called sight size) on several occasions. How doctrinaire he was about sight size I don 't know. I would guess not very. But he evidential did place the canvas up with the sitter, walked back to a fixed point in the room to make an observation and then walked back to the canvas and placed that note on the painting. Then he would return to his observation position and look for the next note. I don't believe he always worked this way, but he certainly did often.
Sargent got many sittings for his portraits. Much has been made about Sargent scraping down and redoing the same head over and over. I believe (based on my own experience) is that this sounds like something different than what it actually may have been. I think two separate actions were observed and recorded by those who saw him paint. The first is that Sargent was likely to take a knife and remove a failed passage before repainting it. That's a good habit to be in. Rather than letting the paint just pile up, Sargent pulled it off with his knife and made another go at it.
Here's where my opinion (and that's all it is too, I am not a Sargent scholar) differs from many who have described his approach. Some of the chroniclers of his time remarked that Sargent would completely remove the painting with a knife before working on it a second or additional sitting. Having been trained by Ives Gammell ( who actually met Sargent, though he did not study with him) in an approach at least partly based on Sargent as transmitted by the Boston school painters who idolized Sargent, I am familiar with something similar but slightly different.
My guess is that Sargent's portrait sittings were far enough apart from each other that he started the sessions with a dry or nearly dry canvas. Sargent worked with lead white which makes the paint dry very quickly. Lead is an excellent dryer and has been used as that historically. I was taught that something you can do when beginning work over a dry canvas was to scrape it with the side of your palette knife. That takes the edges down and even though the fineness of the painting may be lost, more than a ghost of the image remains. I believe this was what Sargent may have been doing. To a casual observer the painting might have looked ruined, but there was a fine if slightly degraded image left on the canvas. Flake white scrapes very differently than the titanium white with which you are no doubt more familiar. Lead white scrapes beautifully to a smooth surface. The side of the knife actually cuts into the lead in a way that it doesn't with titanium which is more rubbery. Any paint that sticks up proud of the surface, slices neatly away with the edge of a palette knife.
Because he still had a good deal of his image left, not a lot was really lost by doing this and Sargent could work on his canvas without having to worry about raised edges of paint or too much buildup. This way each days painting was a rehearsal for the next. The paintings look like they were rattled off in one go, because they were rattled off in one go. However there were any number of rehearsals under the image before that one go.
I expect that there were times when Sargent scraped back a painting that was still wet at the end of a sitting, but even then, I expect he scraped enough to get a decent surface and blur his image some, but there was a pretty good ghost of an image left on the canvas to guide his next attack on the problem.