I find I think with my fingers more than with my brain, I suppose; so, for example, I mean in the case of the Lamb shift, what I would do is to sit down and do some calculations and just use whatever tools I had available and see whether they worked or not. So I would flounder around for a long time, sort of taking bits out of Heitler's book and trying to figure out why they didn't work.

[Q] On paper?

Yes. So it was all being done in writing, so that I would gradually work out how to do it by doing it. I'd never be able to plan ahead, and I think that's true of everything I do in maths or physics or writing prose. Essentially I can only make the plan as I go along, and so the difficult part is always this floundering at the beginning.

[Q] But once you are finished floundering you more or less know how long it will take you to actually get through the problem?

Well, usually I rather tend to overestimate the time. I mean, what I tend to think is, that'll take a month and it actually takes a couple of weeks or something...

[Q] But you know more or less what is involved?

Know more or less, yes. I mean, I think once I get going I am sort of one-tenth as good as Bethe, which means pretty good.

[Q] And you just can, so to say, plough ahead then?

Yes. I had this ability to go through a calculation and really get it done, which of course is the superlative quality of Bethe.