“…It is found that the edge length increases linearly, as reported in the preceeding paper [5], corresponding to a constant growth velocity v , 7.5 · 10\ nm/s in this case. It turned out that v varies considerably from day to day, as already observed in other experiments [2,5], with values between 0.01 and 1 nm/s in our experiments. Figure 3 shows the evolution of another triangle, but now the scanning was interrupted several times: for 700 s, as the triangle had reached a size of about 40 nm (interval I), for 600 s at a size of 60 nm (interval II), and for 2500 s at a size of 80 nm (interval III).…”