It's all Sunday afternoon TV's fault. I have never until today run the Gutmann 35 pass secure deletion method, but idle fingers, etc.
In Recuva I set the deletion method to Gutmann 35 pass. I ran a scan, and chose a file which had no overwritten clusters. I ran secure delete against it. In the header display of the first 128 bytes were hex zeroes. I am assuming that the rest of the file is the same.
Now I tried this (separately) on a 3 kb, a 4 kb and a 29 kb file. All did the overwrite as quickly as a normal one-pass would do, as far as I could tell. These aren't large files, but 35 writes of 29k is over 1 mb. Not a lot to overwrite in one go, but a lot of writes if the write cache were flushed after every pass to force each write to disk.
Secondly all the files ended up filled with hex zeroes. But the Gutmann method starts and ends with four passes of random data, and only one pass, the tenth, is hex zeroes.
I don't really mind, or care, whether the Gutmann method is being used or not, as I shall revert to normal one-pass deletion once again. But it is interesting, no? No?