Thanks for your reply Liz. The methodology you use is the same as mine. So in this case, using her blanking maiden name revealed 2 hits and as both had identical mother's maiden name, I assumed she had married twice. As neither were LAWRENCE, I thought then that she must have been married 3 times. But where was the entry for LAWRENCE? Then noticing that the 2 hits had different death dates, like you I concluded they must have been 2 different women. So, that's when I put LAWRENCE in as surname rather than blank, suspecting that I would find an entry with a death date matching 1 of those 2 hits. There is indeed an Ann LAWRENCE with mother's maiden name DUNCAN, but she died in 1922. I have now concluded that the Ann GRAY I am looking for died in 1922. That is, unless this Ann LAWRENCE is another person altogether and just coincidentally also had mother's maiden name DUNCAN, and the death of the Ann GRAY married to Charles LAWRENCE was never registered.
How did Scotlandspeople neglect to index her 1922 death under Ann GRAY. Also what a coincidence having 3 different death registrations for an Ann GRAY and all showing mother's maiden name DUNCAN, particularly as all 3 deaths were prior to 1974.
I think I will write a note to Scotlandspeople reporting the index oversight re her missing maiden name entry.