The assumption we need to make is that this is 8-bit 2's complement. So the valid range of integers is -128 to +127 (decimal).
Originally Posted by anderson
Provided we don't overflow this range, we can work to 8 bits and ignore any 9th bit that may arise.
So in (i) when you add the 2's complement numbers you get:
...11100011 +Ignoring the overflow 9th bit (1), converting the remaining 8 bits back to decimal gives the answer -66.
This situation with the overflow bit doesn't arise in (ii) and (iii), so there's no problem with your working.