The reason Pina says on page 65 that it is a 512K board, is because there is typically no other reason to remove the ROMs from a 128K board.* That is not always indicative of the problem. As techknight points out, there could be several other reasons for this problem, which had not yet come to light at the time Pina wrote that book, but only became problematic 20 years later. Heck they could have even popped out during shipping.

Also, the 41256 chip pin 1 being connected to the high address bit, has absolutely no affect on a 128K and causes it to function as a 64k dram. Since pin 1 is not connected on a 128K, or at least not connected to a decoding matrix, the extra RAM of the 256k is not recognized. However, I have never mixed and matched these chips. They should function properly together, but why needlessly create potential problems. Also, I'm not sure you should mix and match any chip types in any kind of permanent way. As Pina suggests, if you do have to replace any chips, replace them with sockets, that way you can experiment with whatever chips you had ... Although i seem to recall someone mentioning that all of the chips should be docketed if any.

*The other reason Pina does not cover for some reason, is that there was a ROM revision necessary for early Macs that had their 400K drives replaced with newer 400K drives.