Thank you everybody for the suggestions. Here is how things finally shook out. The Pentium machine had a real, 360KB drive in it borrowed from a PCjr. (I had stated that in the original note.) Formatting issues between 1.2MB drives and 360KB drives were not part of the equation. I tried formating the diskettes (double density, double sided) in both the 360KB drive on the Pentium and the 360KB drive in the PCjr. This did not seem to make a difference. Diskettes formatted and written by the Pentium machine were usually unreadable in the PCjr. In a few tests they were also unreadable in a second PCjr. Sometimes I could get a directory or copy a file, but usually I was getting the familiar "Abort, Retry, Ignore." A diskette "refresh" utility was reporting lots of read errors on sectors, some recoverable and some not. Diskettes written by the Jr were readable on the Pentium. Old diskettes written years ago were readable everywhere. (These were written by yet a different drive.) I deduced that since the Pentium 360KB drive was writing the diskettes and two PCjrs couldn't read them, that it must be the Pentium 360KB drive causing the problems. So I swapped it out and put a 1.2MB drive in it's place. I know about the head width problems, but would you believe that one of the Jrs was able to read diskettes created on the 1.2MB drive? To make a long story short, the 360KB drive in the Pentium was ok, the 1.2MB drive can write 360KB diskettes that are readable on the PCjr, and one of the PCjr drives is flakey. I think I was unlucky in my early testing when both Jrs balked at the diskettes created on the Pentium. One of the Jrs drives is know known to be bad, one is confirmed good, and the 360KB drive on the Pentium is good, yet a diskette created on the Pentium caused read errors on both Jrs, which led me to suspect the wrong drive. I've not been able to reproduce that error again. Now for the bad drive ... If I clean the rails with alcohol and lube them with silicone spray, it does a slightly better job of reading foreign (i.e.: not its own) diskettes. Based on this, methinks that the head is not being positioned exactly where it should be. However, those rails are pretty clean now and the problems persist. It's an old Qume drive, 1/2 height. I'm saving the drive for now, but a flakey 5.25 inch drive isn't of much use. And I've also learned that DOS 2.1 does NOT tolerate diskette errors during the boot process. :-)