I can't say for sure if Gary was the first to use the term BIOS, but it was
being used by Gary in CP/M long before 1978 -- I'd say in late 1975,
probably, and definitely by 1976. A friend bought a copy of CP/M in 1975 (I
know it was 1975, because I was in Atlanta at the time with the friend (who
happens to be Dale Heatherington, a partner in the modem firm of DC Hayes,
for whom the "Heatherington patents" were named), and I moved from Atlanta
to Charlotte in late 1975). In 1976, Imsai was offering CP/M version 1.3
with the dual-Calcomp disk system. All of the documentation for all of
these used the term BIOS, as did the source code for the BIOS' supplied with
CP/M on disk (for the Intel development system). The BIOS resided on disk,
as part of the system tracks, it was the last 7 sectors of the 2nd track
(track 1).