API Std 607 Fifth Edition, June 2005 appears to be a complete revision from the fourth edition 1993. My new pdf file is bigger and includes more pages. The table of contents reflects a complete revision as the section numbering and descriptions are completely different. The standard number is coordinated identical to ISO 10497-5:2004 and includes only SI units.
RR05,
about the same issue, please take also a look at thread408-163494, thread408-163274 and/or thread408-114236 within this Forum.
In addition to the information by JLSeagull, the main practical difference is that in certain cases [see ISO 10497:2004, clause 5.6.4-b)], the maximum test pressure from 30 psi[sub]g[/sub] = 2 bar[sub]g[/sub] is increased up to the 75% of the valve ambient temperature rated pressure.
This results in a more complicated test rig (which requires a closed loop) and, most of all, in more delicate safety issues (some national laws may require special systems and permissions from authorities such as Fire Brigades, etc...).
Personally I'm not aware of any "famous" valve Manufacturers or testing Laboratories that have performed fire tests according to the standard edition in object, and even searching the web I found little interesting results (regarding small manual ball valves only).