Well actually I tried that since I posted that message. Tried a number of things last night.
1) With The "De Fault" latency timing settings of 5-5-5, 18, 3, 3. and the voltage at its "De Fault" setting, running MemTest 4 Deluxe 64 bit version, I got 57 memory errors in one pass.
2) With the voltage bumped by .1v and the timings still at their "De Fault" settings, the number of errors was reduced by approximately 1/2. I didn't bother to finish the test pass though, because although it was helping it was not the solution.
3) With the timings changed to 6-5-5, 18, 3, 3. and the voltage at its "De Fault" setting it was able to run MemTest error free for over 19 hours.
What that suggests to me is : This is commercial quality RAM rather than "Top Drawer" and the Gigabite GA P35 DS3L is apparently a little FUSSY about the quality of RAM being used. I suspect that there are other boards out there that this RAM would work just fine in, but this particluar kind of "Giggle Bite" isn't one of them. I have heard of similar problems with the ASUS P5E3 PRO LGA 775 X48 DDR3 board.
I know from past experience with Super 7 boards that some were completely UNFUSSY about RAM, (ANYTHING world work), e.g. ASUS P5A-B; others were moderately FUSSY, - (about 50-50 things would work well with any given type of RAM ), e.g. DFI K6BV3+ /66 and others were EXTREMELY FUSSY, (Only the VERY BEST quality RAM could be run at its most aggressive timings), e.g. FIC VA 503+
Whether or not this RAM may be considered "Defective" because it is incapable of running at 5-5-5 timing in this particular type of board is an open issue. I suppose the thing to do is for Buffalo to test some in one of these boards, and if they can't find samples that work then it is a case of the board being too FUSSY for this quality of RAM. If on the other hand they find that it does work in the majority of cases, then my RAM could be RMAed.