Tom Ruane of Retrosheet looked into this after I asked him about it. Here is the link:
Best/Worst Month for a Team's Pitchers
Tom mentions "Since 1935, only the 1954 Orioles had a lower HRA rate than those 1969 Mets." Tom also looks at the best months in hits allowed and strikeouts.
Prior to Sept. that year, the Mets' HR rate was .023842 (using batters faced - IBBs - SH). In Sept. it was .002841. So the Sept. rate was only about 1/8 of what it had been before that year.
Using the cumulative binomial distribution and assuming the following:
Number of occurrences: 3
The probability of getting 3 or fewer HRs was about 1 in 36 million (I welcome any comments or corrections on this). Just cutting the HR rate in half or less that month (about 13 HRs allowed) has chance of only 1 in 183. 6 HRs or less is 1 in 225,000.
The Mets did allow 3 HRs in the last two days of the season in Oct. But by then they had clinched the division. They allowed no HRs from Aug. 30 to Sept. 18 in a total of 22 games (that is what a quick check of Retrosheet shows).
On Sept. 19, game 2, Stargell of Pit. hit a 2-run shot in an 8-0 Pirate victory. But by then, the Mets already had a 5 game lead with 13 games to play (the Cubs had 11 left).
On Sept. 21, game 1, Pagan of Pit. hit a solo shot in the 4th inning. It made the score 4-3 but the Mets won 5-3
On Sept. 21, game 2, Stargell hit a solo shot in the 4th inning to make it 4-1 but the Mets won 6-1.
It seems like none of these HRs was significant. Only one even had a man on.
The Mets also allowed only about 31% of their HRs for the whole year with runners on while the rest of the league was about 43%. The Mets turned 30 DPs in Sept. Their next highest month was 25 and the next after that was 18. Don Cardwell had an ERA under 1.00 in Sept. It was .039 in 23 IP. It was 3.25 going into Sept.