In David Roberts’ story about Henry Waxman’s long struggle to strengthen the Clean Air Act (part one, two), some important lessons were unavoidably overlooked, because Waxman inherited, struggled with, and never did manage to remedy a serious architectural flaw embedded in the original 1970 version of the law.

When I first lobbied on clean air in 1970, there was a crucial but often forgotten fork in the road. Environmentalists urged Senator Edmund Muskie, who was leading the charge, to require that all polluting facilities, new and old, be modernized and cleaned up. When the business community pushed back and told Muskie that old power plants, refineries and factories were not worth cleaning up because they would shortly be retired, Muskie compromised.

Muskie believed that his compromise didn’t matter, because the plants over time would be phased out, and because he had a back-up mechanism — every airshed was nominally required to meet health-based air quality standards. Cleaning up existing plants was anticipated to result from this state-based process.

Unfortunately, Muskie misjudged. Companies found ingenious ways to continually upgrade and modernize facilities, turning them into virtual vampires: polluters that cannot be killed. As I write, there are 145 operating coal-fired power plants built before 1950; two-thirds of the coal fleet was constructed before Muskie passed the Clean Air Act.

The state-based process for meeting air quality standards also floundered. When states refused to clean up existing polluters and couldn’t meet the standards, Congress just gave them more time to do the job. Today, 40 years later, the nation faces an enormous air pollution problem, almost all of which results from emissions from these “grandfathered” vampire power plants.

All of this wasn’t a case of the law failing because its initial standards were too weak — it was a case of a basic design flaw which made it almost impossible to ever recover.

Compare this with the treatment of motor vehicles: Although the initial emission standards set for cars were far too weak, in this case the design was solid, the foundation strong — old cars really do wear out and get replaced by new ones. As a result, the Clean Air Act has been far more successful in cleaning up automotive pollution than emissions from power plants.

Over the years, Waxman has indeed been able to make tremendous progress in the face of this flaw. Over time he imposed more and more stringent emission controls on many grandfathered facilities, and in 1990 he applied the cap and trade sulfur program to all power plants, including the old ones. But if the 1970 act has set a deadline for old plants to clean up — even what would have seemed like a very slow deadline in 1970 — the air we breathe would be far cleaner today. (And, incidentally, since old coal plants are the biggest source of CO2, our global warming pollution would also be much smaller.)

In writing the sulfur emission and trading rules in 1990, Waxman made sure that his architecture was tight. Every plant has an emission standard — and it could only obtain additional permits if another power plant cleaned up its sulfur.

That’s how Waxman designed the original cap and trade program for carbon dioxide this year. But what he reported out of the House Commerce Committee had some enormous loopholes in it, in the form of billions of tons of offsets via which projects like tree planting or supporting clean energy overseas could be substituted for cleaning up one power plant or another.

This is the kind of “leak” in the system that Muskie allowed in 1970 — and if we don’t fix it in the Senate climate deliberations, it could be fatal. We don’t have forty years to start cleaning up coal-fired power plants and other grandfathered carbon emitters.

Fortunately, we can fix it without getting into the wrangle about allocating permits. Congress or EPA should simply specify that every boiler, on its 50th birthday, loses its status as an “existing facility” and must be modernized the meet the pollution control standards for new power plants.

The key lesson, in my view, is that a solid foundation is essential. Even if we don’t build the whole structure at once, we must get the basics right. And cleaning up every grandfathered power plant as it reaches its 50th birthday is the key to having a strong foundation for solving global warming in the 21st century.