HERE COME the S&Ls again. It's not enough that the epidemic of S&L failures is now going to cost the taxpayers more than $100 billion and perhaps much more. The S&L lobbyists are back working the administration, plucking at the sleeve of Congress, bleating that the reforms enacted last summer are too sudden, too harsh, too awful in their impact to be endured without a generous period of transition from the slack rules of previous years to the moderately tighter ones now in law.

The only astonishing thing about this performance is that it seems to be having an effect. A delay of one crucial rule is under discussion at the Treasury. That's grotesque. The Treasury's responsibility is to say no, flatly and immediately.

Dozens of congressmen are sponsoring legislation to postpone the new requirement. Doesn't anybody remember that it was precisely this kind of forbearance, inspired by this kind of lobbying, that lies behind the failure of hundreds of S&Ls at inordinate cost to the public?

The issue here is the limit on the amount of money that an S&L can lend to one borrower. In the past, it could lend as much as 100 percent of its capital -- meaning, as some of the regulators have pointed out, that the failure of one big loan could (and sometimes did) throw an institution into immediate bankruptcy. With its capital wiped out its further losses would (and did) fall directly onto the taxpayer, who stands behind the deposit insurance. In last year's reform law, Congress held that no S&L could lend more than 15 percent of its capital. With that limit, it would take seven bad loans to bankrupt the lender. The 15 percent rule is entirely reasonable. It has applied for years to banks.

But Congress has heard from a lot of real estate developers and builders who claim that the new rule will leave them without the credit to carry on their business. That plea itself ought to make any congressman wary -- not to mention the secretary of the Treasury. Some S&Ls became captives of big builders or groups of builders who used them to finance projects that could not survive arm's-length credit analysis. One purpose of the reform was to end this kind of excessive and dangerous mutual dependence between borrower and lender.

The proposal under consideration at the Treasury would phase the new limit in gently and gradually over a couple of years. At a time when the S&L industry is continuing to run huge losses, any procrastination on these reforms only invites more risky lending and, ultimately, higher costs to the public.