Democrat superdelegates, pay heed.

An Indian astrology website , previously named the No. 1 astrology website by PC World magazine, is predicting a Sen. Barack Obama (D-Ill.) victory. The logic, in summary form, seems impeccable to me: “In Navamsa, Lagna lord Venus and 7th lord Mars are in 8th. … The present period is Jupiter MahaDasa Sun antar … indicating the great power of sun in tenth — which is making him move from strength to strength. During election time he is having the antar dasa of Moon, the tenth lord, in exaltation from Sept. 24, 2008 to Jan. 17, 2010. Thus the combinations and periods are very favourable for him to get the Democratic nomination and win the election.”

I am particularly drawn to the “strength to strength” argument. It’s something to admire in a candidate. But should influential Democrats worry, nonetheless, that Barack’s antar dasar doesn’t begin until after Labor Day? Could that be too late?

This is nutty speculation, of course, but I cannot help but wonder whether some superdelegates might seek any insight possible to do their ostensible jobs — namely, saving the Democratic Party from non-superdelegates.

Most elected officials are likely to look beyond astrology to scientific public opinion polls. But which ones should they pay attention to? Democratic primary trial-heats polls pitting Obama against Sen. Hillary Rodham Clinton (N.Y.)? Or the results of general election head-to-head contests between Sen. John McCain (R-Ariz.) and each of the two Democrats? And which polls should matter most? Those taken in their own states? Or polls taken nationally? And whose numbers should they pay most attention to? Gallup? Associated Press/Ipsos? Geoff Garin or Joel Benenson polls?

These are critical questions because they expose the impossible choices that lie before these superdelegates. The premise behind the greater inclusion of superdelegates in the process, after the Democratic disasters of George McGovern in 1972 and Jimmy Carter in 1976, was that elected officials could add some “ballast” to the Democratic ship that would prevent it from being blown off course by sudden winds of passion for populist candidacies. Precisely how they were supposed to accomplish this feat has never been clear. Accounts of the 1982 contest lauded the newly empowered superdelegates for explaining Walter Mondale’s positions better than he could himself. Superdelegates also played a role in furthering a “whisper campaign” about Gary Hart’s private life. So the first-ever superdelegates of the modern era continued acting like old-fashioned power-brokers, throwbacks to the era before Democrats empowered their masses through delegate-selection rules changes.

Today’s power brokers are far more likely to look at polls, I suspect, than were their predecessors. Polling data are more widely available today. And candidates and their advisers are more sophisticated consumers of polls. Yet while these modern politicos may be empiricists practicing their science in smoke-free rooms, they are still politicians, so the process may not deviate as much as you think from the power-broker days. There are so many different polls out there that I suspect every superdelegate can find one survey that confirms the outcome he or she intuitively prefers for the Obama-Clinton fight. And every politician must make his or her own personal world safe for reelection before worrying about the whole party. The superdelegate, trying to serve as “ballast,” cannot be expected to endorse a candidacy that will eventually create a vortex that even takes other elected Democrats down with the ship.

So reading the polls is more of a challenge than you might suspect. It’s far more than simply picking a candidate who might win next November. It’s also incumbent on Democrats to choose someone who can do for the party what the superdelegates are supposed to do, protecting the liberal Democratic franchise.

That may be asking too much.

Hill is director of Hill Research Consultants, a Texas-based firm that has polled for GOP candidates and causes since 1988.