“There is something very wrong here,” Elias Zerhouni, director of the US National Institutes of Health (NIH), told the annual conference of the National Postdoctoral Association (NPA) in Berkeley, California, on 30 March. He was lamenting that most principal investigators are 40 or older when they get their first NIH grant. It's a problem the agency is trying to address with the K99 grant, which started last year.

Intended to bridge the gap between postdoc and independent investigator, K99s fund individual postdocs for two years and then travel with them to an independent appointment for three years. In November 2006, the NIH awarded 58 K99s; that should go up to 170 this year, a big commitment given its flat budget.

But the very grant designed to help young scientists gain a foothold shuts many of them out. To be eligible for a K99, researchers must have received their PhDs within the past five years. “Second postdocs are becoming a common thing, and as the length from the defence date grows, you're losing opportunities,” says Scott Nowak, who is on his second postdoc at the Memorial-Sloan-Kettering Cancer Center in New York.

It may also fail to serve another crucial role: attracting promising foreign postdocs to first-time faculty positions in the United States. Non-US scientists are eligible for K99s, but the lure of alternative career paths and attractive options abroad is diverting foreign talent, according to Rajika Bhandari of the Institute of International Education, a non-profit research group. A 2007 report by the Lumina Foundation for Education estimates that by 2025, the United States will be short of 16 million workers holding an associates or higher degree.

Zerhouni defended K99 as a “defining experiment” in how best to prepare the next generation of scientists. He said it was not intended to make broad changes.

Pushed by the NPA, the NIH and the US National Science Foundation recently settled on a joint definition of postdoctoral scholar, stating that postdocs are in a period of “mentored advanced training” designed to move scientists along their “chosen career path”. That establishes a research sponsors' duty as adviser and implicitly acknowledges that not all careers are academic, says Diane Klotz, chair of the NPA's board of directors. But Zerhouni, noting that postdocs fill highly specialized roles within labs, urged the NPA to do more.