Multisourced

Defense Tech has a mindbloggling article on the process by which mobile biological weapons labs became a plausible part of Saddam's alleged doomsday arsenal.

It's worth reading in full, with all the links, but the gist is that then-UNSCOM inspector Scott Ritter floated the mobile-lab idea to Ahmed Chalabi in 1998, as a possible explanation for why UNSCOM couldn't find Saddam's fabled weapons cache. Not long after, Chalabi's worthy constituent "Curveball" began prattling about BW trucks to German intelligence, who dutifully passed the information to the CIA.

In September of 2001, the United States actually created a prototype mobile lab, either as a proof of the mobile-lab concept or in order to train troops (or both):

In July 2003, there were three press reports about a truck platform that Dr. Steven Hatfill arranged to have constructed towards the end of 2001 under government contract at A.F.W. Fabrication, a metalworking plant in Frederick, Maryland. These reports describe this project as a contract to SAIC, the Science Applications International Corporation, by the Defense Threat Reduction Agency, US Department of Defense, for the US Special Forces Command at Fort Bragg, North Carolina. It was supposedly intended for use in training members of the Delta Force.

If the name Steven Hatfill sounds familiar, it's because John Ashcroft named him a person of interest in the 2001 anthrax mailings.

Here's what happened next:

The agency develops graphics drawn by a U.S. contractor based on Curveball's story and might have known of the mock-up BW lab built for SOCOM, both of which "confirms" the concept that Iraqi mobile BW labs exist, which leads to SecState Powell's speech at the UN in February 2003 and the media's echo chamber agreeing with the president that there's enough evidence to go to war against Iraq.

This, apparently, is what Colin Powell meant when he said that the information on Iraq's mobile labs had been multisourced.