U.S. Rep. Edolphus Towns seemed to be getting angry last week, and I was beginning to understand why.

For several weeks, Towns' House Committee on Oversight and Government Reform has been examining the latest recall by McNeil Consumer Healthcare, of Fort Washington, and the response by the Food and Drug Administration.

At least three times now, Towns (D., N.Y.) and his committee's staff have encountered what appears to be smoke blown in their - or the public's - direction.

The first puff shrouded the magnitude of the April 30 recall. Though it was clear from the start that this was a massive undertaking, McNeil seemed oddly vague about how much product was affected.

McNeil said initially that it was recalling "all lots that have not yet expired" of pediatric formulations of four brands: Tylenol, Motrin, Benadryl, and Zyrtec.

But how much was that? At a hearing June 1, Towns said his staff had been told that the recall involved six million bottles of infant and children's medicines. But in testimony that day, an FDA official said the recall covered 136 million bottles.

The company's explanation? A simple misunderstanding.

"The numbers responded to two different questions," said McNeil spokeswoman Bonnie Jacobs. McNeil was recalling six million bottles from stores, and 136 million bottles already in consumers' hands. The total: 142 million bottles.

Then there was the so-called "phantom recall." According to documents obtained by the committee, McNeil's current problems with actual recalls were preceded by a non-recall recall of adult versions of Motrin that did not dissolve properly.

Although the FDA eventually insisted on a formal recall of the Motrin, a brand version of ibuprofen sold in eight-tablet vials, McNeil began with an unusual first step. Instead of a recall, it turned to a contractor that specializes in "reverse logistics" and directed it to go to about 4,500 stores to buy back the suspect product.

"You should simply 'act' like a regular customer while making these purchases," according to one document disclosed at the June 1 hearing. "THERE MUST BE NO MENTION OF THIS BEING A RECALL OF THE PRODUCT!"

In a statement Friday, Jacobs sought to distance McNeil from the document, which she said "did not come from McNeil. It was created by a subcontractor with whom McNeil had no direct contact."

The statement also challenged the significance of the decision to pull back the product:

"Given that there was no safety risk, the objective was to remove the affected product from a unique distribution channel - mainly convenience stores and gas stations - with as little disruption and consumer confusion as possible. The FDA was kept informed of the survey and product retrieval, and after it was essentially complete the FDA requested that we conduct an additional step, which was to conduct a formal recall, and we agreed."

But it's not clear whether the FDA was even on board initially. In a July 2009 e-mail that lawmakers made public Friday, one FDA investigator complained that McNeil was "doing a recall even though you are calling it a 'retrieval.' "

The investigator added: "The agency's position is that your company should do a voluntary recall of the product, since it appears to be that you already are doing a recall of the product."

Other documents support the notion that the "phantom recall" was meant to protect McNeil's reputation - a fair suspicion, since it has to be costly to buy back product at convenience-store prices.

One April 2009 e-mail said Inmar Inc., McNeil's contractor, initially had planned to "buy all products as mystery shoppers," but then decided to limit its efforts to particular lot numbers.

"We will need to clarify to the field what the response will be if the store asks questions as we are reviewing products searching for specific lot codes," Inmar's Rob W. Small wrote to a colleague and a McNeil staffer.

It appears Small had some ideas. He wrote:

"Some typical responses:

" 'We are looking at packaging changes/conditions of the product and will be purchasing some samples.'

" 'We are conducting random quality and/or packaging checks.' "

Smoke? Towns seems to think so.

"These documents are extremely troubling," he said in an e-mail statement Friday. "We cannot tell where the spin ends and the truth begins."

But there was a third incident that I find even more troubling - perhaps because, this time, the same smoke had been blown directly at me and at Inquirer readers.

In late May, committee sources say, McNeil president Peter Luther told investigators that the Fort Washington plant did not make children's medications for other labels. But just four days later, the FDA announced a recall of four versions of PediaCare - a competing product to McNeil's children's medications - that had been made at the Fort Washington plant. (For more information: http://go.philly.com/pediacare)

It turns out I'd asked the same question myself, on May 3, when I first wrote about the recall. Because the FDA was suggesting that parents could turn to alternatives, I e-mailed a McNeil spokesman, Marc Boston: "Does McNeil make any of the generic versions of these products, or products sold under other brand names, at the same facilities?"

Boston replied: "McNeil Consumer Healthcare does not manufacture other versions of the recalled liquid pediatric products."

Jacobs said Friday that the discrepancy was just a quirk related to McNeil's sale of the PediaCare brands last year to Blacksmith Brands Inc., their current owner.

"For a short period of time after the sale of PediaCare, McNeil continued to manufacture PediaCare until the acquirer was fully operational," Jacobs said.

It's reassuring to remember that the FDA and McNeil continue to say that there is little risk to children from the defects involving the medications' quality, purity, and potency. Blacksmith called its recall "a precautionary step."

As for me, my eyes are still stinging from McNeil's smoke.

Contact columnist Jeff Gelles at 215-854-2776 or jgelles@phillynews.com.