You have to wonder if Cephalon ( CEPH) might act more shareholder-friendly by mailing motion-sickness medication to investors.

In recent years, Cephalon has produced a stream of upside surprises and downside shocks -- financial, regulatory, research-related -- which jolt the stock and keep stockholders queasy. Just in the past 15 months, shares bounced from the mid-$50s to the high $70s, then slipped to the mid-$60s, then rose to low $80s, then dropped to the high $60s.

Now, they're in the low $70s as Cephalon prepares to release third-quarter earnings on Nov. 8.

Forecasting the just-completed quarter and 2008 requires "sorting through many moving parts," said Gary Nachman of Leerink Swann, in a late-October research report to clients.

Despite the stock's bumpy ride and uncertainty over several issues, many analysts endorse Cephalon. Thomson First Call says buy ratings outnumber hold ratings by 17 to 10. Nachman is one of optimists, but he also injects some caution into his outperform rating.

His most immediate concern is an investigation by the U.S. Justice Department and another by Connecticut regulators of Cephalon's marketing practices for certain drugs. Cephalon has set aside $56 million in reserves to cover what is says is the minimum cost of a settlement.

Nachman says Cephalon might announce extra reserves on Nov. 8. In an August filing with the Securities and Exchange Commission, Cephalon said it was "reasonably likely" that any settlement or fines would "materially exceed" $56 million.

Nachman expects an agreement by year-end. Assuming the stock stays at its current level, Cephalon could absorb a settlement costing $300 million to $400 million "if it takes this major overhang away," he says. Nachman doesn't own shares. His firm is a market maker.

The analyst likes Cephalon for the new cancer-pain drug Fentora and the new sleep-disorders drug Nuvigil. "We believe Cephalon's oncology pipeline, especially the opportunity with Treanda, is very underappreciated," he says.

Treanda provided a boost on Oct. 23 when a late-stage clinical trial met its goal in treating a form of non-Hodgkins lymphoma, a cancer of the lymphatic system. Cephalon will seek approval from the Food and Drug Administration later this year. In September, Cephalon asked the FDA to approve Treanda for chronic lymphocytic leukemia.

Analysts had barely enough time to cheer -- comments ranged from "impressive" to "home run" -- when Cephalon and the FDA announced a strengthening of warnings for Provigil on Oct. 24. The sleep-disorders drug accounted for $415.5 million, or 47% of corporate revenue, for the first half of 2007. Patients and doctors are being warned about a rare, dangerous and sometimes fatal skin rash and about some reports of psychiatric problems, such as hallucinations and suicidal thoughts.

The back-to-back revelations for Treanda and Provigil is just one example of why some analysts get queasy, or at least prefer to sit on the sidelines until the swirling sensation subsides.

"A good deal of uncertainty surrounds Cephalon's new drugs, and the failure of a few launches could change the face of the company," says Karen Yiu, of the independent research firm Morningstar, in a report to clients late last month.

She cut her fair value to $64 a share from $77, thanks primarily to the poor performance of the alcoholism treatment Vivitrol. Initially, she said Vivitrol could produce peak annual sales of $500 million, but now she says it may average $200 million a year.

Cephalon markets Vivitrol for its developer Alkermes ( ALKS). Vivitrol, a once-a-month injectable drug, was approved by the FDA in April 2006.

Although Cephalon has an "impressive" marketing record and roster of experimental drugs, Yiu worries about Nuvigil and Fentora. These drugs "may not succeed in the marketplace, and there is no guarantee that the firm's other drug programs will be able to pick up the slack if one of these drugs fails to live up to its potential," she says.

Fentora, approved by the FDA in September 2006, is the successor to Actiq. Both treat severe pain in cancer patients and both contain the powerful painkiller fentanyl.

Hit by generic competition, Actiq's sales dropped to $129.8 million for the first half of 2007 from $289.7 million for the same period last year. Cephalon also sold $64.9 million of its own generic Actiq, as well as $68 million of Fentora during the first half of this year.

Some analysts are edgy about Fentora's start. They wonder how sales growth will respond to a recent FDA alert about some severe side effects and deaths due to improper use of the drug. The FDA has asked Cephalon to strengthen the drug's warnings.

The market-expanding break will be if Cephalon can convince the FDA to approve Fentora for other types of pain. Analysts expect Cephalon to seek FDA approval by year-end, but a concern about side effects could affect the regulatory review.

The other big drug of the future is Nuvigil, a chemical cousin of Provigil. Both treat sleep disorders such as narcolepsy and problems experienced by people whose work shifts change constantly.

The FDA approved the longer-acting Nuvigil in June, but Cephalon has delayed marketing to prevent cannibalizing Provigil's sales. Cephalon can afford to do this because it negotiated deals with six companies between late 2005 and August 2006 to halt patent litigation and delay generic competition until the spring of 2012.

Cephalon will probably introduce Nuvigil in 2010. If Cephalon had lost any patent challenge, it would have had to push Nuvigil into the market last year.

The agreements were controversial, and the Federal Trade Commission continues to investigate the Provigil deals. Some analysts say the FTC could allege anticompetitive behavior.

"We are cooperating fully with the FTC," Cephalon says in an August SEC filing. If the FTC cites one or more deals, "we believe such a challenge would take years to resolve," Cephalon says.