POSNER, Chief Judge.

Tregenza v. Great American Communications Co., 12 F.3d 717, 721-22 (7th Cir. 1993), holds that the one-year "borrowed" statute of limitations applicable to fraud suits brought under the SEC's Rule 10b-5 begins to run when the plaintiff first learned of facts that would have caused a reasonable person to suspect the possibility of a misrepresentation or misleading omission. (All other reported appellate cases are in accord. See references in id. at 718, plus Dodds v. Cigna Securities, Inc., 12 F.3d 346 (2d Cir. 1993), which, having been decided only nine days before Tregenza, we overlooked; and cf. Freeman v. Laventhol Horwath, 34 F.3d 333, 341 (6th Cir. 1994).) In the present case, the district judge granted the defendants' motion to dismiss for failure to state a claim on the ground that it was plain on the face of the complaint that the plaintiffs had in their possession, for more than a year before they filed suit, information that should have alerted them to the possibility of fraud and impelled them to begin to investigate.

The complaint, our only source of facts, is long and intricate, but it can be summarized very briefly. The principal defendant, Medco, is a pharmaceutical company that develops drugs for the diagnosis and treatment of heart disease. In 1988 it had licensed a predecessor of Fujisawa-USA to manufacture Medco's first product, "Adenocard." The following year the Food and Drug Administration had approved Adenocard for sale in the United States. Sales grew rapidly, but Medco was anxious because all its revenues came from the sale of this one product and the product was beginning to encounter competition. Medco had to develop another successful product — fast. The candidate of choice was a related product that Medco dubbed "Adenoscan" and that is designed to simulate the physiological effects of a treadmill stress test for cardiac patients who are for one reason or another unable to take the test or perform adequately on it. Again Medco licensed Fujisawa-USA to do the actual manufacturing. Before Adenoscan could be sold in the United States, it (like Adenocard before it) had to receive the FDA's approval. The fraud alleged in the complaint is that the defendants represented that approval was imminent even though they knew that, because of problems with safety and quality control at Fujisawa's plant, the application for approval was in serious trouble and the approval would be substantially delayed. On April 7, 1992, about a week after Medco had temporarily withdrawn its application for the approval of the sale of Adenoscan, Medco was telling its shareholders that its application was on track and that Adenoscan could be expected to be approved by the end of the year. In April of 1993, a year later, with Adenoscan still not approved (it was not approved until December 14, 1993), Medco announced that it was going to sue Fujisawa. The suit was filed the following month. The announcement of the impending suit was publicized in the business press, revealing to investors — for the first time, say the plaintiffs — the problems at Fujisawa's plant that had been responsible for Medco's inability to obtain timely approval of Adenoscan from the FDA and the existence of which contradicted the representations that Medco and the other defendants had made that approval was imminent.

The present suit, a class action on behalf of persons who bought stock in Medco during the period of the alleged fraudulent concealment of the difficulties Medco had encountered in seeking approval to market Adenoscan, was filed on September 1, 1993. The question whether the suit is untimely depends therefore on whether the plaintiffs should have suspected before September 1, 1992, that the defendants were engaged in fraud. In arguing that they should have suspected, the defendants rely on two "critical facts," in their words. The first is a precipitate decline in the price of Medco's stock. That price had peaked at $31.75 per share on January 24, 1992. By April 24, 1992, it had plunged to $15.25, a decline of slightly more than 50 percent, and while it rose a bit afterward it was still only $15.50 at the end of August of that year. The second allegedly critical fact is that on June 9 the FDA recalled several batches of Adenocard because of "short filling": presumably as a consequence of a manufacturing problem at Fujisawa's plant, some vials did not contain as much of the drug as they were supposed to.

We have great difficulty in seeing how these two facts, even in conjunction, would cause a reasonable investor to suspect fraud. Everyone knows that the process of obtaining the FDA's approval for a new drug is fraught with uncertainty; and when a drug company's entire earnings are from one declining product and it is counting on a new product for its financial salvation, unexpected delays in obtaining an approval without which the product cannot be sold are likely to have a devastating impact on the company's stock price. Stated differently but equivalently, much of the price of Medco stock during the relevant period must have represented simply a capitalization of the expected earnings from the sale of Adenoscan, and those earnings in turn depended critically on the timing of the FDA's approval. If that approval was long delayed, then as had happened with Adenocard the earnings from the new drug would be eroded by competition in the market for new drugs, which is dynamic. It is misleading to point to the January 24, 1992, peak of $31.75. There can be no inference that it represented a stable value of Medco's shares. Three weeks before, the price had been only $25. Two and a half months before that it had been only $16.75. And in January of 1991, only a year before it reached its peak of $31.75, it had been only $6.75. What goes up fast sometimes comes down fast — without fraud.

As for the recall of Adenocard, product recalls have become a familiar feature of the American economy. How common they are in the drug industry we do not know. There is no evidence on the question, although the well-publicized recalls of pacemakers and other medical appliances and drugs are a matter of common knowledge and lead us to suspect that recalls in the industry are not all that uncommon. No doubt the recall of Adenocard was an indication that the Fujisawa plant had problems, problems that might lap over to Adenoscan, which was being manufactured at the same plant. It was an indication, in other words, of looming difficulties with the approval of Adenoscan. It does not follow that it was an indication of fraud.

The defendants rely heavily on the portion of our opinion in Tregenza in which we said that the plaintiffs should have been alerted to the possibility of fraud by the precipitate plunge in the price of the defendant's stock. 12 F.3d at 720. The defendants come close to arguing that a big decline in the price of a stock is notice per se of the possibility of securities fraud. They overread Tregenza. The case involved a secondary offering of stock at a price of $12.50 a share. The defendants represented that the stock was greatly undervalued at that price, that the price should rise to at least $20 within two years, that the downside risk was less than 10 percent, and that prominent investors held large blocks of the stock. The offering was in October of 1989. Immediately afterward the price of the stock began a long and unbroken decline. By October of the following year it had fallen to $1.50, and later it fell to $1 — a fall of 92 percent from the price of the secondary offering. Given the strong representations that the defendants had made concerning the undervalued character of the stock, its glowing prospects for appreciation, the lack of downside risk, and the support of prominent investors, the enormous, unexplained drop in the stock's price — a much greater (although also less rapid) drop than in this case — should have rung an alarm bell. Far from being undervalued, as had been represented, the stock was greatly overvalued — seemingly by a factor of ten. Its upside potential was nil, its downside a free fall. The people who bought at the offering price and later filed suit should have been awake to the possibility that the defendants' representations had been intended to puff up what the defendants knew was a dog.

No similar inference of probable misrepresentation leaps out from the facts set forth in the complaint in this case. Medco's stock price had a history of volatility. Medco was essentially a one-product company (first Adenocard, then, as Adenocard began to fade, Adenoscan — it hoped), and its very right to market its one product depended on the vagaries of the bureaucratic process within the FDA as well as on the uncertainties inherent in the manufacture and sale of a new product in the rapidly changing, highly competitive pharmaceutical industry. A steep decline in the price of a stock cannot without more be considered evidence of fraud sufficient to start a statute of limitations running. ("Most losses occur without fraud of any kind." Eckstein v. Balcor Film Investors, 8 F.3d 1121, 1128 (7th Cir. 1993).) The significance of the decline depends on the circumstances, which in Tregenza supported an inference of fraud but in this case does not. Had the defendants in this case as in that one represented that the stock had no appreciable downside risk, that the FDA's approval was certain to be forthcoming on or before a particular date (cf. Pommer v. Medtest Corp., 961 F.2d 620, 622-23 (7th Cir. 1992)), or that Fujisawa's manufacturing operation was flawless, then the facts on which the defendants rely to show "inquiry notice" (the term used for knowledge of facts that would lead a reasonable person to begin investigating the possibility that his legal rights had been infringed) — the drop in the stock's price and the recall of the predecessor product — would be inconsistent with the defendants' representations. The inconsistency would suggest that those representations had been false, possibly deliberately so. The defendants did represent that the approval process in the FDA was going along smoothly, with approval expected by the end of 1992; and so it might have been. Yet the stock market might have anticipated even earlier approval, leading it to revalue the stock downward when approval was delayed, as in Wielgos v. Commonwealth Edison Co., 892 F.2d 509 (7th Cir. 1989). It was obvious that the defendants did not control the pace of the FDA's consideration and therefore could not guarantee approval by any date. Even a delay of a few months might greatly reduce Medco's earnings from Adenoscan. Competitors might be rushing a competing drug to market, as apparently had happened with Adenocard.

A fuller factual inquiry might of course cast the "critical facts" in a more ominous light. If for example the stock price of Medco's competitors rose or remained steady during the period when Medco's stock price was losing half its value, this might be a reason to believe that fraud was afoot — though an innocent alternative inference would be that, being essentially a one-product company, Medco was more at the mercy of the uncertainties of the FDA's protracted processes than its competitors were. And if on the contrary the stock price of Medco's competitors had fallen at the same time as Medco's price was falling, this would weaken the inference of fraud, since presumably the whole pharmaceutical industry wasn't practicing fraud at the same time. We do not know what the stock prices of other pharmaceutical companies were doing during the relevant period. Tregenza was decided on summary judgment. The plaintiffs had therefore had an opportunity to explore through pre-trial discovery the circumstances surrounding and preceding their discovery of the fraud. The present case was decided on the complaint. We are reluctant to lay down a rule that the conjunction of optimistic forecasts with a sharp drop in price establishes inquiry notice as a matter of law, but without such a rule the defendants cannot prevail on a record as barren as this. Even in Cooke v. Manufactured Homes, Inc., 998 F.2d 1256, 1262 (4th Cir. 1993), perhaps the strongest case for the defendants, the case proceeded as far as summary judgment, and the defendants won because they presented evidence not only of declining stock value and financial distress but also of a flood of negative publicity "questioning the financial integrity" of the company and indeed demonstrating that the company was experiencing "general financial failure" in the face of its rosy representations. Id. at 1259, 1262. It is the stark contrast between representation and (known) reality, as in Tregenza, Cooke, and other cases in which a securities suit was held untimely at an early stage in the proceedings, such as Davidson v. Wilson, 973 F.2d 1391, 1398, 1402 (8th Cir. 1992); Buder v. Merrill Lynch, Pierce, Fenner Smith, Inc., 644 F.2d 690, 692-94 (8th Cir. 1981); Koke v. Stifle, Nicolaus Co., 620 F.2d 1340, 1342-43 (8th Cir. 1980); Gaudin v. KDI Corp., 576 F.2d 708, 712 (6th Cir. 1978), and Rosenberg v. Hano, 121 F.2d 818, 821 (3d Cir. 1941), that is missing in this case.

The defendants urge us to affirm the dismissal of the suit on alternative grounds not reached by the district court. These grounds amount to attempts to nit-pick the complaint to death. The complaint is sufficient. The suit was dismissed prematurely. The judgment is reversed and the case remanded for further proceedings consistent with this opinion.

REVERSED AND REMANDED.