¶1James E. Kassis (argued), of Pease, Mayhew & Kassis, Sacramento, Cal., for appellant.
¶2William H. Shubb (argued), Asst. U. S. Atty., John P. Hyland, U. S. Atty., James H. Daffer, Asst. U. S. Atty., Sacramento, Cal., for appellee.
¶3Before MERRILL, DUNIWAY and HUFSTEDLER, Circuit Judges.
¶5Appellant Smith, having suffered injuries while undergoing tests on behalf of the Government, seeks recovery under the Tort Claims Act, 28 U.S.C. § 2674. He has taken this appeal from judgment of the District Court rendered in favor of the Government.
¶6The United States has not seen fit to accept strict liability for injuries incurred by those voluntarily engaging in tests in the public interest. The Government’s *792liability is limited to cases: of negligence. Dalehite v. United States, 346 U.S. 15, 44-45, 73 S.Ct. 956, 97 L.Ed. 1427 (1953). The sole question on appeal is whether appellant has established negligence on the part of the Government as matter of law. We conclude that he has not.
¶7Appellant, while in flight school training in 1940, voluntarily submitted himself to the tests of the Pensacola Study of Naval Aviators. The Study involved follow-up testing in subsequent years, and in 1963 appellant went to Pensacola for the purpose of taking further tests. One of the tests was a postural equilibrium test that involved walking heel-to-toe on a three-quarter inch rail placed two-and-one-half inches above the floor, with hands folded across the chest. The idea was to see how far the testee could walk the rail before losing his: balance and stepping off. The testee was given five trials and was scored on his three best efforts.
¶8Appellant tried once, lost his balance and stepped off. He tried again, lost his balance and fell injuring himself, for which injuries he has brought this suit. Seven hundred forty persons had taken the test up to 1963. Until appellant's fall no one had fallen.
¶9The District Court found no negligence. We hold that this finding was not clearly erroneous. We agree with appellant that those in charge of the tests should have anticipated that he would lose his balance. They did so anticipate. The tests contemplated that all those tested eventually would lose balance. The question was not whether, but when. We disagree with appellant’s contention that the Government should have foreseen that having, not unexpectedly, lost his balance he would be unable safely to step down from a height of two-and-one-half inches.
¶10Judgment affirmed.
¶12I regret that I cannot concur in the majority opinion.
¶13Were this ease one in which the relationship between the parties was that of a business invitor to his invitee, there would be some support (but not enough) to sustain the District Court’s judgment. But that was not the relationship between the parties.
¶14Appellant was seriously hurt while performing physiological tests for the Government. He was not a Government employee, and he received no compensation for his participation in the tests, other than reimbursement for his travel expenses. The scene of the injury was not the market place, but the laboratory. The object of his visit was not commerce, but research upon human beings. The benefits to be derived from the relationship were not mutual, but unilateral. Surely the duty owed by the Government to a human volunteer upon whom the Government performs nontherapeutic tests for its sole benefit is sterner than the duty owed by a grocer to his patron. Yet the District Court held the Government to less than the grocer’s standard, and the majority affirms.
¶15As I read the majority opinion, the District Court’s conclusion that the Government was not negligent is sustained on the theory either that the test involved no risk of personal injury, or, if there were some risk, that the Government did not foresee it in fact and was not obliged to foresee it in law. I disagree with both.
¶16At the time of the injury, appellant was 46 years old. He was required to walk heel-to-toe along a narrow rail raised 2% inches above the floor. As part of the test he was expected to lose his balance, and to help him to do so, he was required to keep his arms folded to his chest during the walk. Can it be truly said that a test designed to make one lose his balance does not entail a risk that the testee will fall? Suppose a grocer knowingly left a can in the patron’s aisle of the market. Would we heed his argument that he was not negligent because, though he knew someone might stumble over it, he did not anticipate that a customer would fall? Or, indeed, that he should not be found negligent because 739 people had *793stumbled over it, but not one of them fell until the plaintiff came along ?1
¶17Florida substantive law controls this case. Florida has not decided a case on the precise facts here, and neither has any other forum. But Florida does apply established principles of negligence, which are flexible enough to meet this new challenge.2 Florida agrees that the definition of the duty which one person owes another varies with the nature of the relationship between them and the degree of risk which is involved by the conduct in which the actor is engaged. (See, e. g., Vanlandingham v. Florida Power & Light Co. (1944) 154 Fla. 628, 18 So.2d 678.)
¶18I have little difficulty in concluding that Florida would hold that one who subjects volunteers to nontherapeutic tests involving some risk of personal injury has, at a minimum, the duty fully to inform the volunteer about the tests 3 and to secure his informed consent and the further duty to take all reasonable precautions commensurate with the risk to prevent injury. There is no evidence which would have sustained a finding (and none was made) that the Government performed either of those duties. There was no evidence that the Government took any precautions to prevent injury to one whom it had to anticipate might fall as he lost his balance. There was no evidence that safeguards were not feasible.
¶19The judgment should be reversed and the cause should be remanded for a new trial.