McCluskey in 1918 while in the Army had influenza followed later by pneumonia, and bronchitis. His War Risk term insurance lapsed on his discharge in April, 1919. In 1921 he applied for compensation, and was found partially disabled, and his past due compensation kept his insurance in force till August IS, 1921. After that he farmed by renting land and hiring croppers and sub-renting, and by this and his compensation allowance he supported himself and mother till 1929, but he did no manual labor. In 1929 he entered a veteran’s hospital and was treated for pulmonary tuberculosis, active, moderately advanced, and after about four years was discharged, supposedly cured. On May 27, 1931, while in the hospital he made his first claim for insurance benefits, and his claim was denied September 16, 1932. After his discharge from the hospital in 1933 his disease became again active. His suit for insurance, filed in 1932, came to trial in 1941, when the jury gave verdict in his favor, the judge having refused the motion of the United States for a directed verdict. This ■appeal followed.

The sole question argued is whether there is substantial evidence that on August IS, 1921, the veteran was permanently disabled because of tuberculosis. His delay for ten years to make claim puts á heavy burden on him. Lumbra v. United States, 290 U.S. 551, 54 S.Ct. 272, 78 L.Ed. 492; United States v. Spaulding, 293 U.S. 498, 55 S.Ct. 273, 79 L.Ed. 617. While there was no positive diagnosis of pulmonary tuberculosis until 1929, Dr. Ellis suspected it in 1921, though he specified then only bronchitis, and his testimony and the symptoms testified to warranted submission to the jury of the question whether at that time McCluskey had active tuberculosis. •For present purposes we assume he did. Complete rest is, under the medical testimony, a prime essential in the treatment. Dr. Ellis prescribed rest, but complete rest *510was neither prescribed, nor taken. Since any exertion is harmful, it is of course true that active tuberculosis, so long as it is active, is totally disabling. But establishing the existence of active tuberculosis on Aug. 15, 1921, does not show a permanent disability at that date, because under the testimony, with prompt and proper treatment pulmonary tuberculosis can in a majority of cases be arrested and the patient enabled to resume work of a substantial sort. Not until the disease in its earlier stages has been treated in vain long enough to show that recovery is not likely in the particular case, can permanency be said to exist. Of course cases may occur in which the tuberculosis has gone so far in lung destruction when discovered that recovery is at once seen to be unattainable. But such is not this case, for no certain diagnosis was made at all till 1929, and then it was put down “Moderately advanced”, was treated not only with hopefulness, but with temporary success. The fact that this patient has, after long treatment therefor, finally not been cured, very well proves that he cannot now be cured, but it does not prove he was incurably diseased in 1921. The Treasury definition of permanency, coeval with the insurance, made by Regulation, T. D. 20 W. R., is “Total disability shall be deemed to be permanent whenever • it is founded upon conditions which render it reasonably certain that it will continue throughout the life of the person suffering from it.” The existence of active tuberculosis in August, 1921, is not enough, for that condition does not make it reasonably certain-that the disease will disable for life. Another condition must be added: either a period of unsuccessful treatment sufficient reasonably to show in-curability at that time, or proof that the ravages had already gone, too far for the sufferer’s working ability to be restored. Neither of these latter conditions was established or sought to be as of August, 1921. Reliance is put on the fact that the disability has now turned out to be permanent. Both Dr. Ellis and Dr. Donaldson, on whose testimony this part of the case must rest, say they are of opinion that the active tuberculosis existed in August, 1921, and that from the case history both before and since that date to the trial in 1941, “it is reasonably probable the condition will continue during the remainder of his natural life.” That is not saying that conditions were such in August, 1921, that it was “reasonably certain” that disability would continue for life. Under such circumstances as this, where one becomes totally disabled by a curable disease and the permanence of it is-doubtful, he must pay premiums until such time as his condition changes the doubt into a reasonable certainty that the disability will be lifelong; and then it becomes the duty of the insurer to begin paying. United States v. Brewer, 5 Cir., 97 F.2d 899.

We are not saying that the history of the insured’s health subsequent to the date of lapse may not be proved, to throw light on what disease he really had and what his condition really was at that date. Later developments may show much that was at first hidden. We are saying that all the evidence here shows is that if McCluskey was totally disabled in August, 1921, it was because of a curable disease, tuberculosis of the lungs, and none of it shows that at that date his condition was such as to render it reasonably certain he could not recover. United States v. Bryan, 5 Cir., 82 F.2d 784. Since another trial with a better understanding of the issue may possibly produce clearer evidence, we award a new trial. The judgment is reversed and the cause remanded for further proceedings consistent with this opinion.