Plaintiff corporation sues on three identical insurance policies on the life of its late president, Paglia. The single indemnity was paid, but defendant resists the double indemnity, though, for reasons that we do not understand, apparently conceding the death was accidental within the policy terms. The policies provided,

The Company agrees to pay ... [if] the death of the Insured has resulted directly and independently of all other causes from bodily injury caused solely by external, violent and accidental means ... and did not result from a risk stated herein as not assumed.

Death either directly or indirectly resulting from or contributed to by any of the following is a risk not assumed under this rider: (a) disease or bodily or mental infirmity or medical or surgical treatment therefor.

The facts, as developed at the trial, were these. Paglia, in the winter of 1986, applied for additional insurance, and three urine tests showed abnormal protein therein, a condition known as proteinuria. Tested by his own physician, he showed 775 milligrams where normal would be 0-100. His physician recommended an intravenous pyelogram, a diagnostic procedure to evaluate the kidneys based on tracing an injected dye. When this test was performed Paglia suffered a fatal reaction to the dye and died.

Stopping here, although some courts permit recovery under such a policy on such facts, holding that an unexpected result is an accident, see cases collected in 10 M. Rhodes, Couch on Insurance 2d, (Rev. ed.) Secs. 41:112 et seq. (1982), Rhode Island adopts the rule that gives weight to the entire policy language, and holds that "external, violent and accidental means" defines the means, and not simply the result. The policy insures against accidental physical acts, not merely unexpected results that might, colloquially, be called accidental. The only external means here was the administration of the dye, and this was fully intentional.

If counsel had called the court's attention to Kimball v. Massachusetts Accident Co., 44 R.I. 264, 117 A. 228 (1922), that should have been the end of the case. Instead, defendant went to trial on the Risks Not Assumed clause. Even here the court ultimately directed a verdict for defendant, and plaintiff appeals. We affirm.

The special facts relating to the Risks Not Assumed clause were these. An autopsy performed to ascertain the cause of death disclosed that Paglia's right kidney was small and blocked, and functioned poorly, if at all. Plaintiff's expert at the trial, however, testified that the left kidney had enlarged, to compensate, and that the urinary system as a whole, unless possibly for the proteinuria, was fully functional. Over objection, the expert testified that the hypoplastic kidney itself was neither diseased, nor an infirmity. Whether this was an impermissible attempt to contradict plain and accepted meanings might be an interesting question. We need not, however, reach it, but will concern ourselves with the proteinuria.

First, we return to the Risks Not Assumed language. Plaintiff says that a purely diagnostic procedure was not a "medical or surgical treatment." The district court held otherwise. See McKay v. Bankers Life Co., 187 N.W.2d 736 (Iowa 1971). If the policy is construed sufficiently seriously against the insurer, plaintiff may be correct. Is not inquiry commonly thought of as something that precedes a decision to treat? Do you "treat" something before you know what it is? Again, however, we need not reach this question.

Where plaintiff comes up against a hard place is a different part of the policy language: "... directly or indirectly ... contributed to by ... disease." If proteinuria was, or indicated, a disease, can it be thought that its existence and the symptoms it evinced, did not contribute, either directly or indirectly, to Paglia's doctor's submitting him to the dye, regardless of whether its administration constituted "treatment?" Indeed, the proteinuria would seem to have been the true activating cause of the diagnostic procedure.

Plaintiff points to its expert's testimony that proteinuria could be entirely benign. If this were the fact, the decision should be for the plaintiff: the diagnostic procedure had not been caused by disease. The difficulty is that the doctor conceded that there could have been disease, and that he could not say here because there had been no testing.1 Here plaintiff runs into the Rhode Island law to the effect that the burden was on it to disprove contribution by a disease. In Renault v. John Hancock Mutual Life Ins. Co., 98 R.I. 213, 200 A.2d 588 (1964) the insured had a diseased kidney and suffered an automobile accident; the strain on his urinary system caused something to break. A medical expert said that he was unable to answer the question whether death would have occurred but for the kidney trouble. Held, that the plaintiff had not sustained her burden of proof. In holding for the defendant the court said,

It was plaintiff's burden to prove by a fair preponderance of the evidence that death was caused directly, independently and exclusively of all other causes, by a bodily injury sustained solely by external, violent and accidental means, and did not result directly or indirectly, or wholly or partially, from any bodily or mental disease or infirmity.

Equally, in the case at bar, it was plaintiff's burden to exclude proteinuria as being or indicating disease. A disease does not have to be diagnosed to qualify. See Palumbo v. Metropolitan Life Ins. Co., 296 Mass. 358, 5 N.E.2d 836 (1937). If the doctor could not tell whether there was a disease or was not, manifestly the jury could only speculate, which can not be enough. Margo Lynch v. Merrell-National Laboratories, 830 F.2d 1190 (1st Cir.1987). We believe our dissenting brother misreads the testimony. The verdict was correctly directed.

Affirmed.

COFFIN, Circuit Judge, dissenting.

I think the jury should have been permitted to resolve the question whether Paglia's proteinuria was a disease or bodily infirmity. Although plaintiff's expert, Dr. Davis, could not rule out this possibility, it was nevertheless his opinion that the proteinuria was not a disease or infirmity. Asked whether Paglia's proteinuria in any way impaired his bodily functions, Dr. Davis replied, "I have an opinion based on the autopsy report because I think there's substantial evidence that except for hypoplasia, he had no renal infirmity." (As the insurer had never asserted that the proteinuria resulted from anything other than renal dysfunction, the jury reasonably could have understood Dr. Davis as having ruled out proteinuria as an infirmity of whatever sort.) Dr. Davis also testified that in his opinion Paglia did not have any bodily infirmity that directly or indirectly contributed to his death. This, in my view, was enough to create a jury question.

Nor am I able to concur in the result based on the panel's suggested alternative ground: the distinction between "accidental means" and "accidental results." An increasing number of jurisdictions are rejecting this distinction as artificial and confusing. See 10 M. Rhodes, Couch on Insurance 2d, Sec. 41.31 (Rev. ed. 1982) (collecting cases). Interpreting the two terms as synonymous "is clearly the preferred construction of policy language." Id. Sec. 41.29 at 45. This issue has never been raised by the parties to this case, either below or on appeal. I think it would be unfair for us to base our decision on this ground without first having afforded the parties an opportunity to argue whether Rhode Island, which has not considered the issue since 1922, would follow the trend and abolish the distinction.