Appellant, Mesa Oil Company (Mesa) was the named beneficiary of a life insurance policy issued by appellee, Business Men’s Assurance Company of America (BMA), on the life of Cletus Chamberlain. The accidental death provision of that policy contained the following exclusion :

Chamberlain, the insured, died as a result of the combined effect of barbiturates and alcohol. BMA refused to pay the accidental death benefits, and Mesa brought this suit.

The District Court, assuming its diversity jurisdiction, awarded BMA summary judgment. The question presented is whether the meaning of the word “drug,” as used in the insurance policy, so clearly includes barbituric acid that there is no room for doubt on that point among reasonable men.

The policy fails to provide a definition or explanation of the term “drug.” Arizona law, which is, of course, here controlling, requires undefined insurance policy terms to be defined in the common everyday language of the average layman rather than in either technical medical terms or as statutes would define them. Malanga v. Royal Indemnity Co., 101 Ariz. 588 , 422 P.2d 704 (1967); Dickerson v. Hartford Accident & Indemnity Co., 56 Ariz. 70 , 105 P.2d 517 (1940). Further, undefined terms in an insurance policy are to be construed strictly in favor of the insured. United American Life Insurance Co. v. Beadel, 13 Ariz.App. 196 , 475 P.2d 288 (1970); Malanga v. Royal Indemnity Co., supra, and Brenner v. Aetna Ins. Co., 8 Ariz.App. 272 , 445 P.2d 474 (1968). Finally, if the undefined term appears in an exclusionary clause, an even stricter standard of construction must be applied. Brenner, supra.

*493 Mesa, after noting that there is no definition of the term “drug” in the contract, offers three dictionary definitions of the term in attempting to demonstrate that even these common sources are in disagreement over the term’s definition. 1 Mesa then contends that the absence of a policy definition, combined with the differences in dictionary definitions, required the court to deny the motion for summary judgment. We disagree.

In Malanga v. Royal Indemnity Co., supra, the insured decedent died from the combined effect of alcohol and barbiturates. The insurance company refused to pay the accidental death benefit to the decedent’s wife, relying upon the following terms of the insurance policy:

422 P.2d at 706 (emphasis in original).

As in the present ease, the key term (“bodily injury”) was not defined in the policy. The insurance company there claimed that bodily injury was limited to those cases wherein some cut, bruise, or rupture appeared on part of the body. The court declined to accept this limitation and ruled that the decedent’s death was so clearly included within the meaning of the term “bodily injury” that “ . . .we think it is impossible to conclude otherwise.” 422 P.2d at 707 .

In United American Life Insurance Co. v. Beadel, supra, the insured also died from the synergistic effect of alcohol and barbiturates. The insurer there involved also refused to pay accidental death benefits, claiming, inter alia, that an exclusionary clause in the policy denied recovery “if death results directly or indirectly . . . from any poison . . . accidentally or otherwise taken. . . ” 475 P.2d at 290 . The term “poison” was not defined in the policy.

The Arizona court, in determining whether the poison exclusion limited recovery, noted that the term poison, “[c]ommonly speaking, . . . conjures up such names as arsenic or cyanide.” 475 P.2d at 291 . The synergistic effect of alcohol and barbiturates, however, was held to be so far from the mainstream connotations of the term “poison” as to defeat the insurance company’s claim. Thus, the court held that “this accident is covered by the policy terms and is not excluded as poison.” 475 P.2d at 292 .

In the case at bar it is suggested that the undefined term “drug” is so ambiguous as to leave reasonable doubt whether barbiturates fall within its parameters. The trial court, however, was of the opinion that “anyone who has lived long enough to be exposed to the ‘boob tube’ for any length of time has a pretty good idea of what a drug is. * * * I believe that barbiturates are a ‘drug’ within the meaning of the policy.” Further, appellant’s own brief notes that “In the present case, the word ‘drug’ normally suggests to the average layman such substances as heroin, morphine, cocaine, etc.”

We are convinced, and believe that the Arizona courts would so hold, that barbiturates also constitute a par *494 adigm substance that the common layman would classify as a drug. The use of barbituric acid in the form of a depressant, often referred to as a “downer,” has become, unfortunately, extremely commonplace in our society.

Thus, while it may have been questionable whether barbiturates constitute a “poison,” or whether the effect produced by their use in combination with alcohol constitutes “bodily injury,” we think it is beyond dispute that any ordinary layman would consider barbiturates to be “drugs” within the meaning of the term.

In reaching this decision we are not unmindful of the requirement to construe undefined policy terms strictly in favor of the insured, nor of our duty to apply an even stricter standard of construction where, as here, the undefined term appears in an exclusionary clause. We are nonetheless of the opinion that the District Court’s dispositive conclusion below was correct. This opinion is reinforced by the proposition, which has been frequently emphasized, that “We are required to attach great weight to the district judge’s determination as to the law of the particular state in which he sits.” Insurance Co. of North America v. Thompson, 381 F.2d 677, 681 (9th Cir. 1967). See also Turnbull v. Josephine Bonkowski et al., 419 F.2d 104 (9th Cir. 1969); Owens v. White, 380 F.2d 310 (9th Cir. 1967).

Affirmed.

. Mesa’s brief sets out the following examples of dictionary definitions of the term “drug”:

. ABA Special Comm, on Crime Prevention and Control, New Perspectives on Urban Crime 25 (1972).