This suit was tried in the court below as a test case to determine whether an insurance policy covering loss caused by hail includes loss caused by sleet. The Court of Appeals certified the case to us as presenting an issue of public interest, because a number of other pending cases may be affected by the decision. Rule 29 (4) (b).
The plaintiff Southall, the insured, introduced proof to show that on January 11,1978, sleet fell all day long and into the night. The next morning Southall found that the accumulation of about four inches of sleet on top of his chicken house had caused it to collapse. In this suit upon the policy the court instructed the jury that the term “hail” as used in the policy should be given the meaning ordinarily applied to that term in the everyday affairs of life. The jury’s verdict was for the insurance company.
The extended-coverage section in the policy provides that “the coverage of this policy is extended to include direct loss by . . . hail. ...” Later in the same section is this limitation: “Provisions Applicable Only to Windstorm and Hail: This Company shall not be liable for loss caused directly or indirectly by (a) frost or cold weather or (b) ice (other than hail)
At the trial there was undisputed testimony that the word “hail” includes large hail, small hail, winter hail, and sleet. A witness employed by the Weather Service testified that hail is so defined in á recognized meteorological dictionary. Webster’s Second New international Dictionary *60(1939) is to the same effect, subdividing hail into summer hail and winter hail. In a case directly in point it was held that a policy covering loss by hail includes loss by sleet. Evana Plantation v. Yorkshire Ins. Co., 214 Miss. 321, 58 So. 2d 797 (1952). No case to the contrary is cited. In fact, at the trial the appellee’s vice-president conceded on the witness stand that the sleet on the roof of the plaintiff’s chicken house was small hail.
The trial court was in error in instructing the jury that . the term “hail” was to be given its everyday meaning. An insurance policy is to be construed strictly against the insurer, who chooses its language. The construction and legal effect of written contracts are matters to be determined by the court, not by the jury, except when the meaning of the language depends upon disputed extrinsic evidence. Ark. Rock & Gravel Co. v. Chris-T-Emulsion, 259 Ark. 807, 536 S.W.2d 724 (1976); Security Ins. Co. v. Owen, 252 Ark. 720, 480 S.W.2d 558 (1972). In the interpretation of a contract negotiated between individuals no doubt there might be conflicting testimony presenting an issue of credibility for the jury with respect to the meaning of the language used, but there was no such issue in this case. The court should, if appropriate, have instructed the jury that the word “hail” as used in the policy included sleet.
The appellee’s vice-president testified that he construed a “direct loss” by hail to mean only the damage caused by the initial impact of the hail, but we have held that a direct loss is one proximately caused by the hazard insured against. Farmers Union Mut. Ins. Co. v. Blankenship, 231 Ark. 127, 328 S.W.2d 360 (1959). If the weight of the hail damaged the chicken house, the hail was the proximate cause of the loss, absent some other possible cause. It is also argued that the policy did not cover loss caused by ice, but that argument is refuted by the language we have quoted, that the company shall not be liable for loss caused directly or indirectly by ice other than hail.
The j udgment must be reversed, but the record does not support the appellant’s argument that we should enter judgment here. There may be an issue of proximate causa*61tion to be submitted to the j ury upon a retrial, because it does not appear to be an undisputed fact that the collapse was caused by an accumulation of sleet rather than of snow.
Reversed and remanded.
Purtle, J., dissents.