dissenting. The opinion of the majority effectively does away with the requirement that an injury arise out of employment to be compensable. The twin tests for a workers’ compensation award are that the injury 1) arise out of employment, and 2) occur in the course of employment. Ark. Code Ann. §§ 11-9-102(4); 1 l-9-401(a)(l) (1987). The majority correctly concludes that Earl Angus’s injury occurred in the course of employment. I cannot agree, however, that his injury, which was caused by a tornado while he was not performing work duties and was eating supper at his home, arose out of his employment.
Angus’s mobile home was on the premises of his employer, where he was the manager, but he worked in a separate office. That office was not destroyed by the tornado. He had left his office at 9:15 p.m. to return to his home to have supper. That is what he was doing when the tornado struck. There is no question that he had quit work for this period of time, though he knew that a trucker named Billy Harris would be coming in sometime later that evening:
Q. Was Mr. Harris working that day?
A. Yes, he was loaded that morning. I was in the process of cooking oil that night and I knew Billy wouldn’t be in until late, so I went home to eat and wait for Billy to get back in and load him when he got back in that night.
Q. What time was this when you quit?
A. It was about 9:15 when I got home.
Q. And then what happened?
A. I was talking to my wife and I heard some loud pops and noises and I had heard a transformer explode before, the lights were flickering and I heard the roar and everything. I grabbed my wife and my child and we started to get out of the trailer and everything exploded. That is all I remember.
Q. But, before the tornado demolishing your quarters there, would you have had work to do that night?
A. I had to load Billy when he come back in. The tornado then hit at about 9:30.
Angus testified that normally he went to work at 6:45 in the morning and would be finished at about 7:00 or 8:00 p.m. His typical routine was to go to the house trailer after 7:00 or 8:00 in the evening and then return to his office after supper to turn the alarm system off and perform other duties. When the tornado hit, Angus admitted that he had gone home and was not performing any duties for his employer while he was eating supper with his family:
Q. But you were not performing of (sic) your employment at the time you were in the trailer?
A. Well, not any duties, no, but my work wasn’t done until Billy got back in.
Q. So, you might have gone back in the yard to help Billy when he got back in, but as far as what you were doing in the trailer, you testified that all you were doing was eating.
A. Eating dinner.
Q. And you were staying there with your family?
A. Right.
Q. And that is when the tornado hit?
A. Correct.
In sum, we have a neutral risk — the tornado — unrelated to Angus’s employment that caused his injury while he ate supper in his home during a period of time when, by his own admission, he was not performing duties for his employer. The fact that a truck would be arriving later that night and that he would have to leave his home and perform some duties does not convert Angus’s injury by the tornado into an injury arising out of his employment. He might have been operating in the course of his employment because his mobile home was on his employer’s premises and he was on call, but the injury did not arise out of it.
The majority, accordingly, stretches to reach this result. In my judgment, this court was correct in Parrish Esso Service Center v. Adams, 237 Ark. 560, 374 S.W.2d 468 (1964), which also involved a wind-related injury. In that case, Adams was employed as a night manager of a service station, and at the time in question he was the only person working on the shift. At about 3:30 a.m., Adams decided to go fasten and secure items on the service station islands. It was while performing these duties that he was injured by a gust of wind. We stated in that case:
Certainly, there was a duty upon Adams, as an employee, to protect the property of his employer, and the protection that Adams was seeking to afford, could not have been done without leaving the building. The acts being performed were as much a part of his duties as though he had been waiting on a customer when the wind struck. There is absolutely no evidence that Adams was not engaged in the work that he testified to at the time the injury was sustained.
237 Ark. at 568, 374 S.W.2d at 473. Those facts are substantially different from the facts in the present case and exhibit a clear injury arising out of employment. Adams was unquestionably performing work-related duties when he was injured. The same cannot be said in the case before us.
We granted review of this case because of a three-to-three division in the Court of Appeals sitting en banc. The evenly divided court amounted to an affirmance of the Workers’ Compensation Commission’s decision. Did the origin and cause of the injury emanate from Angus’s employment? I do not think so. Angus did not perform work-related duties twenty-four hours a day at Deffenbaugh Industries, and he was not working at the time the tornado struck. Indeed, he had left his office and was in his mobile home having supper. The mere fact that he lived on the premises cannot sustain the requirement of work-related causation and this award. Had the tornado injured him when he was in his office or assisting Billy Harris unload his truck, I would have no hesitancy in affirming the award. Here, though, that is not the case. I would reverse the Commission’s decision.