MELVIN MAYFIELD, Chief Judge.

Bill Reader was killed in an airplane crash on May 21, 1980, while working as a crop duster for Farm Air Corporation, an uninsured employer. His widow, Doris Reader, filed a claim in the Workers' Compensation Commission for dependency benefits and was awarded $126.00 per week.

The only issue on appeal is the amount of the award. It is Farm Air's contention that the deceased was a seasonal employee who had worked only three days the year he was killed and that the Commission's award based on earnings for that short period was improper.

The appellant operated a flying service to apply various substances to soybeans, wheat, and rice. The work generally began around the second week in May when they started applying herbicides and continued until the middle of September when they finished applying fertilizer to rice fields. Occasionally there was some work in the late fall applying defoliants. Reader had worked as a pilot for the corporation since its inception in 1978. He earned 25% of the gross charges for the work he performed.

In 1980, Reader started working on May 19 and had flown five jobs when his plane crashed. There was testimony that his commission on these jobs would have been $376.50. This did not include the bob he was flying at the time he crashed. The president of the corporation testified that he gave Mrs. Reader $300.00 for Mr. Reader's work on that job.

This case involves the application of the following portions of Ark. Stat. Ann. 81-1312 (Repl. 1977):

Compensation shall be computed on the average weekly wage earned by the employee under the contract of hire in force at the time of accident, and in no case shall be computed on less than a full time work week in the employment. . . . If, because of exceptional circumstances, the average weekly wage cannot be fairly and justly determined by the above formulas, the Commission may determine the average weekly wage by a method that is just and fair to all parties concerned.

It is clear that the law judge based the amount of his award upon the amount earned by Reader during the three days he worked in 1980, and the Commission adopted the law judge's finding. The appellant argues that this was unfair to appellant and that the Commission should have applied the last sentence of the section to make a determination of Reader's average weekly wage by a method that was "just and fair" to all parties.

The appellant, however, is not very specific in its suggestions as to the method that would be "just and fair Implied in its argument is the contention that whatever amount of earnings is used — whether those of the previous year or only those earned during 1980 — the amount should be spread out over a full year. In other words, if earnings for 1979 are used, the appellant would find the average weekly wage by taking the total amount that was earned in the approximately four months worked in 1979, and divide that amount by 52; and, if only the earnings for the three days in 1980 are used, the appellant would convert that into the amount that would have been earned in the four months that Reader expected to work in 1980, and would divide that amount by 52.

The only cases cited by appellant are Travelers Insurance Co. v. Perry, 262 Ark. 398, 557 S.W.2d 200 (1977) and Ryan v. NAPA, 266 Ark. 802, 586 S.W.2d 6 (Ark.App. 1979), neither of which is really on point. In Travelers the employee was not even a seasonal employee. At times he made himself available for work to Manpower, Inc., an organization that provided workers for employers who needed temporary help, but he had no contractual arrangement that required him to work or to be available for work. In Ryan the employee was a four-hour per day worker whose claim that she was required to be available for a full eight hours was rejected.

The appellee defends the Commission's award in this case on the basis of assumptions which it says the evidence would support. Thus the appellee argues that we should affirm because the Commission's decision is supported by substantial evidence. It is clear, however, that the appellee agrees with the Commission's refusal to spread Reader's seasonal earnings over a 52-week period and use that as the basis for the calculation of average weekly wage.

We recognize that there is force in appellant's argument, but we have concluded that the Commission's award should be affirmed. Our decision is based upon the following considerations.

First, we agree with the appellant's contention that the Commission could have used a better basis in making the determination of the employee's average weekly wage. Our problem, however, is the same problem that the appellant had in suggesting what method the Commission could have used under its "just and fair" authority that would be of help to appellant. The president of the appellant corporation testified that Reader was paid $13,101.68 for his work in 1979, that Reader started working around the first week in May that year, and that the flying season ended the last of August or first of September. There were 123 days in the full four-month period involved, of which 17 were Sundays, and since they did not work Sundays, that leaves 106 working days. Dividing 106 into $13,101.68, we get $123.60 per day. Using the six days worked per week, we get an average weekly wage of $741.60. It is conceded that Mrs. Reader is entitled to benefits of 35% of the average weekly wage, Ark. Stat. Ann. 81-1315(c) (Repl. 1977), subject to the maximum weekly benefit of $126.00, Ark. Stat. Ann. 81-1310(b)(A) (Supp. 1983). Since 35% of $741.60 would be greater than the maximum allowance, it is of no help to appellant for the Commission to base the average weekly wage on the 1979 earnings unless, as appellant has suggested, these earnings are spread out over the entire year.

At this point we are faced with the second reason that impels our affirmance. The appellee cites and relies upon Gill v. Ozark Forest Products, 255 Ark. 951, 504 S.W.2d 357 (1974). The employee in that case seldom worked a full five-day week. This was because of the nature of the timber industry in which he was employed. His work agreement was for a five-day week whenever work was available, but his employer made work available based upon weather conditions and the timber supply.

In determining the employee's average weekly wage, the Commission used the previous 52-week period, but ignored one week in which the employee worked only one day and 18 weeks in which he worked only two or three days per week. The Commission then used the balance of 33 weeks, in which the employee worked either four or five days per week, and calculated the average weekly wage on the earnings for those 33 weeks. On appeal the Arkansas Supreme Court held that the average weekly wage should be based on a full week's pay.

The statutory law in effect during the period involved in the Gill case is still in effect now. While the factual situation there was different from the factual situation in the case at bar, the principle is the same. Obviously, the court in Gill did not think there were exceptional circumstances which required the Commission to devise some "just and fair" method to replace "the average weekly wage earned by the employee under the contract of hire" as the basis on which compensation benefits should be computed. In the present case we do not approve of the method used by the Commission. It used less than a "full time work week" contrary to the statute, and even if there had been a full work week to use, there probably still would have been exceptional circumstances to require the employee's 1979 earnings to be used in determining the average weekly wages. We do not, however, think there are exceptional circumstances in this case that would make it just and fair to both parties to take the wages earned in 1979, divide that amount by 52, and use that figure as the average weekly wage.

In discussing average wage computation, 11 Schneider, Workmens' Compensation Law 2175 (perm. ed. 1957), states:

The computation of such average wage, under particular fact situations, often becomes difficult in one's effort to arrive at a result just and fair to both employee and employer. It is generally considered contrary to that concept and against public policy to so compute an employee's wage that it will result in a compensation award that pays the employee more during his period of disability than he is accustomed to earn in his usual or normal year around activity.

In the instant case, $126.00 per week for 52 weeks will pay this widow $6,552.00 per year. This is certainly not mole than her husband was earning at the time he was killed. So while we do not approve of the method used by the Commission in the determination of the average weekly wage of the deceased employee, we see no other method, which we think would be just and fair to both parties, that would make any difference in the amount of benefits to be paid the appellee.

A third consideration in our decision to affirm is the fact that our Workers' Compensation Act contains no specific provision dealing with seasonal employment. To hold that the 1979 earnings have to be spread out over the whole year for the purpose of determining the employee's average weekly wage, it seems to us, would go further than the use of a more appropriate basis to compute the wage, and would, in this case, be tantamount to legislating policy in regard to seasonal employment.

Finally, we note the statement in Gill, supra, that "it is well established in workmen's compensation cases that when doubt exists we must remember the Workmen's Compensation Act is remedial and should be construed liberally to effectuate its purpose."

Based upon the above considerations, we affirm.

CLONINGER and CORBIN, JJ., agree.