The average of the 4 synoptic temperature 
readings was not calculated and recorded until April '96.  In monthly 
reports issued from Palmer prior to that time, the reported "average" 
was a calculated column in a spreadsheet that was just an average of 
the minimum and maximum temperatures.