Like last week's environment variable, this week's variable also provides a different value each time it's referenced.  The value stored in SECONDS indicates the number of seconds since the shell was invoked:

$ print $SECONDS; sleep 30; print $SECONDS
8
38
$

If a value is assigned to SECONDS, the sum of that value and the number of seconds since its assignment is displayed when the variable is referenced:

$ SECONDS=10; print $SECONDS; sleep 30; print $SECONDS
10
40
$

One possible use for this variable is to time how long a particular operation within your shell script takes to complete.  For example, you may want to know how long it takes to process a unique set of data during each pass through a loop construct.  The following  script illustrates this concept:

#!/bin/ksh


FILELIST="file1 file2 file3"

sleeptime=5

for i in ${FILELIST}
do
  # record the value of SECONDS before processing begins
  start=$SECONDS

  print "Processing $i..."

  # Normally, processing statements would replace
  # the following sleep statement
  sleep ${sleeptime}

  # record the value of SECONDS after processing ends
  stop=$SECONDS

  # calculate processing time
  (( runtime=stop-start ))
  print "Processing time for $i was $runtime seconds.\n"

  # increase the sleep time for next iteration of the loop
  (( sleeptime=sleeptime*2 ))

done

exit 0

$ ./seconds
Processing file1...
Processing time for file1 was 5 seconds.

Processing file2...
Processing time for file2 was 10 seconds.

Processing file3...
Processing time for file3 was 20 seconds.

$