Re: Deleting bytea, autovacuum, and 8.2/8.4 differences

There does not appear to be much WAL activity.  Here's the insertion of 100 rows as seen by iotop:
   4.39 G  0.00 %  9.78 % postgres: writer process          
   5.34 G  0.00 %  5.93 % postgres: postgr~0.5.93(1212) idle
  27.84 M  0.00 %  1.77 % postgres: wal writer process      
 144.00 K  0.00 %  0.00 % postgres: stats collector process 
   0.00 B  0.00 %  0.00 % postgres: autova~ launcher process
   0.00 B  0.00 %  0.00 % postgres: postgr~0.5.93(4632) idle
                                                            


.. and the deletion:
288.18 M  0.00 % 37.80 % postgres: writer process          
  3.41 G  0.00 % 19.76 % postgres: postgr~0.5.93(1212) DELETE
 27.27 M  0.00 %  3.18 % postgres: wal writer process      
 72.00 K  0.00 %  0.03 % postgres: stats collector process 
  0.00 B  0.00 %  0.00 % postgres: autova~ launcher process
  0.00 B  0.00 %  0.00 % postgres: postgr~0.5.93(4632) idle

So, the original 1.9 GB of useful data generate about 10GB of IO, 5 of which end up being written to the disk  The deletion generates about 3.8 GB of IO all of which results in disk IO.  WAL activity is about 27MB in both cases.
 

I read all of the above, but it does not really clarify why deletion generates so much IO.