<< As a rule of thumb diesels reject more heat to both the coolant and the oil
 than do spark-ignition engines.  >>

Heat rejection is far more of a function of specific fuel consumption, EGT,
and the exhaust placement vs. the cooling flow than any comparison of engine
type.  I can definitely assure you that a TIO-540U2A at 263 HP rejects far
more heat and requires far more cooling than a modern 300 HP compression
ignition engine at the same power simply because of the difference between a
.59 SFC and the C-I's .36.  For given amount of output HP, the lower the SFC
the lower the rejected heat.  

Since C-I requires high compression ratios to work, this leads to higher
expansion ratios and lower EGT's.  Lower EGT also leads to lower cooling
requirements so the nod goes away from the S-I engine in that area.

If the exhaust is placed close to the intake or in the way of the cooling
flow to vital components, more air (and drag) must be used to keep things
cool.  So arranging the engine in the fashion of the TSIO-550N is better than
the TIO-540 U2A.  Best of all (for opposed engines) is the TIO-540S1AD with
updraft cooling, intake on the bottom, and exhaust on top;  this allows
convection to aid cooling instead of fight it.

A C-I engine will reject a higher percentage of its waste heat to the coolant
and oil that an equivalent S-I engine due to the low amount of heat rejected
in the exhaust.  However, modern S-I engines are so much more efficient than
our current Lycoming and Continental offerings that the cooling load would
likely be less for a properly designed aircraft C-I engine with an SFC of .36
or less.

Whether any given cooling load equates to less cooling drag is far more a
function of system design, and the GA aircraft industry is rife with bad
system design.

Eric Ahlstrom