Early work of Ricardo is described, in which squish is used in flat-head en
gines to generate turbulence levels comparable to those in overhead-valve e
ngines, leading to rapid flame propagation, and suppressing knock. Work by
NACA before World War II is described, in which turbulence levels were meas
ured in overhead-valve engines, indicating indirectly that surprisingly hig
h levels were achieved just before ignition, possibly due to a tumble insta
bility. Finally, work of Obukhov of 30 years ago is described, in which ins
tabilities of tumbling flow are investigated in ellipsoids crudely modeling
the engine cylinder as the piston rises; this suggests that there is an in
stability leading to intense small-scale motion just before ignition. Sugge
stions for further work are given.