The great majority of present-day motorcycle engines have plain sleeve bearings of the type adopted by Honda on their 1969 CB750 four-cylinder. In such bearings oil is pumped into a low-pressure region of the clearance between a cylindrical journal and its supporting sleeve or bearing shells. The oil is then swept by a combination of its own viscosity and the relative motion between journal and bearing into the loaded zone, where pressure can rise as high as several thousand kilogram per square centimeter. The minimum clearance in this loaded zone can be of the order of 2.5 microns, which is why the surface finish of the journal must be extremely smooth, and the oil very clean.
The earliest internal combustion engines operated so slowly that railroad-derived lubrication pratice was sufficient. Bronze sleeves supported crank journals, lubricated by grease cups that were given a twist every little while. When engines turned faster, crankshafts had to be fully enclosed to prevent the lubrication from being thrown out by parts motion. A half-cup of oil in a fairly close-fitting crankcase was tossed about by the crank, lubricating the rod big-end directly, lubricating the main bearings by drain-back, and splashing up to lube piston and cylinder.
As oil passed the simple piston rings of the time and was consumed by combustion, more was needed. This was made up 'on the fly' by the rider, administering a stroke of the hand pump, which transferred oil from a tank into the crankcase. The rule was to look back. If you didn't see smoke, you stroked the pump.
When engines turned faster yet, auto engineers began to adopt circulating oil systems, but the motorcycle was at first too basic a device for such sophistication. It was given rolling-element bearings whose need of oil was minimal. This allowed the total loss system to persist into the mid-1920s. The engines of racing cars likewise adopted rolling-element bearings – under the intuitive but wrong assumption that rolling must always generate less friction than sliding. By 1927 Fiat's engineer Tranquilo Zerbi was ready to adopt plain babbit bearings in his 406 Grand Prix engine.
No less a person than Harry Ricardo had in 1923 enunciated that operation above 4000rpm would not be possible with babbit bearings. That opinion was partly the result of poor big-end lubrication – usually provided not directly by pump but by collecting oil thrown off of main bearings, then channeling this through drilling to the crankpins. But as full pressure lubrication was adopted, cooler operation resulted, as the ability to deliberately push extra oil through bearings carried away the heat of viscous friction. In the four years from 1923 to 1927, crankshaft lubrication evolved sufficiently to allow Fiat's 406 engine to operate at 8500rpm.
Meanwhile there had been many problems with roller big-end bearings. Rollers skidded, ploughing the oil off crankpins, then overheated and seized. Roller cages fatigued under the battering of large rollers - and broke. Without cages, centrifugal force pinched the rollers against each other, generating extra heat. At Royal Enfield, Tony Wilson-Jones had expressed misgivings about roller big-ends as early as 1929.
In the early 1920s in the US, a collaboration between the Army's air development center at McCook Field, Ohio, and the Allison Division of General Motors developed a superior plain bearing material – copper-lead. A strong matrix of copper contained regions of lead. The harder copper supported the load, while the lead provided seizure resistance and the soft embedability required to immobilize foreign particles. Bearings of this type would be used to carry crank and rod loads in Allied liquid-cooled aircraft piston engines through WWII, copper-lead bearings find application to this day.
In 1934 Wilson-Jones devised a floating white-metalled bush big-end for Cecil Barrow's TT250, and it was successful, potentially putting an end to traditional roller big-end troubles. It was made standard on production models in 1939. Students of racing history will recall that it was mainly problems with roller big-ends that impeded successful development of Moto Guzzi's ambitious 500 V8 of 1955-57. The last Norton to win the Sr. TT (1961) bypassed the chronic big-end problems of the time by use of a Jaguar plain-bearing con-rod.
Despite this, the assumption that rolling bearings were essential for sports and racing engines persisted until definitive studies in Formula One, established that even at high revolutions there is no useful difference in the friction losses of the two bearing types. When a new generation of liquid-cooled sporting motorcycle appeared in the 1980s, all had plain bearings. Rolling bearing persist mainly where pressure lubrication isn't possible – in two-stroke and in lightweight off-road four strokes that carry only rudimentary oil systems.