Warm any oil and it thins. That is not the interesting part; every liquid does it. The interesting part is that two oils can sit at the same viscosity at room temperature and thin at very different rates as they heat up. One stays usable across a wide temperature span; the other turns to water in a hot engine. Viscosity index, or VI, is the single number that captures that difference — and, like any single number standing in for a curve, it is worth knowing what it hides.
What VI actually measures
Viscosity falls with temperature along a curve, not a straight line. VI does not describe the whole curve; it places your oil against two reference series measured at the same two temperatures. Those reference oils are long gone; what survives is a set of fixed tables derived from them. In the 1920s, oils refined from Pennsylvania crude changed little with temperature and were assigned VI = 100; oils from Gulf Coast crude changed a lot and were assigned VI = 0.1 Your oil’s kinematic viscosity at 40 °C and 100 °C is compared to those tables, and the result is its VI.2 Higher VI means a flatter curve — less change with temperature.
From the oil’s 100 °C viscosity, the D2270 tables give two anchor values: L, the 40 °C viscosity a VI = 0 oil of that same 100 °C viscosity would have, and H, the 40 °C viscosity a VI = 100 oil would have. Your oil’s own 40 °C viscosity, U, is placed between them: VI = (L − U) / (L − H) × 100. A separate extrapolation formula covers VI above 100, where the scale is open-ended — many synthetics reach 150–200, with no theoretical ceiling. The scheme rests on one assumption: that all oils follow the same family of viscosity-temperature curves between the references. For some synthetics, esters, and silicones that assumption bends, and VI becomes a looser guide.
An engine oil has two jobs that pull in opposite directions: thin enough to pump and flow at a cold start, thick enough to keep a protective film at operating heat. A low-VI oil cannot do both — set it thin enough for the cold and it is too thin when hot. A high-VI oil holds a usable viscosity across the whole range, which is what makes possible a multigrade oil like 10W-40.
What the number leaves out
VI is a summary, and summaries lose things. It says nothing about absolute viscosity — two oils of wildly different thickness can share a VI. It is defined only from the 40 °C and 100 °C points, so it does not see behaviour below or above that span, where a real cold start or a hot bearing actually lives. And the most important omission: much of the high VI in multigrade engine oils comes from VI improver polymers, which raise the bench VI but shear-thin under the stress of a running engine. The oil in the bearing can be thinner than its VI suggests, temporarily or — if the polymer is permanently cut — for good.
Density enters only through the kinematic viscosity, which is dynamic viscosity divided by density; but a meter that reports viscosity and density together (the subject of an earlier piece) gives you the inputs for VI and the density in one pass.3
The limits of one number
Treat VI as what it is: a useful, standardised shorthand for temperature stability, not a full description of an oil. A high VI is genuinely valuable, but confirm the absolute viscosities at the temperatures you care about, and remember that a VI-improved oil’s real in-service viscosity is a shear-rate question, not a bench number — which is exactly where the next piece, on Newtonian and non-Newtonian behaviour, picks up.