Gas mark ½ is not half of anything
Gas marks run ¼, ½, 1, 2, 3 … 9, and the two fractional ones look like they ought to be fractions. They are not.
The marks are evenly spaced 25 °F apart, and mark 1 is 275 °F. That gives a tidy formula — °F = 250 + 25 × mark — and it is correct for every mark from 1 to 9. Apply it to mark ½ and it says 262.5 °F. The real mark ½ is 250 °F.
What ¼ and ½ actually are is the two 25 °F steps below mark 1: 250 and 225 °F. They continue the spacing perfectly and simply break the numbering, because the scale needed low settings and there was no room below 1 without inventing a zero.
The full gas scale, with the fan equivalent for each mark| Mark | °F | °C | °C fan |
|---|
| ¼ | 225 | 105 | 85 |
|---|
| ½ | 250 | 120 | 100 |
|---|
| 1 | 275 | 135 | 115 |
|---|
| 2 | 300 | 150 | 130 |
|---|
| 3 | 325 | 165 | 145 |
|---|
| 4 | 350 | 175 | 155 |
|---|
| 5 | 375 | 190 | 170 |
|---|
| 6 | 400 | 205 | 185 |
|---|
| 7 | 425 | 220 | 200 |
|---|
| 8 | 450 | 230 | 210 |
|---|
| 9 | 475 | 245 | 225 |
|---|
One further thing follows from the spacing: a gas mark is only 13.9 °C. That is a small step, and it becomes important below.
The recipe does not say whether it means fan
A fan oven circulates air, which transfers heat to the food faster. So it cooks as though it were hotter than it says, and the conventional allowance is 20 °C: a recipe written for 180 °C conventional wants 160 °C in a fan oven.
The difficulty is that recipes very often do not say which they mean. Get the direction wrong and the error is not 20 °C but 40, because you have moved the wrong way — and 40 °C is nearly three gas marks. That is the difference between a cake and a burnt one.
Some rules of thumb that usually hold. A recipe that gives a gas mark is conventional, always, because gas ovens do not have fans. British and European recipes generally mean conventional unless they print “fan” explicitly, and increasingly they give both. American recipes in Fahrenheit almost always mean conventional, since convection is usually flagged as an option rather than assumed.
One genuine complication to check in your manual: some fan ovens apply the offset internally, so the number on the display is already the conventional equivalent and you should set what the recipe says. Ovens differ, and this is the one case where the general rule will send you the wrong way.
Timing shifts too. At the correctly reduced temperature a fan oven still tends to cook a little faster and much more evenly, so start checking around 10% earlier than the recipe says, and expect to stop rotating trays.
Your oven is out by more than a gas mark
Here is the thing that makes all of the above less important than it sounds. A domestic oven thermostat is commonly wrong by ±20 °C, and one gas mark is only 13.9 °C. The appliance’s own error is larger than the step you are converting between.
It is worse than a single offset, too. The temperature varies through the cavity — typically by 20 to 25 °C between the top and the bottom — and it cycles as the thermostat switches the element on and off, so the actual temperature at any moment swings around the setting rather than sitting at it. Opening the door drops it sharply, and it takes minutes to recover.
Which means converting to a tenth of a degree is precision that has nowhere to go. The calculator above rounds its headline figures to whole degrees for exactly that reason.
The genuinely useful response is a cheap oven thermometer. Put it in the middle, preheat properly, and see what your oven really does at each setting — you will probably find a consistent offset you can simply apply, and you will find out which shelf runs hot. That single measurement is worth more than any conversion table, and it also explains a lot of past baking.
Why the tables never quite agree
Compare two published oven charts and they will differ by five degrees here and there. Gas mark 4 appears as 180 °C on one and 175 °C on another; mark 6 as 200 or 205.
The reason is that the gas scale is defined in Fahrenheit and converts to inconvenient Celsius. Mark 4 is exactly 350 °F, which is 176.67 °C — a number no chart wants to print and no dial can be set to. So every publisher rounds, and they round to different conveniences: some to the nearest 5, some to the nearest 10, some to the round number nearest the oven dial’s own markings.
None of them is wrong, and the spread between them is a few degrees — well inside the error of the oven itself. The figures on this page come from the underlying 25 °F spacing rather than from any one chart, which is why they may not match the table on a packet exactly and why that does not matter.
The same is true in the other direction. A recipe saying 180 °C is not naming a gas mark: it falls 3.3 °C above mark 4 and 11 °C below mark 5. Mark 4 is the right answer, and it is not an exact one.
What the old descriptions meant
Older recipes often give no number at all, only a description — a moderate oven, a slow oven, a hot oven. These were not vague at the time; they were the only available vocabulary before thermostats, and they map onto the modern scale reasonably well.
Very cool or very slow is around 105 to 120 °C, gas ¼ to ½ — meringues, drying, and very long slow cooking. Cool or slow is about 135 to 150 °C, gas 1 to 2, for rich fruit cakes and long casseroles. Warm is around 160 °C, gas 3.
Moderate is 175 to 180 °C, gas 4 — the default for most cakes and biscuits, and the temperature a recipe means when it gives no other clue. Moderately hot is 190 °C, gas 5. Hot is 200 to 220 °C, gas 6 to 7, for pastry and roasting. And very hot is 230 °C and above, gas 8 to 9, for bread, pizza and searing.
A caution worth adding for old recipes generally: the descriptions are the least of the differences. Ingredients, flour strength, oven behaviour and tin sizes have all changed, so a nineteenth-century “moderate oven” converted precisely to 180 °C is still only a starting point for a recipe that assumed a cooling bread oven and a different flour.
Sources and methodology
The Celsius-to-Fahrenheit arithmetic is a definition and needs no authority; the gas-mark scale is a British trade convention rather than a standard, which is exactly why its two fractional marks behave oddly. What the sources establish is the part with consequences: that a thermostat setting is a nominal figure measured at one defined point under test conditions, not a uniform cavity temperature — and that whether food is safely cooked depends on the temperature reached inside it rather than on the number on the dial.