My Outer Jars in a 100-Candle Batch Show Different Adhesion from the Centre Jars — Could Uneven Cooling Explain the Difference?

What the pattern most likely is, and the first check

When a hundred freshly poured jars stand in one tight block, they do not cool together. Each jar loses heat through the glass, and how fast it loses it depends on what is on the other side of that glass. A jar in the middle of the block is surrounded by other jars that are just as hot, so its walls face warm air and warm neighbours. A jar on the edge has one side facing the open room; a corner jar has two. In a 10 × 10 block that is 4 corners and 32 edge jars — 36 jars, 36% of your batch, cooling from at least one side much faster than the 36 jars in the core. Wet spots are commonly described as places where wax has separated from the glass as it contracts while it cools and sets, and uneven or fast cooling is one of the most commonly cited contributors. So yes: uneven cooling across the block is a plausible and common explanation for what you are seeing.

Plausible is not proven, and there are two impostors that produce the same picture. The first is fill order. Most makers fill a block row by row, starting from one corner, so the jars on the outside are often also the first or the last jars filled. The wax in the jug cools while you work, so the first jar and the last jar of a hundred did not get wax at the same temperature. A position pattern and a pour-temperature pattern can sit exactly on top of each other. The second impostor is where the jars came from. If the outer jars were the ones lifted from the top of a carton, from a carton that sat by the door, or from a different delivery, they may have gone into the block colder or from a different glass lot than the centre ones.

So the single cheapest first check costs nothing: on the next batch, write down for every jar its row and column, the order in which it was filled, and which carton it came from. Photograph the block from above before anything is moved. When the spots appear, score each jar and look at which of those three lists the spots follow. If the spots follow position and not fill order or carton, cooling is your answer and the fix is a layout change, not a purchase. The rest of this page shows how to do that check properly, how to confirm it with a one-variable test, and what each outcome tells you to buy — usually nothing.

Why the outside of a block cools differently from the inside

Picture the block an hour after the last jar is filled. Every jar is giving heat away. A core jar gives its heat mostly to its neighbours, which are giving theirs back — the whole core behaves like one large, slowly cooling mass. An edge jar gives heat to its neighbours on three sides and straight to the room on the fourth. A corner jar faces the room on two sides. If the room has moving air from a fan, an open door or an AC unit, the open sides lose heat faster still. The result is that the glass on the outward face of an edge jar can be colder, sooner, than any glass in the core.

Why should that matter for adhesion? In general candle-making practice, wax shrinks as it cools and crystallises. Where it sets against glass that is cooling quickly, and the wax further in is still contracting, the outer skin can be pulled away from the glass, leaving a thin air gap that looks dark and wet through clear glass. Makers commonly report that faster or more uneven cooling makes this more likely. CSI publishes no contraction figure for any of its waxes, and we will not give you one; the mechanism is described here as it is commonly described, and the test below is how you find out whether it applies to your block.

Look closely at where on each outer jar the spots sit. If cooling from the open side is the cause, you would expect spots to favour the side of the jar that faced the room, not the side that faced the block. That is a free piece of evidence most makers never collect, because they rotate jars while lidding and labelling. Mark the outward face of every edge and corner jar with a small dot on the base before you move it. If the spots cluster on the dotted face, the open side is doing it. The post on perfect adhesion on one side and wet spots on the other goes deeper into one-sided patterns on a single jar.

Step one: map the batch you already have

You may already have the evidence sitting on a shelf. If the batch has not been moved, or you can reconstruct the layout from photographs, score every jar with the house scale further down and write the score into a 10 × 10 grid on paper. Then group the jars: the 4 corners, the 32 edge jars, the 28 jars of the second ring and the 36 jars of the core. For each group, divide the number of jars with any wet spot at your house threshold by the number of jars in the group. Four numbers, one per ring, tell you more than any amount of looking.

Now do the same count twice more, grouping the same jars differently. First by fill order: the first 25 jars filled, the next 25, the next 25 and the last 25. Then by carton. If the four ring rates step down neatly from corners to core while the fill-order quarters and the cartons look alike, you have a position effect. If the fill-order quarters step and the rings do not, you have a pour-temperature effect that only looked like position. If one carton stands out, you have a jar question. Very often, with a row-by-row fill, two of these lists step together — which is exactly why the next step exists.

Step two: separate position from fill order and carton

You do not need extra materials to separate the three suspects; you need a fill plan. On your next production batch, decide in advance where each jar will stand, and make sure that early-filled and late-filled jars end up in every ring. A simple way: number the hundred positions in the grid, then fill in an order that jumps between the outside and the inside — for example, the first jar to a corner, the second to the core, the third to an edge, the fourth to the second ring, and so on round the block. Write the plan on a sheet before you melt and tick each position as it is filled.

Do the same for cartons. If your jars arrive in packs of 10, as CSI's Clear Glass Jar with Golden Lid (150 gram) does, take one jar from each pack in turn rather than emptying one pack into one corner of the block. Then every ring holds jars from every pack. Finally, read the wax temperature in the jug with a thermometer at the first fill and at the last fill, and write both down. If you work in Luxury Soy Wax Chunks, its page gives a pour of 75–80°C in its spec table (the steps also show 70–80°C); CSI's Clear 150 g jar page says the jar is safe for hot pour up to 85°C. Both figures are what the product pages state, not tested limits; stay inside both, and note how far the jug drifts across the batch.

This step costs time, not money: a fill plan and some extra writing. It also gives you a result you can trust. If the ring rates still step down from corners to core after fill order and cartons have been spread evenly across the rings, position has survived its two most common impostors and cooling is now the leading suspect. If the stepping disappears once fill order is shuffled, your original pattern was a jug-temperature pattern, and the post on finding a pour temperature for both adhesion and good tops is where to go next.

Step three: the one-variable cooling test

Once position has survived step two, test the cooling explanation directly by changing one thing: how much the outer ring is exposed to the room. Pour the next hundred not as one 10 × 10 block but as four 5 × 5 blocks, placed well apart on the same surface in the same room. Each 5 × 5 block has 16 outer jars and 9 inner jars. Two of the blocks are the control — open to the room, exactly as you cool now. The other two are the test arm — each surrounded by a simple collar that shelters the outer faces, such as a cardboard wall standing just clear of the jars on all four sides (a labelled example; any clean, non-touching barrier does the job). Everything else is identical: the same melt, the same fill plan from step two, the same jug, the same surface, the same room and the same checkpoints.

Then compare like with like. The 32 outer jars of the two open blocks against the 32 outer jars of the two sheltered blocks. The 18 inner jars of each pair are your internal check: they should look alike in both arms, because the collar barely changes their surroundings. If the sheltered outer jars look like the inner jars and the open outer jars do not, the open sides were doing it. If both outer groups look alike, sheltering did nothing, and cooling from the sides is not your main cause.

Two blocks per arm, rather than one, matter more than they look. A single block can be unlucky — closer to the door, nearer a sunny window, under a fan blade. With two blocks per arm, placed so each arm has one block on each side of the table, a difference that appears in both blocks of an arm is far more convincing than one that appears in one. Decide your rule for "a real difference" before you look at a single jar: for example, the sheltered outer rate must be lower than the open outer rate in both pairs of blocks, and at both the 48-hour and 7-day checkpoints. Write it at the top of your sheet.

The scoring sheet: your house scale

No industry figure exists for how much wet spotting is acceptable, so you set your own house standard and then never change it between tests. The example below grades each of four marked sides of every jar from 0 to 4 and adds them, so a jar scores 0–16. The grades describe what you see through the glass against a plain, evenly lit background, looking straight at the jar from a fixed distance. Use a printed grid or a marked acetate sleeve around the jar if you want to estimate the share of each side that is spotted; the sleeve keeps two people's estimates closer together.

Side 1 is always the dotted, outward face for edge and corner jars, so the sheet shows at a glance whether spots favour the side that faced the room. The "any side ≥2" column is the one you tally by group, because it matches what a customer sees: a jar either has a noticeable spot somewhere or it does not. The total score is for comparing severity once the counts agree. What to record when comparing the glass adhesion of different waxes has a longer recording template if you want to reuse one sheet across every test you run.

What the outer ring costs you

It is worth knowing what this pattern is worth before you spend time or money on it. Here is the series default candle — 150 g in CSI's Clear 150 g jar at 8%, so 12 g of oil and 138 g of wax — priced with Luxury Soy Wax Chunks at the 5 kg bag rate (₹0.60 a gram), Lavender at the 1 kg rate (₹5.22 a gram), an Eco Thin C1 wick at ₹5.90 as the example wick line (its size for this jar must be burn-tested — CSI publishes no wick size for any jar) and the jar at ₹94.40. That is ₹245.61 of CSI materials per candle, before labour, packaging, freight and GST, and ₹24,561.40 for the batch of 100.

The rows are scenarios to show the arithmetic, not results we have measured: your count comes from your ring tally. But the shape is clear. If the whole outer ring fails a premium standard, every saleable candle carries ₹138.16 more material cost than it would if none did. And those 36 jars do not simply vanish. Some makers hide them behind an opaque label and some sell them at a discount; each route has its own cost, and you should count them separately because they behave differently.

Set those numbers against the test. The in-production test on this page costs a fill plan, some cardboard and an hour of extra recording. Even if your real figures are half the examples, a fix that moves a few outer jars per batch back to full price pays for itself on the first batch. Calculating the cost of cosmetic rejects builds the full reject-cost model with complaints and discounts included.

The priced screen, if you test outside production

Most makers with a hundred-jar batch should run the test inside production — it is free and it tests the real layout. If you would rather not risk a production day, or you are about to scale up to blocks this size for the first time, the same comparison works at a smaller scale: two 4 × 5 blocks of 20 jars, one open and one sheltered, poured from one melt. Each 4 × 5 block has 14 outer and 6 inner jars. That is fewer jars per group than the full test, so treat it as a screen: a clear difference is worth acting on; a small one is worth repeating in production. Grams use the CSI convention (oil = W × L) and a 5% pour-loss working allowance.

Two points of honesty about that total. First, it is mostly not a test cost: forty finished candles come out of it, and every one that meets your house standard is stock. The real cost of the screen is the materials in the jars that fail, plus your time. Second, if you already make this candle, do not buy new wax or oil for it — use your production lot. A screen poured from a fresh bag of wax adds a second variable (the wax lot) to a test that is supposed to have one. Each extra jar you add to an arm costs about ₹247.76 in CSI materials at small-pack prices, which is the arithmetic to use if you want more than 14 outer jars per group.

If cooling is confirmed: layout fixes in order of cost

If the sheltered outer jars clearly beat the open outer jars, you have three ways to make every jar in the block cool more alike, and they cost progressively more. Try them one at a time, each against a control block in the same batch, and keep only what earns its place.

Spacing deserves a word because it changes the question entirely. When jars stand apart, there is no core and no edge: every jar has open air on every side, and the block stops having neighbourhoods. That makes jars more alike, which is what you want for consistency, but it may make every jar cool faster than your old core jars did. Whether that is better or worse for your wax in your jar is exactly the kind of thing you test, not assume. Should freshly poured candles be spaced consistently? works through spacing as its own variable, and reproducing slow cooling at commercial scale deals with slowing the whole batch down.

The decision rule: what to buy, or not

Apply this after the tally and the collar test. Stop at the first line that is true; it is written so two people reading the same sheets reach the same decision.

Most readers with this exact question end on the first three lines, and all three say buy nothing from us. That is not modesty; it is the logic of a within-batch pattern. A wax change is the right move only when the problem is in every jar, and the cooling and jar checks have been cleared — and even then it re-opens wick, fragrance-load, cure and hot-throw validation, which is why it sits at the bottom of the rule.

The CSI products on this page

Frequently asked questions