Are Baby Bottle Measurement Lines Accurate? What the Study Found
The short answer
Not always. In a 2016 survey of 91 bottles from 28 brands sold in Australia, 19 of the 88 hard-sided bottles had at least one marking outside the EN14350 tolerance, and bottles that claimed compliance were not less likely to be off. The bottles were bought in 2013 and 2014, so the survey cannot tell you about yours.
Key takeaways
- The survey (Gribble and colleagues, 2016) checked 91 bottles from 28 brands bought in Australia between December 2013 and February 2014.
- Nineteen bottles (22%) had at least one measured marking outside the EN14350 tolerance. In our reading, that is 19 of the 88 hard-sided bottles. The 37 markings involved were 21 over and 16 under the true volume.
- Bottles that claimed EN14350 compliance were not less likely to be off: 28% against 17% without a claim, not statistically significant (p = .23). Price made no significant difference either (p = .13).
- The tolerance is ±5 ml for marks under 100 ml and ±5% from there. At the 120 ml mark that is 114 to 126 ml (our arithmetic).
- The authors suggest checking the markings with a scale accurate to 1 g. Philips Avent support says a weight scale gives accurate results and the bottle scale is a guide.
- The survey tested one bottle of each type, bought in 2013 and 2014. It says nothing about your bottle.
If your food scale and the lines on the bottle disagree, this page explains what one study found and how to check a bottle at home. If you keep a log, the free bottle feeding log has a place for the number you measured.
What did the survey of baby bottle markings find?
A 2016 survey checked the volume markings on 91 baby bottles sold in Australia. By our reading of the paper’s counts, 19 of the 88 hard-sided bottles (22%) had at least one marking outside the tolerance of the European standard EN14350. The table lists the figures as the paper reports them.
Table 1. What the survey found (Gribble, Berry, Kerac and Challinor, Maternal & Child Nutrition, 2016)
| Measure | Result | Where in the paper |
|---|---|---|
| Bottles and brands | 91 bottles, 28 brands, 19 outlets, bought December 2013 to February 2014 in Australia. 88 were hard-sided and 3 were disposable liner systems | Abstract, Results |
| Marks checked | 50, 60, 90, 100, 120 and 150 ml, filled with deionized water and weighed to 0.1 g at 25°C | Methods |
| Bottles with at least one marking outside EN14350 tolerance | 19 bottles (22%). Our reading of the base: 19 of the 88 hard-sided bottles | Abstract, Results |
| Inaccurate markings | 37 markings on those 19 bottles (1 to 5 per bottle): 21 overestimated the volume, 16 underestimated it | Results |
| Missing markings | 39 bottles (44%) lacked at least one marking for a volume in manufacturers’ instructions for infant formula available in Australia, for infants 2 months of age and younger (1 to 3 markings missing per bottle). 50 bottles (57%) had an inaccurate or a missing marking | Results |
| Claimed EN14350 compliance | 10 of 36 bottles with a claim (28%) and 9 of 52 without (17%) had an inaccurate marking. Not significant (p = .23) | Results |
| Price | 9 of 53 bottles over US$5 (17%) and 12 of 38 under US$5 (32%) had an inaccurate marking. Not significant (p = .13) | Results |
| Disposable liner bottles | 3 bottles with “particularly large volume inaccuracies (up to 43% outside the marked volume)” | Abstract, Results |
| Embossed or printed lines | 40 bottles embossed (45%), 45 printed (51%), 3 both. Embossed-only bottles were no more likely to be inaccurate than printed-only (p = .78). Not significant | Results |
| Ounce lines | Most bottles had ml and ounce lines. 54 bottles (61%) did not say which ounce. Only ml marks were checked | Results, Methods |
The paper does not print the base next to its sentence “Nineteen bottles (22%)”. We read it as the 88 hard-sided bottles because 10 of 36 and 9 of 52 add up to 19 of 88, and 19 / 88 = 21.6%, which rounds to 22%. 19 / 91 would be 20.9%.
The paper’s Results text also says “Ninety-one of these bottles were hard sided”, which conflicts with its Abstract and Table 1 (88). We use 88.
The price comparison in the paper adds up to 91 bottles (53 + 38) and 21 inaccurate ones (9 + 12), so its counts do not match the 88-bottle analysis. We give each figure as the paper does and do not reconcile them.
What does the EN14350 tolerance mean in milliliters?
As the paper states it, the tolerance is plus or minus 5 ml for marks under 100 ml, and plus or minus 5% for larger marks. EN14350 itself is paywalled, so we cite it through the paper.
Table 2. EN14350 tolerance in ml for the marks parents use (our arithmetic from the tolerance as the paper states it)
| Mark | Tolerance | Accepted range | Tolerance as % of the mark |
|---|---|---|---|
| 50 ml (1.7 fl oz) | ±5 ml | 45 to 55 ml | 10% |
| 60 ml (2.0 fl oz) | ±5 ml | 55 to 65 ml | 8.3% |
| 90 ml (3.0 fl oz) | ±5 ml | 85 to 95 ml | 5.6% |
| 100 ml (3.4 fl oz) | ±5% = ±5 ml | 95 to 105 ml | 5% |
| 120 ml (4.1 fl oz) | ±5% = ±6 ml | 114 to 126 ml | 5% |
| 150 ml (5.1 fl oz) | ±5% = ±7.5 ml | 142.5 to 157.5 ml | 5% |
| 4 fl oz (118.3 ml) | Not stated | Not stated | Not stated |
Arithmetic: 100 x 0.05 = 5 ml, 120 x 0.05 = 6 ml, 150 x 0.05 = 7.5 ml.
For the small marks, 5 / 50 = 10%, 5 / 60 = 8.3%, 5 / 90 = 5.6%.
The paper’s Table 2 footnote says “over 100 ml” for the percentage rule. At 100 ml both readings give 5 ml. The fl oz figures in the Mark column are the ml mark divided by 29.5735, rounded to one decimal.
The paper tested ml marks only, and none of the sources we opened gives a tolerance for ounce lines, so the last row says “Not stated”. The conversion is 4 x 29.5735 = 118.294, about 118.3 ml (NIST Guide to the SI, Appendix B.8, page updated 2025-08-18, opened 2026-11-01).
What do bottle makers say about the lines on their bottles?
The one maker page we could open gives no numeric tolerance. Its advice matches the paper’s: weigh water to check. Maker pages are the maker’s own guidance, not evidence of how accurate a bottle is.
Table 3. Manufacturer guidance (not evidence)
| Source | What it says | Last checked |
|---|---|---|
| Philips Avent support, How accurate is the scale on my Philips Avent bottle? | “Using a weight scale to measure the volume of water will provide accurate results, while the scale on the bottle can provide a guide (even when slight distortion has occurred).” No number is given | 2026-11-01 |
| Same page | Says plastics “can shrink or distort, particularly when heated”, and that “a small deviation is acceptable” under guidelines | 2026-11-01 |
The page shows no date for the answer. Two other pages we tried to open, a second maker’s support page and a university news item about the study, could not be opened on 2026-11-01, so they have no row.
How can you check the lines on your bottle against a food scale?
Weigh water to the line and compare the grams with the mark. This is based on the paper’s method (water weighed at each mark, meniscus level with the middle of the mark), and its authors suggest testing markings “using a scale accurate to 1 g”. The tare step and the comparison with Table 2 are ours.
- Pick the line you use most. The paper checked 50, 60, 90, 100, 120 and 150 ml.
- Put the empty, dry bottle on a food scale that reads in 1 g steps or finer. Press tare so the scale shows 0.
- Pour in water until the lowest point of the water surface sits at the middle of the line. The paper levelled the base of the meniscus with the midpoint of the marking.
- Read the grams. One gram of water is about one ml (OpenStax Chemistry 2e, water properties table, opened 2026-11-01).
- Subtract the mark from the grams. A plus sign means the bottle holds more than the line says. A minus sign means it holds less.
- Compare the result with the accepted range in Table 2, and write it down.
Water is 0.9982 g/mL at 20°C and 0.9970 g/mL at 25°C, so reading grams as ml makes the volume look about 0.2 to 0.3% smaller than it is. The example below ignores that.
Example, hypothetical: water filled to the 120 ml line weighs 128 g. 128 - 120 = 8 ml more than the line says, which is 8 / 120 = 6.7% of the mark. By Table 2 the accepted range at 120 ml is 114 to 126 ml, so 128 ml is outside it. A reading of 124 g would be inside: 124 - 120 = 4 ml, 4 / 120 = 3.3%.
This is a check of one bottle on one day, not a verdict on a brand. The paper tested one bottle of each type, and Philips says plastic can distort when heated, so a reading describes the bottle as it is today.
Table 4. A sheet to keep your own offsets (the first row is the hypothetical example)
| Mark | Scale reading (g) | Difference (ml) |
|---|---|---|
| 120 ml | 128 | +8 (example) |
| 60 ml | (your reading) | (your difference) |
| 90 ml | (your reading) | (your difference) |
| 150 ml | (your reading) | (your difference) |
What do you write down when the line and the scale disagree?
Write the number you measured and say how you measured it: “line” or “scale”. The method tells you later what the number means.
Example: you fill to the 120 ml line and the scale reads 128 g, so you offered 128 ml by scale. When the feed ends, the milk reaches the 30 ml line, which you write as 30 ml left by line. Offered 128 ml (4.3 fl oz), left 30 ml (1.0 fl oz), drank 128 - 30 = 98 ml (3.3 fl oz). Put “offered by scale, left by line” in the note.
Mixing the two methods adds a small error, because the 30 ml line has its own offset. The paper tested no mark below 50 ml, so it cannot say how big that offset is. This page does not correct for it. If you want to read the leftover milk the same way every time, see how to read what is left in the bottle. If one bottle holds breast milk and formula, write the method next to each number too, and our guide to breast milk and formula in one bottle shows what the official pages say about it.
The printable log takes the number you write, exactly as you wrote it. Rememilk, the bottle log we are building, is in development. The printable works without it.
Does this matter for formula preparation?
The paper’s authors link marking errors to the volume of water used to make formula. They write that inaccurate markings are “a previously neglected but potentially important source of error in the reconstitution of infant formula.” They also state that the effects in real life “have not been studied.” This page gives no formula instructions: follow the label on your formula, and ask your pediatrician.
Where can you write the number down?
Get the free printable: it has columns for offered, left and drank, and room for a note about how you measured.
What this page is, and what it is not
The survey bought one bottle of each type in Australia between December 2013 and February 2014. Its authors write that different production batches could be more or less accurate, that bottles in other countries may differ, and that the effects of the errors on infants "have not been studied." The survey tested marks from 50 ml up, so it says nothing about a 30 ml line. On 2026-11-01 we ran two web searches for a study published after 2016 on the accuracy of bottle markings and found none. Two searches are not proof that none exists.
This page names no bottle brand as inaccurate and is not a recall list. The paper does not name brands either. Philips Avent appears only as manufacturer guidance. We are parents, not medical professionals, and this page gives no medical advice. The source method page explains how a source gets in, and who writes here states the author's role.
Frequently asked questions
Are baby bottle measurement lines accurate?
In one 2016 survey of bottles sold in Australia, 19 of the 88 hard-sided bottles had at least one marking outside the EN14350 tolerance. By our reading of the paper's counts, the other hard-sided bottles had every measured marking inside it. The survey cannot say where your bottle falls.
Do expensive baby bottles have more accurate lines?
In the survey, 9 of 53 bottles priced over US$5 (17%) had an inaccurate marking, against 12 of 38 priced under US$5 (32%). The paper reports no statistically significant link between price and accuracy (p = .13). These price counts add up to 91 bottles, not the 88 behind the 19 (see the note under Table 1).
Does EN14350 on the box mean the lines are accurate?
In the survey, 10 of 36 bottles that claimed EN14350 compliance (28%) had a marking outside the tolerance, against 9 of 52 without a claim (17%). The authors found no significant difference (p = .23), so a claim on the box did not predict accuracy there.
How do I check the measurement lines on a baby bottle?
Weigh the empty bottle on a food scale, fill it with water to the line, and read the grams. One gram of water is about one ml, so the difference from the mark is your offset. The numbered steps above are based on the paper's method.
Are the ounce lines as accurate as the ml lines?
Not stated. The paper checked ml marks only, and none of the sources we opened gives an ounce tolerance. In the survey, 54 of 88 hard-sided bottles did not say which kind of ounce their lines used.
Sources
- Gribble K, Berry N, Kerac M, Challinor M, Volume marker inaccuracies: a cross-sectional survey of infant feeding bottles
- Philips Avent support, How accurate is the scale on my Philips Avent bottle?
- NIST Guide to the SI (SP 811), Appendix B.8
- OpenStax Chemistry 2e, Appendix E: Water Properties (LibreTexts copy)
Rememilk is a bottle milk log in development
It keeps offered, left and drank for every bottle, and nursing you cannot measure.