A roselle drink can lose a substantial amount of monomeric anthocyanins without an equally obvious change in appearance. In a six-month study of roselle–fruit drinks, the researchers reported no visually observed color change at 4°C, even though total monomeric anthocyanin losses at that temperature ranged from 53% to 75% across the studied 40% roselle blends.[2] That contrast is more useful to a formulator than a small difference between two redness readings: pigment content and apparent color answer different questions.
The observation does not mean that the drinks were instrumentally unchanged. The same paper reports changes in color coordinates, and its visual observation is not a quantified consumer-acceptance result.[2] It does show why an appearance check alone is a poor basis for estimating how much monomeric anthocyanin remains.
What was in the drinks?
The 2014 study blended aqueous roselle calyx extract with mango, papaya or guava juice; detailed bioactive-component comparisons used blends containing 40% roselle extract.[2] A 2015 paper by the same four authors examined roselle–mango blends in glass and plastic bottles, stored at 4°C and 28°C for six months.[3] These are related reports from one research group, not independent replication.[2][3]
The experimental material was a beverage made with a water extract, not a standardized commercial color powder. Both studies used sodium benzoate and citric acid at 1 g/L each, with a water-bath treatment of 82.5°C for 20 minutes.[2][3] Those conditions describe the experiments, not a production recipe: 40% liquid extract is not 40% dry powder, and the studies do not establish additive permission, a safe process or a shelf life for another product.
Temperature, pigment and color: separate results
The 2015 paper measured total monomeric anthocyanins by the pH differential method and evaluated color separately using L*, a* and b* coordinates.[3] Its statistical analysis offers a useful distinction. In Table 5, the storage-temperature main effect is significant for monomeric anthocyanins (P < 0.001), but is marked “not significant” for a*, the red–green coordinate.[3] This is evidence that the two endpoints did not give the same statistical result in this experiment—not proof that temperature has no effect on redness. The temperature-by-time interaction for a* was significant (P < 0.001), so the temperature main effect cannot describe every storage interval.[3]
The actual color readings also deserve a fuller view than a* alone. The following descriptive table reproduces the 40R rows of Table 4, keeping all three coordinates and the reported ± values.[3]
| Package and storage condition | L* | a* | b* |
|---|---|---|---|
| Before storage, both packages | 17.9 ± 0.04 | 18.3 ± 0.53 | 7.8 ± 0.07 |
| Glass, six months at 28°C | 16.4 ± 0.02 | 17.2 ± 0.06 | 6.6 ± 0.07 |
| Glass, six months at 4°C | 16.5 ± 0.06 | 17.4 ± 0.04 | 7.1 ± 0.02 |
| Plastic, six months at 28°C | 14.7 ± 0.06 | 15.3 ± 0.07 | 4.6 ± 0.41 |
| Plastic, six months at 4°C | 15.1 ± 0.04 | 15.2 ± 0.05 | 5.9 ± 0.05 |
The paper reports triplicate testing (n = 3), but its text and Table 4 do not explicitly identify the ± terms as standard deviations or standard errors; they are retained as reported, not relabeled. Table 4 also mislabels 40R as “100% roselle” in its footnote; the formulation and statistical methods identify it as 40% roselle extract.[3]
In plastic bottles, the final a* means were close, while L* and b* also changed from baseline.[3] The 0.1-unit difference between temperatures is not a demonstrated pairwise effect, nor does it establish that refrigeration worsened the color. The visual-versus-chemical observation in the earlier study and the endpoint-specific statistical results already make the distinction.[2][3]
Why the analytical method matters
The pH differential assay estimates monomeric anthocyanins under specified analytical conditions; L*, a* and b* describe different dimensions of measured color.[3] An a* value is therefore not an anthocyanin concentration, and its percentage change is not a pigment-retention percentage. Likewise, a stable-looking drink is not evidence of a stable marker content.
Antioxidant-assay results add another, separate dimension. The 2015 authors reported FRAP losses below 30% despite marked monomeric anthocyanin losses.[3] They suggested that polymeric compounds might partly explain the difference, but did not establish that mechanism in the reported experiment.[3] FRAP here is a laboratory reducing-capacity assay, not a direct measure of redness or evidence of a human health benefit.
Choosing a hibiscus material for an actual beverage
For a color-led project, compare materials in the intended fruit base at a similar starting color. Keep the addition rate visible: a material that needs more powder to reach the target has a different formulation cost, even if its quoted price per kilogram is lower. Botanical identity, calyx-derived material, extract format and any carrier belong in the material brief so that unlike products are not treated as interchangeable.
During storage, track the full color coordinates and compare appearance against an agreed reference. Measure monomeric anthocyanins separately if marker retention is also a product requirement. Use the intended package, process and distribution conditions; neither a raw-material assay nor the glass-versus-plastic result from this experiment substitutes for testing that combination.
The practical decision is not whether chemistry or color is the “better” measurement. It is which requirement the product must meet—and whether the evidence actually measures it.