HAL Id: hal-00929301https://hal.archives-ouvertes.fr/hal-00929301

Submitted on 1 Jan 1992

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

Tristimulus colour reflectance measurement of milk anddairy products

W Kneifel, F Ulberth, E Schaffer

To cite this version:W Kneifel, F Ulberth, E Schaffer. Tristimulus colour reflectance measurement of milk and dairyproducts. Le Lait, INRA Editions, 1992, 72 (4), pp.383-391. <hal-00929301>

Lait (1992) 72, 383-391© Elsevier/INRA

383

Original article

Tristimulus colour reflectance measurementof milk and dairy products

W Kneifel, F Ulberth, E Schaffer

Agricultural University, Department of Dairy Research and Bacteriology,Gregor Mendel-Str 33, A-118D Vienna, Austria

(Received 20 December 1991; accepted 4 May 1992)

Summary - A tristimulus reflectance technique was applied to the objective assessment of the colourof milk and dairy products. A variety of milk and dairy products (Iiquid milk, cultured products, cheese,butter, milk powder) was characterized based on the L" a', b' (CIE-LAS) colour parameters. The b'value was not suitable for estimating the B-carotene content of butter, whereas storage defects(non-enzymatic browning reactions) of whey powder could be monitored using this parameter.

colour 1milk product 1reflectance colorimetry 1tristimulus technique

Résumé - Méthode tristimulus (réflexion colorimétrique) pour la mesure de la couleur du laitet des produits laitiers. La technique de réflexion tristimulus a été appliquée à la mesure objective dela couleur d'un certain nombre de laits et de produits laitiers du commerce. Parmi ceux-ci, on trouvedes liquides, des produits fermentés frais tels que yaourts, diverses sortes de fromages, des beurresd'été et d'hiver et des poudres de lait, caractérisés par leurs valeurs L',a' et b' selon la CIE. La com-posante jaune b' s'est révélée inadéquate à estimer la teneur en fJ-carotène du beurre, mais permetde suivre les altérations subies par une poudre de lactosérum en cours de stockage (brunissementnon enzymatique).

couleur / produit laitier / photométrie de réflexion / méthode tristimulus

INTRODUCTION

ln general, colour and shape are the ma-jor properties that give objects their indi-vidual characters. As far as foods are con-cerned, other sensations su ch as srnell,taste and texturai attributes contribute tothe ove rail quality of these products. Nev-ertheless, in many cases food colour isthe first criterion to be perceived by the

consumer. It is weil known that the re-peated recognition of a particular brand ofa Iood commodity largely depends on itstypical colour. Thus, in the food industrythe assessment of the colour of foods andits components has become an integralpart of total quality control. Since reliablémethodology for the objective measure-ment of colour has been developed, thistechnique has found widespread use in

384 W Kneilel et al

many food sectors. For instance, intrumen-tal assessment of the colour of meat andmeat products (Stolle and Paulick, 1990),egg yolk (McCready et al, 1973) fruits andvegetables (Kader and Morris, 1978;Wainwright and Hughes, 1989), sweetsand chocolate (Ugrinovits, 1987; Kneifel etal, 1990) and coffee (Francis and Clydes-dale, 1975) has been described. Severalreports concerning milk and dairy productscan also be found in the Iiterature (Bossetet al, 1977, 1979, 1983a,b, 1986; Kam-merlehner and Kessler, 1979; Desarzenset al, 1983; Desarzens, 1988; Giangiaco-mo and Messina, 1988, 1989; Kneifel et al,1992).

The purpose of this paper was to dem-on strate the potential inherent in objectivecolour measurement and to present a sur-vey of the colour of different milk productsas estimated by a tristimulus reflectancetechnique.

MATERIALS AND METHODS

Samp/e materia/s

A variety 01 milk and dairy products was pur-chased lrom local retail outlets. Whey and milkpowder samples were provided by different Aus-trian plants. For storage experiments, wheypowders with different water content were pre-pared by conditioning the products in an atmos-phere 01 delined humidity provided by cabinetscontaining delined salt solutions. Retail lard wasused as "relerence material" (matrix) lor colourmeasurements 01 butter. 13-Carotenewas pur-chased from Sigma Chemicals (St Louis, USA).

Heat treatments of milk

Screw-capped glass tubes containing 20-ml por-tions of raw milk (4.1% fat) were immersed in aboiling water-bath until the required heatingtemperature was reached. Therealter, the sam-

pies were held at 70, SOand 90 oC for 1 and 5min respectively in thermostated water-baths.An oil-bath was used for heat-treatment at 100and 120 "C using the same time conditions. Sam-pies were cooled in an ice-bath alter heating.

çotour measurement

A Microcolor tristimulus colorimeter (Dr BrunoLange GmbH, Berlin, Germany) was used forcolour testing. Calibration was performed usingthe Dr Lange "White-standard" UM 076 (stan-dard tristimulus values: X = 69.0; Y = 73.5; Z =77.0) as specified by the manufacturer. Themeasuring principle 01 this apparatus is basedon a d/Sooptical structure, 10° standard observ-er, D 65 standard illuminant. A xenon flash lampwas the light source. Each sample was tested in4 replicates. Results were expressed using theL*, a*, b*-system according to CIE-LAB (Com-mission Internationale de l'Éclairage, 1971). Inthis system, L* defines the position of the sam-pie on the dark-light axis, e: on the green-redaxis, and b* on the blue-yellow axis.

Liquid and semi-solid products were filied tothe engraved mark of a "Iiquid sample" quartzcuvette, and subsequently covered with a PTFEpiston. Care was taken not to include air bub-bles in the liquid. Fruit yogurts were stirred witha spoon and subsequently poured through a 1-mm sieve to remove larger partieles beforemeasurement. Powdered products were trans-ferred into the "powder sample" quartz cuvette.The filled cuvette was then tapped slightly on asolid support in order to ensure a homogeneoussample distribution. Both types of cuvette wereplaced on the head of the Microcolor measuringunit and covered with a Iid before starting themeasuring cycle. In the case of solid samplesIike cheese, the measuring unit was placed di-rectly onto the specimen which had been freshlyeut from the cheese sample. Ali samples weremeasured at 20 ± 1 "C alter an equilibrationtime of at least 1 h (Burton, 1956).

Other physica/and chemica/ parameters

Dry matter of powdered products was deter-mined according to FIL-IDF standard (Interna-

Col our measurement of dairy products

tional Dairy Federation, 1964). Sieve fractions ofmilk powder were collected as described byHaugaard-Sorensen et al (1978). Total hydroxy-methylfurfural (HMF) concentration was deter-mined according to the spectrophotometricmethod of Keeney and Bassette (1959). The ex-tent of homogenization was estimated as out-Iined by Schneider and Roeder (1979). Caro-tene content of butter fat was determinedaccording to Pardun (1969). Melted butter oilwas decolorized with charcoal, following themethodology proposed by Schaap and Rutten(1974).

RESUL TS AND DISCUSSION

Precision of cotour measurement

The precision of the colour reflectancemethod was determined by repeatedlymeasuring pasteurized who le milk (3.6%fat). The within-run relative standard devia-tion (RSD) (N = 10) was 0.06% for the L*value, 2.99% for b*, and 0.65% for a*. Thebetween-run RSD (N = 10) was 1.21 %,3.11 % and 1.44%, respectively.

cotour parametersof milk and dairy products

A characterization of the col our of differentdairy products is given in table 1. As can beseen from these data, Iiquid and culturedmilk products tested had slight 'green' and'yellow' components. The values found forpasteurized milk are partly different fromthose reported by Giangiacomo and Messi-na (1988) (L* = 88.2, a* = -4.35, b* = 5.40)and by Bosset and Blanc (1978) (L = 95.5,a = -2.0, b = 12.6). The observed differ-ences between our results and those re-ported by Bosset and Blanc (1978) wereobviously due to the fact that they used theHunter-L,a,b system. Compared to retailIiquid milk (3.6% fat, homogenized), the b*

385

value of set-style yogurt increased by 1.3units. As demonstrated by Giangiacomoand Messina (1989), this difference iscaused by the acidification and coagulationprocess. Most of the cheese types can becharacterized as exhibiting slight 'red' andpronounced 'yellow' colour components.The L* values of the liquid and culturedmilks indicated a high degree of whitenessand gave a rather consistent pattern, rang-ing from 81.7 to 87.5. Skimmed productstended to be lower in L* than their corre-sponding full-fat products (producing ahigher degree of light scattering). The col-our parameters of fresh and feta-typecheese closely resembled those data ob-tained with Iiquid products. On the otherhand, ripened cheese varieties showed amarked variation in colour values. lt hasbeen shown previously (Bosset et al,1977) that several parameters, eg texture(holes and cracks), surface properties, oilexudation, sam pie thickness and slicingtechnique can influence the results of col-our measurements on cheese samples.Mainly due to the typical colour of the addi-tives, the colour parameters of fruit yogurtsvaried as expected to a great extent.

The colour parameters of full-cream milkpowders differed from those of theskimmed milk powders (table 1). The rela-tive magnitude of this difference was most-Iy pronounced with respect to the b* value.However, it should be taken into consider-ation that in the CIE-LAB system the Y"parameter is used for the computation ofL*, a* and tr, meaning that ail parametersare interrelated. Obviously, the powder col-our is influenced by the fat content via theliposoluble ~-carotene. The colour of pow-dered milk products may also be influ-enced by technological parameters and bythe geographical as weil as c1imatic condi-tions of milk production (table 1). lt is fur-ther evident from the colour data given infigure 1 that there were no marked differ-ences in the L* and a* values of the sieve

386 W Kneifel et al

Table 1. Typical colour parameters of different milk products (mean values of at least 3 different repli-cate samples); C: dark (0), light (100); a*: green (-), red (+); b": blue (-), yellow (+).Composantes typiques de la couleur de produits laitiers différents (valeur moyenne d'au moins 3échantillons mesurés en triple); L *; (foncé) (0), clair (100); a * vert (-), rouge t-): b * bleu (-), jaune (+).

Product type Fat content (%) L* a* b*

Pasteurized milk < 0.1 81.7 -4.8 4.13.6 86.1 -2.1 7.84.5 86.2 -1:7 7.5

UHTmilk 2.5 86.0 -2.0 7.9Cream 36.0 88.1 -0.2 8.8Coffee cream 10.0 86.9 -0.5 8.6Cultured buttermilk 0.1 86.5 -2.6 6.9Cultured milk 3.6 87.5 -1.5 6.5Yogurt (set-style) 1.0 85.9 -2.5 8.8

3.6 86.6 -1.9 9.1Yogurt with fruits

apricot 3.2 82.3 1.3 10.7strawberry 3.2 77.0 9.1 4.9blueberry 3.2 52.9 20.6 -7.3raspberry 3.2 67.8 13.1 2.4

Yogurt dessert productvanilla 7.0 83.7 0.9 11.8coffee 7.0 66.5 5.3 18.7

Fresh soft cheese < 1.0 85.7 -0.9 10.410.0 86.1 -0.9 10.420.0 85.6 -0.3 10.640.0 85.0 1.3 10.7

Gervais 65.0 85.9 1.1 12.0Processed cheese 55.0 91.0 3.0 18.6Camembert (surface) 45.0 95.7 0.1 5.2

(interior) 85.6 3.3 26.9(surface) 60.0 94.6 0.5 5.9(interior) 86.8 3.2 27.7

Brie (surface) 60.0 95.9 0.2 4.1(interior) 90.2 2.7 26.5

Feta cheese 45.0 93.5 -1.1 11.0Roquefort 50.0 92.8 -1.4 14.5Tilsit chee se 35.0 77.2 3.1 28.8Tilsit Swiss type 45.0 79.9 3.1 24.1

25.0 72.9 4.0 27.6Edam cheese 45.0 79.8 4.2 32.2Gouda cheese 30.0 82.6 3.6 27.1Swiss type cheese 45.0 72.7 0.6 20.9Appenzell type cheese 45.0 71.9 2.2 25.5Full-cream milk powder 25.0 95.6 -3.6 19.8Skimmed milk powder

(Austrian origin) 1.0 94.9 -1.7 11.3(American origin) 1.0 92.5 -2.6 18.3(New Zealand origin) 1.0 93.0 -2.2 16.4

Colour measurement of dairy products

Size distribution %60 .------'---'-=-----------,

90 150 200

Mesh

L- 93,6 95,3 95,8 96,3 96,5

a- -3,7 -3,5 -3,5 -3,3 -3,6b· 20,8 20,1 19,3 18 19,8

Fig 1. Particle size distribution and colour para-meters L*, a* and b* of full-cream milk powder.Distribution des particules et des composantesL*, a * et b* de la couleur de la poudre de lait en-tier.

fractions collected from full-cream milkpowder. Only the b* values differed to acertain extent. This observation is also inagreement with the findings of Bosset et al(1979).

387

cotour of butter

Results of colour measurements on buttersamples are listed in table II. Colour differ-ences between summer and winter butterwere apparent and mainly due to differing~-carotene contents. Moreover, it is evi-dent from the data that the L*, a*, b* pa-rameters were strongly influenced by thesample temperature. This effect is obvious-Iy caused by the temperature-dependentextent of fat crystallization (solid/liquid ra-tio). In the case of colour measurementson butter, it is therefore particularly neces-sary to perform the tests under definedtemperature-time conditions to obtain astable crystal modification.

ln another series of experiments, an at-tempt was made to estimate the ~-carotene content of butter based on the b*values as an indicator for the yellow col-our. Different amounts of ~-carotene wereadded to a decolorized butter oil, and forreference purposes also to lard, which isknown to be completely colorless. The b*values measured are graphically present-ed in figure 2. Correlation coefficients (re-gression lines) calculated were 0.95 (y =3.020 + 2.010x; N = 6) for the butter oil,and 0.98 (y = 2.626 + 13.126x; N = 5) for

Table II. Colour parameters of butter samples at different temperatures.Composantes de la couleur des échantillons de beurre mesurés à différentes températures.

Butter type Sample temperature (%) L* a* b*

10 63.2 3.9 31.116 61.8 3.0 30.618 59.9 2.5 30.522 57.1 2.5 30.4

10 70.8 4.7 28.316 65.6 3.7 29.818 63.8 3.4 29.722 61.3 2.8 29.6

Summer butter

Winter butter

388 W Kneifel et al

b·value40.-------------~-__,

30

20

• Butterailo Lard

246 8

D-carotene U9/9la

Fig 2. Relationship between l3-carotenecontentand b* values of butter oil and lard, after addi-tion of known amounts of l3-carotene.Rapport entre la teneur en {3-carotène et les va-leurs b * de la matière grasse liquide du beurreet du saindoux enrichis en {3-carotène.

the lard samples. Although a close corre la-'tion was observed between B-caroteneconcentrations and b* values, an accu rateestimation of the 13-carotene content wasnot possible based on b* value measure-ment. The deviations of the chemically de-termined 13-carotene content from the re-sults obtained by colour measurementsmay be due to the varying crystal structureof the products as weil as to the depen-dence of b* on other colour parameters (egL*). A similar divergence is evident from theresults reported by Desarzens et al (1983)who were unable to find constant relation-ships between the vitamin B2 contents andthe L,a,b parameters of milk samples ex-posed to light for different time periods.

Changes of cotour during processingand storage of milk products

The formation of coloured products in thecourse of heating of milk or milk concen-

12

trates to high tempe ratures has been weildocumented (eg Horak and Kessler,1981). Using laboratory experiments, wewere not able to detect significant changesin the colour of non-homogenized milk un-der time-temperature conditions usuallyapplied for milk pasteurization within therange of 70-90 oC (table III). Pronouncedchanges could be reqistered only on se-vere heat treatment of milk in conditionsresembling autoclaving or UHT treatment.However, Bosset et al (1979) indicated asm ail but significant increase of the Hunt-er-L,a,b components with the temperatureapplied to homogenized milk.

It has been shown that non-enzymaticbrowning reactions also occur during pro-longed storage of dried milk products

Table III. Colour parameters of Iiquid milk hea-ted at different time-temperature conditions.Composantes de la couleur des laits chauffés atempérature et à durée contrôlées.

Heatingconditions

L* a* b * umol HMF/I

70 -c1 min 79.4 -7.5 4.8 3.25 min 79.6 -7.6 5.6 3.1

80 -c1 min 79.0 -7.2 4.7 3.15 min 79.5 -7.4 5.6 3.3

90 -c1 min 78.9 -7.4 4.3 3.05 min 80.1 -7.5 5.7 3.5

100 -c1 min 79.9 -7.4 5.0 3.25 min 80.7 -7.3 5.6 5.2

120 -c1 min5min20 min

80.3 -7.481.5 -5.881.5 -5.4

5.16.97.2

4.317.158.8

Colour measurement of dairy products

5

. .o

20,-----------,

15

3.00/010,!~:::~====~===1

;.-.,..,.5

0 .

20

15

5.0°/010

,

389

40°C HMF ",mol/100g60

.L.........-----~- 40

20

,--------...,.60

~:::::::==~=444O

20

'----'--_--'-_-'-_----'0

20

20 40 60 20 40o

(Renner, 1988; Kneifel, 1989). Hitherto,the HMF value has mainly been used todescribe these alterations. To demonstratethe relationship between powder colorationand HMF content, whey powder sampleswere stored at different temperatures (20,30, 40 oC) and sampled periodically (fig 3).As the water content of the product influ-ences the extent and the velocity of Mail-lard reactions, this parameter was adjustedto 1.5, 3.5 and 5.0% (w/w). The b* valueproved to be the most suitable indicator forthe detection of changes in colour. As canbe seen from these graphs, browning reac-tions of whey powders were pronounced athigh temperatures and water contents, re-spectively. For example, the HMF valuesof samples with a moisture of 1.5, 3.5 and

5.0% which were stored at 30 "C in-creased to 148%, 160%, and 162% of theirinitial values. By contrast, the correspond-ing b* values increased to 108%, 124%and 132%. Although the HMF value wasgenerally more sensitive in detecting thesealterations, reflectance colorimetry was amore rapid and simple means for the as-sessment of storage defects in whey pow-der.

CONCLUSIONS

Tristimulus colour reflectance measure-. ment is a tool which can be utilized to ob-tain additional objective and well-definedphysical data on milk and dairy products.

390 W Kneifel et al

This technique not only yields basic infor-mation on the colour of milk and dairyproducts, but also enables a precise con-trol of the food quality du ring manufactur-ing of selected dairy foods or during prod-.uct development. Depending on the locallegal situation, the colour values can beused as a basis for carrying out correc-tions of a product's colour (eg by additionof permitted food dyes or 13-carotene).

Particularly in the case of fruit yogurtsand dessert products, colour measure-ment may be of assistance in standardiz-ing the desired colour intensity.

REFERENCESBosset JO, Ruegg M, Blanc B (1977) La couleur

du fromage et sa mesure: essai de détermi-nation par photométrie de réflexion. SchweizMilchwirtsch Forsch 6, 1-6

Bosset JO, Blanc B (1978) Die Messung derFarbe von Milch und Milchprodukten mittelsRemissionsphotometrie(Tristimulus-Methode,System Hunter). XX Int Dairy Gongr, Paris

Bosset JO, Martin B, Blanc B (1979) La mesurede la couleur du lait et des produits laitierspar photométrie de réflexion (méthode tri-stimulus). Trav Ghim Aliment Hyg 70, 203-227

Bosset JO, Desarzens C, Blanc B (1983a) Laphotodégradation du lait et de quelques pro-duits laitiers. Partie 1 : Altérations de la cou-leur, du goût et de la teneur en quelques vi-tamines. Lebensm-Wiss Technol 17, 241-247

Bosset JO, Desarzens C, Blanc B (1983b) Laphotodégradation du lait et de quelques pro-duits laitiers. Influence de certains facteurschimiques et chimico-physiques sur l'alté-ration de la seule couleur. Lebensm-WissTechno/17,248-253

Bosset JO, Flückiger E, Lavanchy P, Nick B,PauchardJP, Daget N, Desarzens C, Dieffen-bacher A, Tagliaferri E (1986) Influence de latranslucidité et de la perméabilité aux gaz dedifférents matériaux d'emballage sur la qualitédu yaourt entier nature en cours de stockage.Lebensm-Wiss Techno/19, 104-116

Burton H (1956) Colour changes in heated andunheated milk. V. The effects of temperatureof measurement, pH, and the addition of cer-tain ions on the reflectance of separated milk.J Dairy Res 23, 92-104

Commission Internationale de l'Éclairage (CIE)(1971) Golour Measurement. CIE DocumentN015

Desarzens C (1988) Effet protecteur del'emballage contre la photo-oxydation : III.Étude du pH et de la couleur de diversyaourts en cours de stockage. Trav Ghim Ali-ment Hyg 79,378-391

Desarzens C, Bosset JO, Blanc B (1983) Laphotodégradation du lait et de quelques pro-duits laitiers. Partie 1: Altération de la couleur,du goût et de la teneur en quelques vita-mines. Lebensm- Wiss Techno/17, 241-247

Francis FJ, Clydesdale FM (1975) Food Golor-imetry: Theory and Applications. AVI PublCorp Inc, Westport, USA

Giangiacorno R, Messina G (1988) Determinazi-one oggetiva dei colore dei latte alimentaremediante colorimetria tristimolo. Sci Tec Latt-Gasearia 39, 21-39

Giangiacomo R, Messina G (1989) Applicabilitàdella colorimetra tristimolo per la valutazionedella struttura dei coagulo di yogurt. Ind Latte25,67-79

Haugaard-Sorensen l, Krag J, Pisecky J, West-ergaard V (1978) Bestimmung derTeilgrôl3enverteilung durch Siebung. MethodNo A8a. In: Ana/ysenmethoden tût Trocken-milch-Produkte. NS Niro Atomizer, Copen-hagen, Denmark

Horak FP, Kessler HG (1981) Die Farbmessungais Indikator hitzebehandelter Lebensmittelam Beispiel von Milchprodukten. Z LebensmTechnol Verfahrenstech32, 180-184

International Dairy Federation (1964) Determina-tion of water content of dried milk. FIL-lOFstandard No 26

Kader AA, Morris LL (1978) Tomato fruit colourmeasured with an Agtron E5-W reflectancespectrophotometer. Hortscience 13, 577-578

Kammerlehner J, Kessler HG (1979) Farbmes-sung an Butter zur objektiven Beurteilung ih-rer Farbe und verschiedener Einflüsse. DtschMilchwirtsch 30,758-762

Keeney M, Bassette R (1959) Detection of inter-mediate compounds in the early stages of

Colour measurement of dairy products 391

browning reactions in milk products. J DairySei 42, 945-960

Kneifel W (1989) Lagerungsstabilitat von VolI-michpulver unter besondererBerücksichtigung des Vitaminspektrums. 1.Chemisch-physikalische Parameter und Er-gebnisse von Provokationsversuchen. Mi/eh-wissensehaft 44, 607-611

Kneifel W, Ulberth F, Hoch G (1990) A rapid col-orimetrie method for the estimation of cocoacontent in milk powdsr-cocoa mixtures. Le-bensm Wiss Teehno/23, 289-291

Kneifel W, Ulberth F, Schaffer E (1992) Evalua-tion of coffee whitening ability of dairy prod-ucts and coffee whiteners by means of reflec-tance colorimetry. Mi/ehwissensehaft (inpress)

McCready ST, Fry JL, Hinton CF, Harms RH(1973) Colorimetrie characterization of eggyolk and egg yolk products. J Food Sei 38,175-176

Pardun H (1969) Analyse der Fette und Fettbe-gleitstoffe. In: Handbueh der Lebensmittel-ehemie, Vol 4: Fette und Lipoide(Schormüller J, ed) Springer-Verlag, Berlin

Renner E (1988) Storage stability and somenutritional aspects of milk powders andultra high temperature products at highambient temperatures. J Dairy Res 55, 125-142

Schaap JE, Rutten GAM (1974) Colorimetriemethod, based on the carotene content, forthe estimation of the percentage of solids inbutterfat with crystals. Neth Mi/k Dairy J 28,166-174

Schneider K, Raeder H (1979) Die PraktischeMi/ehprüfung und Kontrolle von Molkereipro-dukten. KJ Wyss Erben AG, Bern, Switzer-land

Stalle FA, Paulick C (1990) Ausblutungsgradbei Schlachttieren. Mêiglichkeiten einerobjektiven Bestimmung und Beurteilung.Fleisehwirtsehaft 70, 274, 279-280, 282-284,300

Ugrinovits MH (1987) Instrumentelle Farbbeur-teilung von Kakaosuspensionen. Alimenta 2,34-39

Wainwright H, Hughes PA (1989) Objectivemeasurement of banana pulp colour. Int JFood Sei Teehno/24, 553-558