Determination of Animal Fat in Margarine by High Pressure Liquid Chromatography of the Sterols

The utility of high pressure liquid chromatography as a routine method to detect animal fats in vegetable margarines on the basis of sterol content was investigated. Cholesterol, which makes up almost 100% of the total sterol in animal fats, constitutes only 0-4% of the total sterols in vegeta• ble oils. The major phytosterols in vegetable oils currently used for margarine production are B-sitosterol, stigmasterol and campesterol. Brassicasterol is less conunon but in rapeseed oil it is present in high proportion. The minor phytosterols are: ~ 5-avenasterol, ~ -stiqmasterol, 7 ~ -avenasterol, 24-methyl cholest-7-enol and cholesterol. A uBondapak c18 column and a yariable wavelength ultraviolet detector along with a variety of mobile phases, have been used throughout this investigation. A method was developed for isolating, extracting and derivatizing sterols from biological material. Optimum chromatographic conditions are described. Lard, soy bean, corn, cottonseed, safflower, sunflower, olive, walnut and rapeseed oils were analyzed. A good separation was obtained for free sterols from vegetable o·ils. Free cholesterol could be separated from the major phytosterols but not from the whole group and in a

The major phytosterols in vegetable oils currently used for margarine production are B-sitosterol, stigmasterol and campesterol. Brassicasterol is less conunon but in rapeseed oil it is present in high proportion. The minor phytosterols are: ~5 -avenasterol, ~7-stiqmasterol, A uBondapak c 18 column and a yariable wavelength ultraviolet detector along with a variety of mobile phases, have been used throughout this investigation. A method was developed for isolating, extracting and derivatizing sterols from biological material. Optimum chromatographic conditions are described. Lard, soy bean, corn, cottonseed, safflower, sunflower, olive, walnut and rapeseed oils were analyzed. A good separation was obtained for free sterols from vegetable o·ils. Free cholesterol could be separated from the major phytosterols but not from the whole group and in a ii blend of 25% lard and 75% corn oil, cholesterol did not show up clearly.
With benzoate derivatives a good separation was obtained for cholesterol from major phytosterols, but brassicasterol from rapeseed oil and some minor phytosterols from other oils interfered with the cholesterol peak. Because of this interference, the detectable ratio of animal fat in vegetable margarine is dependent on the raw material.· Fifty percent of animal fat can be easily detected in all kinds of margarines.
In soy bean oil marqarine, 25% animal fat can be easily detected. The method is reproducible, rapid, and worth further investigation to find a column and solvent system capable of separating cholesterol completely from all phytosterols, so that very low levels of animal fat could be detected in vegetable oil.
iii ACKNOWLEDGEMENT · I wish

INTRODUCTION
Ghee is the name of butter fat obtained from cow's or sheep's milk, to which some flavors are added. That was the multipurpose butter in our country (Saudi Arabia) • Then, _./ the margarine expansion overwhelmed our market and consumers preferred the margarine for its convenience. Most of the import"ed margarine was labeled as "Substitute Ghee." Some labels specified that it had a vegetable and/or animal origin and under the name of "Substitute Ghee" a lot of animal fat was dumped on the country, especially when the prices of vegetable oils were high. To prevent such commercial fraud, to avoid the health problems of cholesterol, and due to the fact that lard is not allowed to enter the country on religious grounds, the Saudi Arabian government has prohibited the import of dripping fat and permits only margarine of vegetable origin. Usually, the margarine manufacturer has pre-set criteria for the finished fats and oils he uses CAnderson andWilliams, 1969 andMassiello, 1978), and different formulas geared to the cost of the oils which are the major ingredient cost in margarine. Prices of oils and fats change a good deal, like all commodity prices. The premium or special margarines are not subject to such adjustments.
The table following shows the fluctuations in lard use as an ingredient of margarine (Anderson and Williams, 1969 (Riepma, 1970).

14 15
Total 1500 1535 1720 Due to the fact that every imported margarine must be checked for the presence of animal fats, our Food Quality contrql Laboratory is in need of a quick routine method for the detection of animal fat in vegetable margarine.
Measuring cholesterol, a sterol formed predominantly in animals, is one of the most frequently performed assays in the laboratory, for which a wide variety of methods are available and used. The classical method is colorimetric assay based on the photometric measurement of the color formed when cholesterol reacts with Lewis acid (Abell et al., 1958).
However, because of the hazards associated with using the strong acid medium in which the color is formed, alternate methodologies have been and are being developed. Most of the newer methods are based on enzymatic hydrolysis and oxidation (Gray et al., 1977 andAnonymous, 1979), or on chromatographic analysis (Driscoll et al., 1971). The enzymatic reactions are followed by colorimetric or electrochemical analysis.
By gas chromatography alone, the determination of steroids in biological samples cannot usually be achieved. It is necessary to employ other procedures to separate crude fractions and isolate sterols by preparative methods. The isolation of such sterols from a biological material is commonly established by fractionation of lipid extracts by column chromatography or thin layer chromatography (TLC) and in many cases by silver nitrate-silica gel TLC of either the free sterols or their acetates lRees, 1976). Although liquid chromatography (LC) has been used in analysis of lipids in general (Aitzetmuller, 1975), its specific application in cholesterol methodology has been limited because • cholesterol and related compounds absorb very little UV radiation in the relatively high wave length in which most UV detectors used · in LC operate. Consequently, LC has been used in assays in which column chromatography is followed by chemical analysis of collected chromatographic fractions (Shin, 1963) or in which the eluate is monitored on-line with other types of detectors. These detectors include the refractive index detector (Werthessen et ai., 1970), the moving-wire flame ionization detector (Worth and MacLeod, 1969), and a laser infrared detector (Freeman, 1974 (Dam •. 1958 Steroids of some kind are apparently present in all living organisms, with the possible exception of bacteria. They include sterols, certain sapogenins, alkaloids, heart poisons and hormones. Each steroid in common possesses a characteristic tetracyclic carbon skeleton, namely, the skeleton of the perhydrocyclopentanophenanthrene molecule ( Figure 1) (perhydro = fully hydrogenated (Dence, 1980).
In completely saturated steroids, rings A, B, ~nd C are cyclohexane rings, and ring D is a cyclopentane ring. Cyclohexane rings are not planar, for if they were, the c-c-c bond angles would be 120° instead of the observed values which lie much closer to the tetrahedral angle of 109-110°.
Aromatic rings are planar.
In naming steroids, we begin by orienting the molecule as shown in Figure 2 for cholestane, a saturated C27 steroid. The numbering ' is as indicated. According to the International Union of Pure and Applied chemistry-International Union of Biochemistry (IUPAC-IUB) 1967, the use of a steroid name implies that groups at positions 8, 9, 10, 13, and 14 are oriented in the BB, 9a, lOB, 13B and 14a configurations, respectively, unless stated to the col'ltrary. We then need only to specify in the name the configuration at position 5 and 17 and any other positions of asymmetry. By long standing tradition, the terms a and Bare also applied as shown in structure of Figure 3. Many steroids have common or trivial names such as cholesterol.
Systematic names, however, are based on certain parent hydrocarbon systems. Elaboration is described using principles from organic nomenclature and expressing this by means of prefixes and suffixes. An example for sterols is given in Figure 4 (Dence, 1980).   frequently used solvents. One very unusual solubility property of steroids has been known for a long time. This is the fact that cholesterol and other 3B-hydroxy steroids are precipitated from solution .by digitonin, itself a glycoside of a 3B-hydroxy-spirostan, whereas 3a-hydroxy steroids are not so precipitated. The 3B-hydroxy steroid is precipitated as a complex of the glycoside. Other saponins will work also but digitonin was found to give nearly quantitative precipitation (Cook, 1958) . For ultraviolet and visible spec-• troscopy, the C=C and C=O bonds are the ones most important in steroids of natural origin. The C=N also plays an important role in derivatives used for analysis. The commonly measurable region is from 210 to 800nm of which the region from 400 to 800 is regarded as the visible and lower wavelength as the near ultraviolet region. Pure cholesterol, for instance, has no significant absorption from 210 to 800 nm. On the other hand. 5,6-dehydrocholestanol (cholesterol) 5 exhibits end-absorption because of a peak from the ~ bond near 200nm (Nes and McKean, 1977) • That is due to the fact that only certain transitions are allowed and they are dependent on the environment, extent of conjugation, degree of substitution and nature of the atoms joined by pi electrons (Pomeranz and Meloan, 1971  for it to appear consequentially in ,their tissues. There is a strong suggestion that this discrimination against sterols with modified side chains begins with the vertebrates. In many sub. sequent investigations cholesterol was found to be the dominant sterol of all higher vertebrates. Except for the gastrointestinal tract, mammalian sterol is nearly but not completely pure cholesterol. Vertebrate tissues usually not only contain no sterols with side chains which have been alkylated, shortened or enlarged, but also no 6 22sterols (Nes and McKean, 1977). However, the sterols of vertebrates are by no means always absolutely pure cholesterol. Conunon animal sterols include small amounts of the 6°and 6 7 -analogs (cholestanol and lathosterol), the 6 24 -derivative "(demosterol), 65 17 -derivative (7-dehydrocholesterol) and others (Nes and McKean, 1977 and plants but also between particular classes of plants (Goad, 1977) .
Cholesterol is the major sterol of animals. The typical plant sterols are: campesterol, sitosterol and stigmasterol ( Figure 7). The addition of a C at C24 of the side chain distinguishes phytosterols f:t0m.animal sterols. It is now recognized that cholesterol (cholest-5-en-3B-ol} is widespread in the plant kingdom, but it usualLy constitutes only a few percent of the sterol mixture (Goad, 1977). Weinrauch   (Riepma, 1970) .

~pes of margarines
First and most typical are margarines composed of vegetable oils (Massiello, 1978). A second is blend margarines which combine animal and vegetable fats in one F'°partion or another. They are commonly referred to as "AV" or "VA" products, depending on the predominance of animal ("A") or vegetable { "V") fat. The "AV" types are required by the standards to be between 50 and 90 percent animal fat.
A typical "AV" weight composition would be 90 percent pure lard and 10 percent soybean oil. The third is that of . margarines composed wholly of animal fats. Usually the fat content is sl ightly hardened pure lard {Riepma, 1970) . Today there are ten different types of margarine produced. There • are regular, ·::whipped and polyunsaturated margarines in both stick and soft forms. There are diet margarines, liquid aargarines and new 60% vegetable oil spreads. These products cater to the needs of many different segments of the population (Massiello, 1978). In recent years, medical research findings have suggested the advisability of increasing the ~roportion of polyunsaturated fatty acids in the diet and reduci.ng the intake of cholesterol. Parodi (1975)

Raw materials
The grease is collected from slaughtered animals. Then ) it is washed, minced and put into autoclaves with double jackets. Steam is blown in at normal pressure. The fat is drawn off, while the residue or greaves are collected on the screens of the autoclave. The fat is then crystallized slowly at about 30°C. The crystallized product, which is a granular mass, is filtered and pressed yielding about 60 percent liquid with melting point between 28 and 34° c.
The vegetable oils are obtained from oil-bearing seeds or fruits. The seed is crushed, heated moistened with a little steam and then pressed. The oil after that is refined, degurnrned, deodorized and hydrogenated. Hydrogenation consists of adding very pure hydrogen to oil, with a finely divided nickel catalyst (Van Stuyvenberg, 1969).

llll?tarine manufacture
The fats and oils in proper proportions are heated to 30 oc or more and oil soluble ingredients (emulsifiers, coloring agents, flavoring agents and vitamins) added. An aqueous phase containing milk or dried protein and water is prepared separately. After pasteurizing and cooling, salt and preservatives are added. The two phases are blended and the resulting emulsion is rapidly chilled in heat exchangers and allowed to stand in crystallizing chambers for 2 minutes until the product is sufficiently stable to be extruded and packaged. For soft tub margarine, the fluid melt is agitated during chilling to prevent fat crystals from growing into a firm network. The emulsifying agents (mono and diglycerides, sometimes lecithin) help to keep the water dispersed as fine droplets within the oil phase. Whipped margarines are produced by introducing nitrogen gas just prior to chilling and ~intaining vigorous agitation during the crystallizing step.
The low fat spreads with a higher water content require more emulsifying agents and more careful control of production conditions in order to maintain the water-in-oil emulsion (Brekke, 1980).

IJ!!lysis of sterols
Measuring cholesterol is one of the assays that have passed through several laboratory investigations and critical discussions since it was discovered and until now.
fenerally, any method for sterol determination involves one or more of the following stages: 1 Both free sterols and esterif ied sterols will be ex-23 tracted from biological tissue by common lipid solvents such as ethylether and chlorof~rm-methanol (Bligh and Dyer) and the alcohol and ester forms can be separated to some extent from other neutral lipids by silicic acid column chromatography and thin layer chromatography (Kates, 1972 andHeftman, 1975

Isolation of sterols
On saponifying the lipids with an alkali in alcoholic solution, the esterified fatty acids of the lipid, including the steryl esters, form alkali salts (soaps) and the liberated alcohols constitute the neutral fraction. After dilution with water, the higher alcohols and sterols are separated from the hydrolyzed material by extraction with an organic solvent such as petroleum ether (Kates, 1972) or ethyl ether (AOAC, 1980) which is immiscible in water.
The material recovered from the solvent forms the unsaponifiable matter. known as cholesterol digitonide (Cook, 1958). Digitonin is used frequently for the separation of 3B-hydroxy (precipi- After the precipitation is complete, it is necessary to split the digitonide into its components and to recover the The evaluation and criticism of digitonide formation method is as follows: A) The advantages are: (l) The low solubility of the complex makes the test very sensitive (2) The reaction has been used directly as a granimetric method for the determination of choleste·rol.
B) The disadvantages are: (1) The formation of insoluble complex is not restricted to digitonin but saponins, tiyonin and digitonin and the alkaloid formation give precipitates w·ith cholesterol.
Commercial digitonin frequently contains gitonin as impurity.
(2) Variable amounts of water of crystallization of the digitonide may affect the gravimetric results.
(3) Digitonin forms molecular complexes with many nonsteroid substances, e.g. phenols and terpene alcohols. After that, digitonin solution was added to precipitate the free cholesteroL. The dried precipitate of cholesterol diqitonide was dissolved in acetic acid. Next, the temperature of the water bath was adjusted to 25° and acetic anhydride was added followed by concentrated sulfuric acid.
The intervals between the addition of reagents was so timed that not less than 27 minutes nor more than 37 minutes elapsa:i between the addition of H2S0 4 and reading the color. -Once immobilized, an enzyme is of ten stable for weeks • or even months.
-It is easy to separate the insoluble enzyme derivative from reactants and/or products for reuse.
-The derivative can be packed in a column similar to those used in liquid chromatography or it can be used in a stirred tank with the product.
-The enzyme derivative has different catalytic properties from the enzyme and these properties can be controlled by immobilization conditions. Bottle (2)  It should be noted that the procedure is not specific for tholesterol since cholesterol oxidase oxidizes any sterols in which the hydroxyl . group at carbon 3 is in the B-position.
'rherefore , phytosterols, such as stigmasterol and sitosterol, also react in the assay.
This is no longer used · in steroid chromatography, except for some polar steroid conj~gates. Other chromatographic sheets have also lost the race, because they cannot resist the aggressive reagents and solvents used for sterols.
Glass-fiber paper is unsuitable because its coarse texture produces excessive diffusion of chromatographic zones llllr layer chromatography.
Thin layer chromatography of steroids has reached a state of perfection where one can hardly expect to devise major improvements in developing and detection methods.
Evaluation of chromatography by computer methods is simply a device for exploiting the wealth of information provided by TLC (Johnson, 1973 Gas cnromatography is a process by which a mixture is separated into its constituents by a moving gas phase passing over a sorbent. It has produced sweeping changes in many fields of research since its introduction in 1952.
Basically, it is a separation process capable of extremely high resolving power. Its direct applicability is limited to compounds which can be voiatilized without decomposition, such as therinally stable, nonionic compounds with a maximum ~lecular weight of around 400-500. This powerful analytical method has contributed significantly to advances in many fields -for example, catalyst research, biochemical studies and flavor analysis. Scaled-up in size, gas chroma-mwraphytography also offers a method for preparing very pure compounds.

38
Grunwald (1970)  Therefore, it is clear that different retention data demonstrate nonidentity, but identical data may be ambiguous. This is best shown for the case of sterols epimeric at C-24, because so far all analysis of such epimeric pairs of compounds by GLC have produced identical retention data for both isomers (Ikekawa, 19681 Although high pressure liquid chromatography has been used in analysis of lipids in general (Aitzetmuller, 1975} and in reference sterols separation (Colin ~ al., 1979}, its specific application in sterols methodology has been limited because cholesterol and related compounds absorb 4 _ Rees, et al. (1976) applied the HPLC to sterols separation. They recommended that the reversed phase separation of sterols on uBondapak c 18 c~n be applied to the preparative separation of sterols mixtures isolated from biological ~aterials by the conventional techniques of columns and thin layer chromatography. Colin, et al. (1979) have described a method for separating standard free sterols by high pressure liquid chromatography, using pyrocarbon modified silica gel column. Due to the limitation of their detector, they could not measure cholesterol at wavelength lower than 254nm, which is quite higher than cholesterol maximum absorption region for UV radiation. However, they recommended using this detector rather than Refractive Index detector. Newkirk, et al. (1981) have used high pressure liquid chromatography for determining cholesterol in foods.

-
They eluted cholesterol as its benzoate ester on uBondapak c 18 column. Detection was carried out with variable wavelength detector. When they compared their results with results obtained from gas chromatographic analysis, it was found that HPLC is a favorable technique. (2) Columns: uBondapak c 18 , 3.9mm x 30cm, and u-Porasil, 3.9mm x 30cm, Waters Associates.

Reagents
(1) HPLC solvents; acetonitrile, methanol, isopropyl alcohol, dichloromethane, and hexane were either Nanograde Dlallinckrodt) or Distilled in Glass (Burdick and Jackson) • UV-grade acetonitrile and isopropanol were used in some experiments. All solvents, including distilled water, were prefiltered through millipore or glass fiber filters.
(S) Petroleum ether and dichloromethane were Nanograde (Mallinckrodt) . l!terials (l} Beef fat, pork fat, vegetable oil and margarines were purchased at local supermarkets. Rapes~ed oil was extracted from crushed rapeseed. (2) Cholesterol, stigmasterol, campesterol and a plant aterol mixture were purchased from Applied Science, Supelco and steraloids.

IJlpo·nif ica tion
One gram of oil and/or margarine was weighed into a 50 ml glass tube equipped with Teflon-lined cap and 15 ml Of l.S M KOH in methanol was added. The tube was then sealed, kept in a water bath at l00°C for 20 minutes and allowed to· cool to room temperature. After cooling, 10 ml water and 1 5 ml petroleum ether were added. The tube was •haken for one minute and put in a centrifuge for three l'f.nutes at 2000 rpm to separate the PE layer from the soap.
The PE layer was then pipetted to a small beaker and evaporated to near dryness on a water bath and then under a stream of nitrogen.
lll"C of free sterols The residue of sterols obtained from PE evaporation was dissolved in 5 ml dichloromethane {DCM) , dried over anhydrous sodium sulphate and concentrated to 1 ml, from which 10 ul was injected into the LC.
Acety la tion • The sterol residue was dissolved in 1.2 ml acetic anhydride:pyridine, 2:1 (v/v) and transferred to a glass tube equipped with a Teflon-lined cap. The reaction solution was kept on a water bath at 37°C for 15 minutes and then on an ice bath for 2 minutes. After that, 0.6 ml PE, 0.6 ml acetone and 0.6 ml water were added. The tube was capped, shaken and put again in the 37°C water bath for 5 minutes.
Then the reaction solution was transferred to a small separatory funnel and washed four times with 2 ml water each.
The PE layer containing the sterols acetates was transferred from the separatory funnel, dried over anhydrous sulphate and concentrated to 1 ml, from which 10 ul was injected.

J!nzoylation
The sterol residue was dissolved in 1 ml acetone and transferred to a 20 ml qlass tube equipped with Teflonlined cap and 10 ml of 7.5 M sodium hydroxide in water and 0.2 ml benzoyl chloride were added. The tube was capped and shaken vigorously on a mixer for , 3 minutes. The reaction •ixture was then transferred to a separatory funnel containing 25 ml water and extracted two times with 15 ml DCM each.
The DCM extract was collected in another separatory funnel and washed two times with 20 ml of saturated sodium carbonate and one more time with 20 ml water. The DCM layer was then dried over anhydrous sodium sulphate and concentrated to 1 ml. Benzoates from standards were made by benzoylating 5 mg of the free sterol and diluting the final DCM extract to 10 ml . (4) The majority of runs were made with an attenuator ~etting of 0.1 and recorder chart speed of 0.5 cm/min.

1Pitial trials
To choose the column for this work both uPorasil and uaondapak c 18 , the two most common HPLC columns were tested for the resolution of free sterols and their benzoates. A variety of solvent systems were employed. All sterols showed sood elution individually with mixtures of isopropanol, acetonitrile, methanol, hexane and/or dichloromethane, but the separation of mixtures of sterols was better in the uBondapak c 18 column as shown later. Figure 9 shows the best separation obtained for free sterols on uPorasil.
Peaks 1 and 2 are solvent peaks, while Peaks 3 and 4 are unresolved sterols. Colin et al. (1979) also tried the nor--·mal phase for free sterols and found that lanosterol eluted first, then a whole group of mono-and di-unsaturated sterols and finally ergosterol which was well separated from others. The resolution of sterol benzoates on uPorasil was also poor {Figure 10). Accordingly, the use of such a column is not effective for routine analysis and the UBondapak c 18 (Figure 14) was the choice for this work.
The determination of sterols in unsaponified oil was attempted by dissolving 0.5 g of rapeseed oil or corn oil in 10 ml dichloromethane and injecting 10 ul into the column. Many peaks · followed this injection and so the chromatogram was confusing, as shown in Figure 11. Not only th e small amounts of free or esterif ied sterols (1% of are the total lipid) masked by other lipid components but injection of crude oil is impractical because it would require extensive analysis time to allow elution of all peaks before injecting another sample. Therefore, saponification was adopted for sterol determination.
Direct injection of a diluted (isopropanoll saponification mixture did not give good results due to the large • solvent peak, as shown in Figure 12. To prepare mobile phase (b} acetonitrile was added l+l to solvent system (a}. Figure 14 showed a good separation for corn oil sterols but stigmasterol unfortunately Still had a retention time similar to cholesterol.  up clearlv, perhaps due to the fact that the proportion of. • • I phytosterols in corn oil (1%} is much higher than that of cholesterol in lard lO .1%} (USDA, 1979) . In Figure 17, 25% lard in corn oil could barely be detected.  Figure 19. Acetates of cholesterol and stigrnasterol were successfully prepared by heating cholesterol and stigmasterol for 15 minutes in a water bath with acetic anhydride alone, but extracts of saponified oils did not chromatograph well a fter acylation. This approach was not pursued further for the following reasons: (1) The acetates of sterols • absorb UV radiation at very low wavelength (Grasselli and Ritchey, 1975) and were less sensitive than the benzoate derivatives, which could be analyzed at 230 nm, a more convenient wavelength, and (2) cholesteryl acetate could not be separated from some phytosteryl acetates. Rees et al. (1979 )  a sensitive detection at wavelength. 230 run and the benzoylation was reproducible. Fitzpatrick and Siggia (1973) used pyridine as solvent for the benzoylation reaction, while Blau and King (1978) describe a procedure using sodium hydroxide. Both procedures were attempted and use of sodium hydroxide was adopted for the following reasons: (1) Pyridine can form a compound with benzoyl chloride that might interfere in the chromatogram, (2) a lot of washings especially with hydrochloric acid are needed when pyridine is there. These are laborious and may cause loss of sample,  and Gardner (1978). The agreement with the theoretical is reasonably good. Unfortunately, the peak of major interest, the "cholesterol peak," shows the most discrepancies.   However, the patterns in different samples for the same kind of oil were shown to be reproducible ( Figure 26 and 27). Table 3 shows the results after analysis of several oil samples • .Analysis of mixtures of vegetable oils and lard 75 Cholesterol constitutes around 95% of the lard sterol content as shown in Figure 28 and in the literature (Dence, 1980) . Cholesterol is present in vegetable oils but at small aI\lOUnts, 0-4% of the totaL sterol (Weihrauch andGardner, 1978 andAppelqvist andOhlson, 1972). Vegetable sterols are not present in lard according to Figure 28 and Nes and McKean (1977). For the purpose of detecting animal fat in vegetable margarines, mixtures of lard and vegetable margarines were prepared at the ratios of 0, 25, SO, 75, and 100% lard. The sterols were extracted, benzoylated and analyzed by HPLC. Figures 29 and 30 and Tables 4 and 5 show the trend of cholesteryl benzoate peak to increase in accordance with the amount of lard added to soya bean oil margarine or corn oil margarine. From these results we can deduce that the cholesterol value is a good indication for the presence of lard in a vegetable margarine. If the peak exceeds 15%, it is highly likely that the margarine contains more than 25% animal fat. If the cholesterol is higher than 30%, the sample certainly contains over 50% animal fat.   lard ana 75% margarine, tel 50% lard am 50% margarine, CD> 75% lard am 25% margarine, am (E) 100% lard. uBomapak c 18 column with nd:>ile phase acetoni trile 100% at flow-rate 2 ml/min am detection at 230 rm.   (2) each.
• Two different margarines. Single determination of

Conclusion
The results obtained through this work have proved that high pressure liquid chromatography is a good tool for the analysis of sterols. Although, the separation was diff icult between cholesterol and some minor phytosterols, the detection of animal fat was possible. The utility of HPLC is practical when the fraudulation is more than 25% and critical at less proportions, especially if rapeseed oil were present in the margarine. However, HPLC performs separation' and detection at the same time without any preparative procedures such as thin layer chromatography or digitonin precipitation, which is a good factor for obtaining more accurate results. It is convenient for such routine analysis and time-savinq. Of its economical disadvantages, the deuterium lamp of the detector is expensive and it has only short life. What is recommended for achieving better results for this work is a column that can separate cholesterol from the entire vegetable oil sterol group.
Sterols standards were not abundantly available and their purity was often poor. So, it is recommended to have more advanced preparative procedures for obtaining less expens ive pure sterols, a task in which HPLC also could be a good tool.