Oxidative stability of soybean oils with altered fatty acid compositions

The oxidative stabilities of one canola oil and six soybean oils of various fatty acid compositions were compared in terms of peroxide values, conjugated dienoic acid values and sensory evaluations. Two of the soybean oils (Hardin and BSR 101) were from common commercial varieties. The other four soybean oils were from experimental lines developed in a mutation breeding program at Iowa State University that included A17 with 1.5% linolenate and 15.2% palmitate; A16 with 2% linolenate and 10.8% palmitate; A87-191039 with 2% linolenate and 29.6% oleate; and A6 with 27.5% stearate. Seed from the soybean genotypes was cold pressed. Crude canola oil was obtained without additives. All oils were refined, bleached and deodorized under laboratory conditions with no additives and stored at 60°C for 15 days. The A17, A16, A87-191039 and A6 oils were generally more stable to oxidation than the commercial soybean varieties and canola oil as evaluated by chemical and sensory tests. Canola oil was much less stable than Hardin and BSR 101 oils by both chemical and sensory tests. The peroxide values and flavor scores of oils were highly correlated with the initial amounts of linolenate (r=0.95, P=0.001). Flavor quality and flavor intensity had negative correlations with linolenate, (r=−0.89, P=0.007) and (r=−0.86, P=0.013), respectively.     

Determination of the Δ15 double bond in fatty acids of hydrogenated soybean oil

A method was developed to determine the extent of hydrogenation of the Δ15 double bond which occurs during partial catalytic hydrogenation of soybean oil. A linear relationship was found to exist between the linolenate content of commonly occurring C₁₈ unhydrogenated oils (containing no tetraene) and the propanal resulting from their ozonization reduction. The amount of propanal so produced is directly related to the amount of Δ15 double bond in these oils, as well as in hydrogenated soybean oils. Soybean oil was treated with ozone in carbon tetrachloride at —20 C and then reduced with triphenylphosphine. The ozonized-reduced sample was injected into a gas chromatograph, operated at 170C and equipped with a 12 ft × 1/4 in. column of 100/ 120 mesh porous polymer beads. The propanal peak was identified and its area used as a measure of the fatty acids containing Δ15 double bonds in unhydrogenated soybean and other oils of known linolenate content. A nearly stoichiometric amount of propanal results from ozonizing, reducing and chromatographing soybean oil as shown by comparison with a standard mixture of propanal and carbon tetrachloride. The relative standard deviation for the method is ±4.4%. We have also found this method applicable to other oils containing the omega-3 double bond. Presented at the AOCS-AACC Meeting, Washington, D.C., March, 1968. No. Utiliz, Res. Div., ARS, USDA.     

Oxidative stability of hydrogenated menhaden oil shortening blends in cookies,crackers, and snacks

The oxidative stability of partially hydrogenated menhaden fish oil (PHMO) shortening/canola oil blends with added antioxidant tertiary butylhydroquinone (TBHQ) and various blended partially hydrogenated vegetable oil (PHVO) shortenings without antioxidant in aged cookies and crackers was analyzed by anisidine value (AV), peroxide value (PV), and Totox value. The results showed no significant differences (P<0.05) for PV, AV, or Totox value between the PHMO shortening containing TBHQ and the PHVO shortening in cookies, crackers, and deep-fried extruded snacks, except for the AV and Totox value of crackers.     

Canola Quality Indian Mustard Oil (<Emphasis Type="Italic">Brassica juncea</Emphasis>) is More Stable to Oxidation than Conventional Canola Oil (<Emphasis Type="Italic">Brassica napus</Emphasis>)

Canola-quality Indian mustard (Brassica juncea) is being developed as a complimentary oilseed crop to canola (Brassica napus) for cultivation in hot and low-rainfall areas, where canola does not perform well. In Australia, several B. juncea breeding lines have been developed for commercial cultivation and for eventual processing as canola oil. However, there still are significant species-based differences in the fatty acid composition with B. juncea containing lower levels of linoleic acid and higher levels of oleic and linolenic acids compared with B. napus. This has raised concern about possible oxidative stability differences between the oils. Oils (unrefined) extracted from different breeding lines of each species were subjected to accelerated autoxidation, and development of oxidative rancidity was assessed by four separate techniques: depletion of polyunsaturated fatty acids, depletion of tocopherol, development of primary oxidation products, and development of secondary oxidation products (propanal and hexanal). All the tests showed that the newly developed B. juncea oils are more stable to autoxidation than conventional canola (B. napus) oil, despite containing marginally higher linolenic levels. Oxidative stability does not appear to be a barrier to using oils from these emerging lines of B. juncea for partial or full replacement of conventional canola oil.     

Linolenic Acid as the Main Source of Acrolein Formed During Heating of Vegetable Oils

Acrolein, which is an irritating and off-flavor compound formed during heating of vegetable oils, was estimated by the gas–liquid chromatography (GLC). Several vegetable oils such as high-oleic sunflower, perilla, rapeseed, rice bran, and soybean oils were heated at 180 °C for 480 min and then the concentration of acrolein in the head space gas was determined by GLC. The formation of acrolein was greatest in perilla oil among the tested oils, while it was much lower in rice bran oil and high-oleic sunflower oil. There was a good correlation between the level of acrolein and linolenate (18:3n-3) in the vegetable oils. To investigate the formation of acrolein from linolenate, methyl oleate, methyl linoleate, and methyl linolenate were also heated at 180 °C, and the amounts of acrolein formed from them were determined by GLC. The level of acrolein was the greatest in methyl linolenate. Acrolein was also formed from methyl linoleate, but not from methyl oleate. Acrolein in vegetable oils may be formed from polyunsaturated fatty acids, especially linolenic acid but not from glycerol backbone in triacylglycerols.     

Frying quality and oxidative stability of high-oleic corn oils

To determine the frying stability of corn oils that are genetically modified to contain 65% oleic acid, high-oleic corn oil was evaluated in room odor tests and by total polar compound analysis. Flavor characteristics of french-fried potatoes, prepared in the oil, were also evaluated by trained analytical sensory panelists. In comparison to normal corn oil, hydrogenated corn oil and high-oleic (80 and 90%) sunflower oils, high-oleic corn oil had significantly (P<0.05) lower total polar compound levels after 20 h of oil heating and frying at 190°C than the other oils. Fried-food flavor intensity was significantly higher in the normal corn oil during the early portion of the frying schedule than in any of the high-oleic or hydrogenated oils; however, after 17.5 h of frying, the potatoes fried in normal corn oil had the lowest intensity of fried-food flavor. Corn oil also had the highest intensities of off-odors, including acrid and burnt, in room odor tests. High-oleic corn oil also was evaluated as a salad oil for flavor characteristics and oxidative stability. Results showed that dry-milled high-oleic corn oil had good initial flavor quality and was significantly (P<0.05) more stable than dry-milled normal corn oil after oven storage tests at 60°C, as evaluated by flavor scores and peroxide values. Although the high-oleic corn oil had significantly (P<0.05) better flavor and oxidative stability than corn oil after aging at 60°C, even more pronounced effects were found in high-temperature frying tests, suggesting the advantages of high-oleic corn oil compared to normal or hydrogenated corn oils.     

Oxidative stability of fat substitutes and vegetable oils by the oxidative stability index method

Oxidative Stability Index (OSI) of carbohydrate fatty acid polyesters, fat substitutes and vegetable oils were measured with the Omnion Oxidative Stability Instrument according to the new AOCS Standard Method Cd 12 B-92 (The Official Methods and Recommended Practices of the American Oil Chemists' Society, edited by D. Firestone, AOCS, Champaign, 1991). The stability of crude and refined, bleached and deodorized (RBD) vegetable oils (soybean, hydrogenated soybean and peanut) were determined at 110°C. In addition, OSI times for sucrose polyesters of soybean oil, butterfat, oleate:stearate and methyl glucoside polyester of soybean oil were determined in the absence and in the presence of 0.02 wt% antioxidants, [Tenox TBHQ (tertiary butylhydroquinone, Tenox GT-2 (from Eastman Chemical Products (Kingsport, TN); and vitamin E (from BASF, Wyandotte, MI)], and the results were compared with those of vegetable oils. Crude oils were most stable (20.4–25.9 h), followed by RBD oils (9.3–10.4 h) for soybean and peanut oils, respectively, and fat substitutes (3.8–6.8 h). Overall, Tenox TBHQ was the best antioxidant for improving the oxidative stability of both vegetable oils and fat substitutes. The sucrose polyester made with oleic and stearic acid was more stable than fat substitutes containing more polyunsaturated fatty acids, such as those from soybean oil, or from short-chain fatty acids, such as from butterfat. Antioxidants enhanced the stability of RBD oils (222% increase) and synthetic fat substitutes (421–424% increase) more than that of crude oils (33% increase). The shapes of the induction curves, not the actual OSI times for fat substitutes and vegetable oils, were similar and sharply defined.     

Oxidative Stability of Conventional and High-Oleic Vegetable Oils with Added Antioxidants

The oxidative stability of conventional and high-oleic varieties of commercial vegetable oils, with and without added antioxidants, was evaluated using the oil stability index (OSI). Oil varieties studied were soybean (SOY), partially-hydrogenated soybean (PHSOY), corn (CORN), sunflower (SUN), canola (CAN), high-oleic canola (HOCAN), very high-oleic canola (VHOCAN), oleic safflower (SAF) and high-oleic sunflower (HOSUN). One or more commercial antioxidants were added to the four most stable oils at supplier-recommended levels: rosemary extract (RM; 1,000 ppm), ascorbyl palmitate (AP; 1,000 ppm), tert-butylhydroquinone (TBHQ; 200 ppm), and mixed tocopherols (TOC; 200 ppm). OSI in hours (h) at 110 °C of the conventional oils were 5.2, 7.6, 8.4, 9.8, 10.9 and 14.3 h for SUN, SOY, CAN, CORN, PHSOY and SAF, respectively. OSI of high-oleic variants were 12.9, 16.5 and 18.5 h for HOCAN, HOSUN and VHOCAN, respectively. Maximum OSI values for the four most stable oils when treated with antioxidants, were 40.9, 48.5, 48.8 and 55.7 h for HOCAN, VHOCAN, SAF and HOSUN, respectively. Addition of TBHQ, alone and in combination with other antioxidants, resulted in the greatest increase in oxidative stability of SAF and other high-oleic oils evaluated. AP had a positive synergistic effect when used with TBHQ, while RM decreased TBHQ effectiveness.     

Flavor and oxidative stability of continuously hydrogenated soybean oils

Soybean oil was partially hydrogenated in a continuous system with copper and nickel catalysts. The hydrogenated products were evaluated for flavor and oxidative stability. Processing conditions were varied to produce oils of linolenate contents between 0.4 and 2.7%, as follows: oil flow, 0.6–2.2 liters/hr; reaction temperature, 180–220 C; hydrogen pressure, 100–525 psig, and catalyst concentration, 0.5–1% copper catalyst or 0.1% nickel catalyst.Trans unsaturation varied from 8 to 20% with copper catalyst and from 15.0 to 27% with nickel catalyst. Linolenate selectivity was 9 with copper catalyst and 2 with nickel catalyst. Flavor evaluation of finished oils containing 0.01% citric acid (CA), appraised initially and after accelerated storage at 60 C, showed no significant difference between hydrogenated oils and nonhydrogenated oil. However, peroxide values and oxidative stability showed that hydrogenated oils were more stable than the unhydrogenated oil. CA+TBHQ (tertiary butylhydroquinone) significantly improved the oxidative stability of test oils over oils with CA only, but flavor scores showed no improvement. Dimethylpolysiloxane (MS) had no effect on either flavor or oxidative stability of the oils.     

Seed roasting improves the oxidative stability of canola (B. napus) and mustard (B. juncea) seed oils

Animal fats and partially hydrogenated vegetable oils (PHVO) have preferentially been used for deep‐frying of food because of their relatively high oxidative stability compared to natural vegetable oils. However, animal fats and PHVO are abundant sources of saturated fatty acids and trans fatty acids, respectively, both of which are detrimental to human health. Canola (Brassica napus) is the primary oilseed crop currently grown in Australia. Canola quality Indian mustard (Brassica juncea) is also being developed for cultivation in hot and low‐rainfall areas of the country where canola does not perform well. A major impediment to using these oils for deep‐frying is their relatively high susceptibility to oxidation, and so any processing interventions that would improve the oxidative stability would increase their prospects of use in commercial deep‐frying. The oxidative stability of both B. napus and B. juncea crude oils can be improved dramatically by roasting the seeds (165 °C, 5 min) prior to oil extraction. Roasting did not alter the fatty acid composition or the tocopherol content of the oils. The enhanced oxidative stability of the oil, solvent‐extracted from roasted seeds, is probably due to 2,6‐dimethoxy‐4‐vinylphenol produced by thermal decarboxylation of the sinapic acid naturally occurring in the canola seed.     

Formation of headspace volatiles by thermal decomposition of oxidized fish oilsvs. oxidized vegetable oils

To understand the reasons for differences in oxidative stability among edible oils, the temperature dependence was investigated for the development of volatile lipid oxidation products in fish oils and in vegetable oils. A rapid headspace capillary gas chromatographic method was developed to determine volatile oxidation products of omega-6 (n-6) polyunsaturated fats (pentane and hexanal) and omega-3 (n-3) polyunsaturated fats (propanal) at different decomposition temperatures. Headspace gas chromatographic analyses of partially oxidized menhaden, bonita and sardine oils could be performed at 40°C, whereas soybean, canola, safflower, high-oleic sunflower and high-oleic safflower oils required temperatures greater than 100°C. Volatile formation by thermal decomposition of oxidized oils had lower apparent activation energies in fish oils than in vegetable oils, and significantly higher apparent activation energies in high-oleic oils than in polyunsaturated oils. The activation energy data on headspace volatiles provided another dimension toward a better understanding of the thermal stability of flavor precursors in unsaturated fish and vegetable oils. Presented at the ISF/AOCS joint meeting, Toronto, Canada, May 10–14, 1992.     

Evaluating Electron Paramagnetic Resonance (EPR) to Measure Lipid Oxidation Lag Phase for Shelf-Life Determination of Oils

Lipid oxidation is one of the major causes of oil deterioration causing off-flavors and consumer rejection. Fast, easy, and dependable assays for predicting lipid oxidation rates in foods are important for shelf-life prediction. In this study, an electron paramagnetic resonance (EPR) spin-trapping technique with N-tert-butyl-α-phenylnitrone (PBN) was tested to determine the lag phase of lipid oxidation in stripped soybean oil (SSO), SSO with added α-tocopherol, and commercial soybean, canola and corn oils. EPR intensity of spin-trapped products from SSO correlated well with lipid hydroperoxides formation for samples stored at 37 and 55 °C respectively. When the antioxidant α-tocopherol was added, the EPR signal intensity of oil samples increased—indicating sample deterioration—after 50–65% of α-tocopherol was consumed. When using the EPR method with commercial soybean, canola or corn oil stored at 55 °C, there was a poor relationship between EPR intensity and lipid hydroperoxides lag phases. However, a linear correlation was found between EPR signal intensity and hexanal formation. For example, EPR signal intensity lag phases were 5, 13 and 27 days for soybean, canola and corn oils, respectively which was similar to the hexanal lag phases of 5, 13 and 25 days for the same oils. The EPR spin-trapping assay method has several advantages over headspace hexanal measurements, especially with regard to easier sample handling and shorter analysis times.     

AOCS collaborative study on sensory and volatile compound analyses of vegetable oils

An AOCS collaborative study was conducted to determine the effectiveness of sensory analysis and gas chromatographic analyses of volatile compounds in measuring vegetable oils for levels of oxidation that ranged from none to high. Sixteen laboratories from industry, government, and academia in Canada and the United States participated in the study to evaluate the flavor quality and oxidative stability of aged soybean, corn, sunflower, and canola (low-erucic acid rapeseed) oils. Analytical methods included sensory analyses with both flavor intensity and flavor quality scales and gas-chromatographic volatiles by direct injection, static headspace, and dynamic headspace (purge and trap) techniques. Sensory and volatile compound data were used to rank each of the oils at four levels of oxidation—none, low, moderate, and high. For soybean, canola, and sunflower oils, 85–90% of laboratories correctly ranked the oils by either analysis. For corn oil, only 60% of the laboratories ranked the samples according to the correct levels of oxidation by either analysis. Variance component estimates for flavor scores showed that the variation between sensory panelists within laboratories was lowest for the unaged oils. As storage time increased, the variance also increased, indicating that differences among panelists were greater for more highly oxidized oils. Between-laboratory variance of sensory panel scores was significantly lower than within-laboratory variance.     

Frying stability of soybean and canola oils with modified fatty acid compositions

Pilot plant-processed samples of soybean and canola (lowerucic acid rapeseed) oil with fatty acid compositions modified by mutation breeding and/or hydrogenation were evaluated for frying stability. Linolenic acid contents were 6.2% for standard soybean oil, 3.7% for low-linolenic soybean oil and 0.4% for the hydrogenated low-linolenic soybean oil. The linolenic acid contents were 10.1% for standard canola oil, 1.7% for canola modified by breeding and 0.8% and 0.6% for oils modified by breeding and hydrogenation. All modified oils had significantly (P<0.05) less room odor intensity after initial heating tests at 190°C than the standard oils, as judged by a sensory panel. Panelists also judged standard oils to have significantly higher intensities for fishy, burnt, rubbery, smoky and acrid odors than the modified oils. Free fatty acids, polar compounds and foam heights during frying were significantly (P<0.05) less in the low-linolenic soy and canola oils than the corresponding unmodified oils after 5 h of frying. The flavor quality of french-fried potatoes was significantly (P<0.05) better for potatoes fried in modified oils than those fried in standard oils. The potatoes fried in standard canola oil were described by the sensory panel as fishy.     

Current Status of High Oleic Seed Oils in Food Processing

Dietary trans fatty acids (TFA) from industrial partial hydrogenation continue to occupy the attention of health and regulatory authorities, prompting renewed recommendations and regulations around the world. Partial hydrogenation of liquid oils was widely used because it increases the oxidative stability and plasticity of vegetable oils. The development of high oleic (HO) oils appeared as an efficient and healthy strategy to replace partially hydrogenated vegetable oils. At present, the main sources of HO oils are the HO varieties of sunflower, canola, and soybean, covering almost the entire market. Although all these HO crops show oleic acid contents higher than 70%, they can be differentiated by their fatty acid profiles. Recent clinical evidence suggests that polyunsaturated content should also be considered since its intake has been associated with coronary heart disease. Comparison of the fatty acid compositions, places HO sunflower as the best option in terms of oxidative stability and beneficial effects on the health of consumers.     

High-temperature stability of soybean oils with altered fatty acid compositions

One canola oil and six soybean oils with different fatty acid compositions were heated intermittently, and bread cubes were fried in them to determine the stability of the oils. Two of the soybean oils were commercial varieties Hardin and BSR 101. The other soybean oils were from experimental lines developed at Iowa State University, and included A17 with 1.5% linolenate (18:3) and 15.1% palmitate (16:0), A16 with 1.9% 18:3 and 10.6% 16:0, A87-191039 with 1.8% 18:3 and 29.1% oleate (18:1) and A6 with 27.7% stearate (18:0). The soybean seeds were cold-pressed and crude canola oil was obtained without additives. Oils were refined, bleached and deodorized under laboratory conditions with additions. Each oil (300 mL) was heated to 180 ± 5°C in a minifryer. Bread cubes were fried at the beginning of heating, and half of the cubes were used for analyses. The second half was analyzed after storage at 60°C for seven days. Heating of the oils was continued for 20 h, cooled for 10 h, and then reheated for another 20 h, after which additional bread cubes were fried and analyzed. Results of sensory evaluation of the fried cubes, the peroxide values (PV) of oils extracted from the cubes and the conjugated dienoic acid values of the oils showed that the A17, A16, A87-191039 and A6 oils had better stabilities than did Hardin, BSR 101 and canola oils. The initial 18:3 contents of oils predicted their oxidative and flavor stabilities under heating and frying conditions (for PVvs. 18:3, r=0.89,P=0.008; for flavor qualityvs. 18:3, r=−0.93,P=0.002; for flavor intensityvs. 18:3, r=−0.91,P=0.004).     

Performance of Regular and Modified Canola and Soybean Oils in Rotational Frying

Canola and soybean oils both regular and with modified fatty acid compositions by genetic modifications and hydrogenation were compared for frying performance. The frying was conducted at 185 ± 5 °C for up to 12 days where French fries, battered chicken and fish sticks were fried in succession. Modified canola oils, with reduced levels of linolenic acid, accumulated significantly lower amounts of polar components compared to the other tested oils. Canola oils generally displayed lower amounts of oligomers in their polar fraction. Higher rates of free fatty acids formation were observed for the hydrogenated oils compared to the other oils, with canola frying shortening showing the highest amount at the end of the frying period. The half-life of tocopherols for both regular and modified soybean oils was 1–2 days compared to 6 days observed for high-oleic low-linolenic canola oil. The highest anisidine values were observed for soybean oil with the maximum reached on the 10th day of frying. Canola and soybean frying shortenings exhibited a faster rate of color formation at any of the frying times. The high-oleic low-linolenic canola oil exhibited the greatest frying stability as assessed by polar components, oligomers and non-volatile carbonyl components formation. Moreover, food fried in the high-oleic low-linolenic canola oil obtained the best scores in the sensory acceptance assessment.     

Effects of frying oil composition on potato chip stability

Potato chips were fried in six canola (low-erucic acid rape-seed) oils under pilot-plant process settings that represented commercial conditions. Oil samples included an unmodified canola oil and oils with fatty acid compositions modified by mutation breeding or hydrogenation. Chips were fried for a 2-d, 18-h cycle for each oil. Chips and oil were sampled periodically for sensory, gas-chromatographic volatiles and chemical analyses. Unmodified canola oil produced chips with lower flavor stability and oxidative stability than the other oils. The hydrogenated oil imparted a typical hydrogenation flavor to the chips that slightly affected overall quality. the modified canola oil (IMC 129) with the highest oleic acid level (78%) had the lowest content of total polar compounds and the lowest total volatile compounds at most of the storage times; however, the sensory quality of the potato chip was only fair. The potato chip with the best flavor stability was fried in a modified/blended oil (IMC 01-4.5/129) with 68% oleic acid, 20% linoleic acid and 3% linolenic acid.     

Flavor stability and quality of high-oleate and regular soybean oil blends during frying

The percentages of oleate (18∶1), linoleate (18∶2), and linolenate (18∶3) in blended soybean oils (SBO) were evaluated for their impact on flavor stability and quality in fried foods. Six SBO treatments, including a control (conventional SBO, 21.5% 18∶1) and a high-18∶1 SBO (HO, 79% 18∶1), were tested. In addition, these two oils were mixed in different ratios to make three blended oils containing 36.9, 50.7, and 64.7% 18∶1, abbreviated as 37%OA, 51%OA, and 65%OA, respectively. Also, a low-18∶3 (LL) SBO containing 1.4% 18∶3 and 25.3% 18∶1 was tested. Bread cubes (8.19 cm³) were fried in each of 18 oils (6 treatments ×3 replicates). The fresh and stored bread cubes fried in 79%OA were second to the cubes fried in LL in overall flavor quality, were the weakest in intensity of stale, grassy, fishy, cardboard, and burnt flavors by sensory evaluation, and contained the least amounts of hexanal, hexanal, t-2-heptenal, t,t-2,4-nonadienal, and t,t-2,4-decadienal in volatile analysis. Other treatments were intermediate in these sensory and instrumental evaluations, as related to their 18∶1, 18∶2, and 18∶3 concentrations. In general, the results suggested that the overall flavor stability and eating quality of foods fried in the six oil treatments from the best to the poorest would be: LL≥79%OA, 65%OA, 51%OA, 37%OA, and control.     

Sensory stability of canola oil: Present status of shelf life studies

Sensory studies on autoxidation of canola oil, stored under several variations of Schaal Oven test conditions, suggest an induction period of 2–4 d at 60–65°C. Similar induction periods have been observed between canola and sunflower oils, whereas a longer induction period has been found for soybean oil. Canola oil seems to be more stable to storage in light than cottonseed and soybean oils but is less stable than sunflower oil. Storage stability of products fried in canola oil is similar to products fried in soybean oil. Storage stability of canola and cottonseed oils that had been used in the frying of potato chips showed that canola oil was more prone to autoxidation during storage at 40°C. The presence of light aggravated the oxidative effects and was similar for both oils. Advances in our knowledge about the shelf life of canola oil would be strengthened by standardization of Schaal Oven testing conditions and by specifying the testing protocol for photooxidation studies. Methods for training of panelists and for handling and evaluating oils and fried foods require definition. Rating scales used in the evaluation of oils need to be evaluated to ensure that reliable and valid measurements are achieved. Further progress is needed in the identification of chemical indicators that can be used to predict sensory quality of oils. Presented in part at AOCS Annual Meeting in Toronto, Ontario, Canada, May 1992.