Functional foods have been described as foods that provide benefits beyond basic nutrition and may prevent disease or promote health. A functional food ingredient that has received much attention is fructooligosaccharides (FOS). Fructooligosaccharides occur naturally in many plants (Mitsuoka et al. 1987, Spiegel et al. 1994, Tashiro et al. 1992) and have been shown to exhibit beneficial health effects by stimulating the growth of bifidobacteria in the human colon, by suppression of putrefactive pathogens and by reduction of serum cholesterol concentrations (Gibson and Roberfroid 1995, Hidaka et al. 1986, Tomomatsu, 1994). Although the health benefits derived from the colonic fermentation of FOS in humans are well documented (Gibson and Roberfroid 1995), little is known of a potentially similar role in companion animals. Research with another group of oligosaccharides (i.e., lactosucrose) has shown positive changes in the colonic bacterial population and significant reductions in the concentrations of fecal odor components in dogs (Terada et al. 1992) and cats (Terada et al. 1993) as a result of lactosucrose supplementation. Dietary supplementation of FOS, if proven effective in promoting a more remedial colonic bacterial population and in decreasing fecal odor components, will depend on FOS contribution from pet-food ingredients. Therefore, the objective of this study was to establish a database on the concentrations of FOS in 25 pet-food ingredients. Lewis (1993) defined FOS as a mixture of 1-ketose (1-kestotriose; [GF.sub.2]), nystose (1,1-kestotetraose; [GF.sub.3]), and [1.sup.F]-fructofuranosyl-nystose (1,1,1-kestopentaose; [GF.sub.4]). These oligosaccharides [ILLUSTRATION FOR FIGURE 1 OMITTED] consist of short chains of fructose units linked by (2[right arrow]l)-[Beta] - glucosidic bonds and carry a single D-glucosyl unit at the nonreducing end of the chain linked (1[right arrow]2)-[Alpha]-as in sucrose. Materials and methods. Ingredients. A total of 25 pet-food ingredients (alfalfa meal, barley, beet pulp, brewer's rice, canola meal, corn, corn distiller's solubles, corn gluten feed, corn gluten meal, hominy, milo, oats, oat groats, peanut hulls, rice bran, brown rice, white rice, rice hulls, seaweed, soybean hulls, soybean meal, wheat, wheat bran, wheat germ and wheat middlings) were obtained from local commodity stores and ground through a 2-mm screen. Samples were dried (AOAC 1984) to allow expression of FOS content on a dry matter (DM) basis. Sample preparation. Approximately 10 g of the ground sample was blended (Waring blender, Waring Products Division, New Hartford, CT) with 50 mL of water for 5 min. The blended mixture then was filtered through a Buchner funnel containing a Whatman # 1 filter paper and the filtrate was subjected to further centrifugal filtration using a 10 Da cut-off filter (AMICON, Beverly, MA). FOS separation. The FOS content of the samples was determined by HPLC. Each sample (25 [[micro]liter]) was injected into a Dionex BioLC HPLC (Dionex, Sunnyvale, CA) fitted with a CarboPac PAl (4 x 250 mm) analytical column and a CarboPac PAl (4 x 50 mm) guard column (Dionex). The degassed mobile phase, consisting of 100 mmol NaOH/L, was initially run for 8 min. From 8 to 30 min, the proportion of 100 mmol NaOH/L and 100 mmol NaOH/L:600 mmol NaOAc/L was altered linearly to a final ratio of 88% 100 mmol NaOH/L and 12% 100 mmol NaOH/L:600 mmol NaOAc/L. This concentration was maintained for 6 min and then the eluants were changed to a concentration of 50% 100 mmol NaOH/L and 50% 100 mmol NaOH/L:600 mmol NaOAc/L, which was maintained for 5 min to clean the column. The column then was reequilibrated for 19 min with 100 mmol NaOH/L. All eluants were prepared with carbonate-free water and purged with helium. The flow rate was constant at 1 mL/min and the elution was conducted at room temperature. Eluted FOS units were detected using a Dionex Pulsed Electrochemical Detector (Dionex) equipped with a gold working electrode. The detector was operating in the integrated amperometry mode. Total run time was 60 min per sample. Appropriate dilutions of a solution containing each of the [GF.sub.n] units (Wako Pure Chemical, Osaka, Japan) were used as the calibration standards. Chromatographic peaks were identified by comparing sample retention times to those of known standard mixtures. Calculations. Components of FOS ([GF.sub.n] units) were quantified by measuring peak areas and comparing them with standard curves generated by plotting area counts against concentration of standards (0-250 [[micro]gram]/mL). Linear regression was used to calculate the calibration curve and the coefficient of determination ([r.sup.2]). The [r.sup.2] value was 0.98 for all [GF.sub.n] units. All samples were analyzed in duplicate with and without spike, thus allowing recoveries to be calculated. The percentage of recovery was calculated as follows: % recovery = [(C1 - C2)/S] x 100 where C1 is the concentration of sample with spike, C2 is the concentration of sample without spike and S is the concentration of spike. Results and discussion. No FOS were detected in corn, corn distiller's solubles, hominy, milo, brown rice, white rice, brewer's rice, rice hulls, seaweed, or soybean meal. Wheat coproducts contained the highest concentrations of total FOS, followed by peanut hulls, alfalfa meal, barley and wheat. The remaining ingredients contained low concentrations of FOS (Table 1). The proportions of [GF.sub.4] (percentage of total FOS) were highest in alfalfa meal, peanut hulls, and wheat germ (96, 83 and 50%, respectively). The proportions of [GF.sub.2] (percentage of total FOS) were 100% for ingredients containing low concentrations of FOS (i.e., beet pulp, canola meal, oats, oat groats, rice bran and soybean hulls). The proportions of [GF.sub.2] in ingredients containing high concentrations of FOS (i.e., wheat bran, wheat germ, wheat middlings and barley) were 95, 46, 100 and 76%, respectively. The proportions of [GF.sub.3] (percentage of FOS) were very low in all ingredients except for barley (22%), corn gluten meal (56%) and corn gluten feed (78%) (Table 1). The ranges for percentage recovery of FOS in the 25 ingredients evaluated in this study were 84 to 122% for [GF.sub.2], 89 to 107% for [GF.sub.3] and 85 to 119% for [GF.sub.4]. The majority of the percentage recovery values were 100 [+ or -] 10%. In summary, accurate quantification of FOS in pet-food ingredients will allow more precise nutritional formulation to be developed such that dogs and cats might derive maximal benefit from the inclusion of FOS-containing ingredients in their diets. KEY WORDS: fructooligosaccharides; companion animals; HPLC; ion exchange chromatography