Polyunsaturated fatty acids (PUFAs), such as arachidonic acid (ARA) and docosahexaenoic acid (DHA) are natural constituents found in human milk, fish oil or egg yolk. Until recently, infant formulas, though providing the essential fatty acid precursors for these PUFAs, did not contain preformed ARA or DHA. In this study the safety of SUNTGA40S as source of ARA, not only for use in infant formulas but also for nutritional products or food supplements, was evaluated in a subchronic study in Wistar rats, preceded by a 4-week pretreatment period of parental (F0) rats and exposure of the F0 dams throughout mating, gestation and lactation. SUNTGA40S was administered at dietary levels of 0.5%, 1.5% and 5% (wt/wt) adjusted with corn oil to 5.76% added fat. An additional group received 3.65% (wt/wt) SUNTGA40S in conjunction with 2.11% (wt/wt) high DHA Tuna oil, providing an ARA:DHA ratio of 2.7:1. High-fat and low-fat controls received basal diet with or without 5.76% corn-oil supplement. The content, stability and homogeneous distribution of the test substances in the diet were confirmed under study conditions. The administration of SUNTGA40S, with or without DHA oil, did not affect health, growth, fertility or reproductive performance of the parental rats, nor pup characteristics (condition, weight gain, viability, number per litter or sex ratio). In the subchronic study with the offspring (F1) rats, no significant differences were found in condition, neurobehavioural observations, ophthalmoscopy, growth, urinalysis or macroscopic and microscopic findings between the test groups and the low-fat or the high-fat controls. In males of the 5% SUNTGA40S and the SUNTGA40S/DHA group, red blood cell counts, haemoglobin concentration and packed cell volume were lower and reticulocytes were slightly higher than in the high-fat and low-fat control groups. Cholesterol, triglycerides and phospholipids in plasma were lower than in the high-fat controls in both sexes in the 5% SUNTGA40S and the SUNTGA40S/DHA group and (for triglycerides only) in the 1.5% SUNTGA group. Due to the administration of extra dietary fat, food intake and prothrombin time (males only) were lower and alkaline phosphatase activity was higher in all the high-fat groups, including the corn-oil controls, as compared to the low-fat controls. The weight of the spleen was higher in males of the 5% SUNTGA40S and the SUNTGA40S/DHA group compared to both the low-fat and the high-fat controls. The effects noted in this study at high dose levels of SUNTGA40S are consistent with previously reported physiological responses to dietary intake of high PUFA containing oils. The present results provide evidence that SUNTGA40S is a safe source of arachidonic acid. Except during lactation when the intake in dams doubled, 5% Suntga40S in the diet was equivalent to an overall intake of approximately 3 g/kg body weight/day in F0 and F1 animals. © 2005 Elsevier Ltd. All rights reserved.