To determine the effect of feeding formula containing long-chain PUFA (LCP) on immune function, healthy term infants were randomised at age 2 weeks to either a standard term formula (Formula; n 14) or the same formula supplemented with the LCP 20 : 4n-6 and 22 : 6n-3 (Formula+LCP; n 16). Peripheral blood was collected at 2 and 6 weeks to measure immune cell response (the rate of [3H]thymidine uptake and cytokine production after stimulation with phytohaemagglutinin (PHA)). Compared with cells from infants receiving only human milk (HM), the rate of [3H]thymidine uptake in response to PHA, but not IL-2 production, was lower for Formula+LCP infants (P < 0·05). Compared with HM-fed infants, Formula-fed infants (but not Formula+LCP infants) produced more TNF-α (unstimulated) and had a fewer CD3+CD44+ cells before stimulation and fewer CD11c+ cells post-stimulation (P < 0·05). However, compared with Formula-fed infants, the Formula+LCP infants had an immune cell distribution (higher percentage CD3+CD44+ and CD4+CD28+ cells) and cytokine profile (lower production of TNF-α post-stimulation) that did not differ from HM infants. Additionally, it was found that feeding infants formula during the first 10 d of life influenced immune function. These infants had a higher percentage of CD3+, CD4+CD28+, and lower percentage of CD14+ cells and produced more TNF-α and interferon-γ after PHA stimulation than HM-fed infants (P < 0·05). These results demonstrate that early diet influences both the presence of specific cell types and function of infant blood immune cells. Since many diseases have a strong immunological component, these immune changes may be of physiological importance to the developing infant.