A new procedure for non-surgical deep intrauterine insemination (DUI) in unrestrained sows hormonally induced to ovulate, has been reported. In comparison with standard artificial insemination (AI), with this procedure, the sperm numbers inseminated can be reduced 20-fold without reducing the reproductive performance of these hormonally treated sows. The present study evaluated, using two experiments, the reproductive performance applying 20-fold different sperm numbers per AI dose using DUI or standard AI in spontaneously ovulating sows, under field conditions. In experiment 1, AI was applied to crossbred sows at 12, 24 and 36 h after onset of spontaneous oestrus using one of the following two regimes: (i) DUI (treatment) with 0.15 x 10(9) fresh boar spermatozoa in 5 ml of Beltsville thawing solution (BTS) extender (n = 95), and (ii) standard cervical AI (control) with 2.85 x 10(9) fresh spermatozoa in 95 ml of BTS extender (n = 95). The farrowing rates of the two groups of sows were statistically similar (NS). However, a decrease (p < 0.002) in litter size and the total number of pigs born alive was observed in sows inseminated with the DUI procedure. In experiment 2, 42 post-weaned oestrus sows were inseminated following the same design described for experiment 1 during spontaneous oestrus. On day 6 after onset of oestrus, the proximal segment of the uterine horns of the sows were flushed under surgery to retrieve eventual embryos and evaluate the success of fertilization per cornua (e.g. occurrence of effective uni- vs bilateral sperm transport rendering uni- or bilateral, complete or partial fertilization). Retrieved embryos were assessed for cleavage and number of accessory spermatozoa. Although identical overall pregnancy rates were achieved in both insemination groups, the percentage of sows with partial bilateral fertilization and unilateral fertilization was markedly higher (p < 0.05) in the DUI group (35%) compared with the control (standard AI) group (5%), with a consequent lower (p < 0.001) percentage of viable early embryos after DUI. The number of accessory spermatozoa in the zona pellucida of the embryos was highly variable, but higher (p < 0.001) in control animals than in DUI-AI. No accessory spermatozoa were found in oocytes retrieved from sows depicting unilateral fertilization. In conclusion, DUI in spontaneously ovulating sows with 0.15 x 10(9) spermatozoa renders similar farrowing rates but a lower litter size compared with use of standard AI with a 20-fold higher sperm dose. The lower litter size ought to be related to a decreased distribution of spermatozoa after DUI leading to a higher incidence of partial bilateral and unilateral fertilization.