We evaluated the results of cryopreserved/thawed embryo replacement (FER) to determine if the outcome following transfer in a natural cycle in a defined group was different to that from a hormone replacement cycle, and also to assess vaginal ultrasonographic features that assist in predicting the timing of the transfer. At the London Fertility Centre, 149 consecutive FER cycles were studied retrospectively. Women with proven ovulation and regular cycles were included during natural cycles (n = 77). The hormone replacement cycle group included women with anovulation, irregular cycles and older women (n = 72). In the natural cycle group, transfer was performed following positive urinary luteinizing hormone (LH) surge and confirmation of ovulation by ultrasonography. With the hormone replacement therapy group, gonadotrophin-releasing hormone analogue was used to induce pituitary down-regulation, oestradiol valerate was supplemented followed by regular ultrasound monitoring, and FER 2 days following the initiation of progesterone, which was started once adequate endometrial development was noticed on ultrasonography. The pregnancy and ongoing/delivery rates were analysed in relation to the treatment cycle, age, number and quality of embryos transferred. Ultrasonographic features were examined to evaluate their relationship with the outcome of treatment. The results showed that no difference existed between natural and hormone replacement cycles in pregnancy rates per cycle (26 and 25%), ongoing/delivery rate (20.8% in both groups), and implantation rate (10.3 and 10.6%). Pregnancy rates were not influenced by the number of embryos transferred, stage at which the embryos were cryopreserved, or whether they were extra embryos from in-vitro fertilization/embryo transfer, or gamete intra-Fallopian transfer. The pregnancy rate was low (7.4%) if the embryos had less than three blastomeres and if the fragmentation was >50% (0% pregnancy rate). With hormone replacement cycles, age did not influence the outcome, and women 40 years and older had a pregnancy rate of 29.4% per cycle. No pregnancies resulted in the natural cycle group if the maximum follicular diameter was > 22 mm before ovulation. When poor endometrial development was noted (thickness < 8 mm and grade C) no pregnancy resulted from FER in natural or hormone replacement cycles. The pregnancy rates were higher when the endometrium was ≥8 mm thickness and grade B (42.4%) or grade A (21.2%). We concluded that FER outcomes in natural cycles were similar to those arising with hormone replacement therapy provided good selection criteria were used, and vaginal ultrasonography can assist in timing the day of replacement and identify cases to be cancelled before the transfer.