We investigated the accuracy of using symphysis pubis-fundal height measurement and ultrasonically derived estimation of fetal weight for identifying small-for-gestational-age (SGA) and large-for-gestational-age (LGA) fetuses. A retrospective cohort study was performed using computerised records of all non-diabetic women referred for an ultrasound growth scan (US) with clinically suspected SGA or LGA singleton fetuses from 35 weeks' gestation between October 2008 and March 2009 (n = 185). Birth weight data were also collected for all births over the study period (n = 3200). One-third (34%) of ultrasound estimated fetal weights were inaccurate by >10%. However, an inaccurate ultrasound estimation did not significantly increase the likelihood of spontaneous or induced onset of labour or delivery by caesarean section. Most (79%) growth scans were performed on well-grown fetuses. The majority (80%) of SGA and LGA babies in our cohort were not identified by clinicians during routine antenatal care. From 3,200 live births, there were a total of 59 infants weighing <2,500 g or >5,000 g. Of these, only 12 had been referred for an ultrasound growth scan, indicating that abdominal palpation and fundal height measurement has a 20% sensitivity in detecting SGA or LGA fetuses. Of the 12, four were detected using ultrasound, indicating a 33% detection rate. Although ultrasound has a slightly higher sensitivity, neither clinical examination using fundal height measurements nor 3rd trimester ultrasound examinations are effective at detecting SGA or LGA fetuses.