Background: Cyclopentolate continues to be the cycloplegic of choice for refracting young children, although many studies of ocular biometry promote the use of tropicamide. Methods: To clarify the role of drug type in biometric measurements, cycloplegia was induced in two disparate age groups using cyclopentolate and tropicamide on two separate occasions. Refraction, phakometry and A-scan ultrasonography measurements were made on two groups of Tibetan children resident in Nepal. Results: Cyclopentolate produced significantly more cycloplegia in the younger group, which was supported by phakometry measurements. However, in clinical terms, the difference between the measurements was not significant. Conclusion: We conclude that although cyclopentolate is more effective than tropicamide in relaxing accommodation in young children, the use of a local anaesthetic prior to instillation of tropicamide produces refractive data virtually equivalent to that of cyclopentolate, regardless of the age group measured. However, biometric measurements may be susceptible to greater error when near fixation targets are used during phakometry procedures. © 1998 Optometrists Association Australia.