We examined the effects of the visual size and the number of digits on reading numerical time information in young adults. Using an adaptive staircase procedure, minimal stimulus presentation duration (MSPD) for 80%-correct responses was determined for visual sizes ranging from 0.1° to 15°, when reading 1 (‘‘mm’’), 2 (‘‘hh:mm’’) or 3 (‘‘hh:mm:ss’’) 2-digit units of time information. All three time types revealed U-shaped relations between MSPD and visual size, with the characteristics of the relation depending on the number of time units. Time type had two different effects. First, longer time types gave rise to longer MSPDs, as more elements needed to be encoded into working memory. Second, longer time types gave rise to smaller ranges of optimal visual character size, decreasing from 0.2–2° for the 1-unit time type to 0.3–0.5° for the 3-unit time type. The lower boundary of the optimal range of visual size may be understood as resulting from acuity limitations. The shift in the upper boundary of the optimal range of visual size is suggested to reflect the change in size of the visual span associated with larger visual character sizes.