A figure like this is likely representing the best frequency for hair cells. In actuality, cells respond to a broad range of frequencies. Perceptual specificity comes thanks to lateral inhibition. Tuning curves look something like this:
image from: http://www2.tulane.edu/~h0Ward/BrLg/AuditoryTransduction.html
The y-axis here is in units of threshold, so a small number means the cell is detecting quieter sounds at that frequency. For example, cell D here has a best frequency around 300 Hz, but was still somewhat sensitive to the lowest frequencies presented (looks like about 180 Hz). You won't see many experimental tuning curves that go below that because a typical speaker doesn't go down that low.
If someone says humans can detect sounds down to 20 Hz, that doesn't necessarily mean any inner hair cell has a best frequency at 20 Hz; if it did, it would probably be somewhat sensitive to some even lower frequencies, too, and you'd have to report a different number. If humans can hear 20 Hz sounds, that just means that for some cell, 20 Hz is on the far left end of its tuning curve, such that a loud enough sound at 20 Hz can be perceived.
You can assume the cells most sensitive to the lowest frequencies are at the apex.