Ditto to what Jennifer says: count'em because the same number of points
produces different characters per inch in different fonts.  Times Roman
10 is too small and can be spotted easily.  Times Roman 11 may not
produce an average count above 15 cpi, but in a phrase like
"illigitimate lilliputian" it can.  Times Roman 12 is always safe, but
it sure is big.  About five years ago we had one application returned by
NIH and another by NSF becasue the type was too small.  (In each case
the SRA let the PI return corrected copies in a few days for the same
review cycle).  None has been returned since, and NIH regularly accepts
our applications in 11 point Times Roman.  But we give a stern warning
to 10 pointers: type that small is unlikely to win the heart of a
reviewer who submits all her applications in 11 and 12 point type--and
the difference is EASY to see.

Why is it that PIs have such difficulty understanding that professional
who daily handle such tricky notions as titer and joules have such a
hard time understading the concept of "characters per inch"?