There are certain disease-causing de novo mutations in the germline with rates per generation orders of magnitude higher than the genome average. Moreover, these mutations have additional characteristics: 1) they result from point mutations and lead to a gain-of function change, 2) they occur exclusively in the male germ line, and 3) older men have a higher probability of having an affected child than younger ones, known as the paternal age-effect (PAE). One of the best known PAE mutations is achondroplasia, caused by a single nucleotide substitution in FGFR3. In order to elucidate what mechanisms might be driving the expansion of this mutation in the male germline, we examined the spatial distribution of the achondroplasia substitution (c. 1138G>A) in a testis of an 80-year old man. Using a technology based on bead-emulsion amplification, we were able to measure mutation frequencies as low as 3x10-6 when assessing the mutation frequency in 192 individual pieces of a dissected testis. We observed that the mutations are clustered in a few pieces with 95% of the mutations occurring in 25% of the total testis. Using computational simulations, we determined that the observed distribution fits a selection model well, where occasionally a spermatogonial stem cell divides symmetrically instead of asymmetrically. A model proposing a mutation hotspot to explain the elevated mutation rate was rejected. Our observations parallel the ones for other PAE mutations such as Apert syndrome and MEN2B and provides further evidence that the PAE might be caused by the selective advantage conferred to mutant cells by means of altering tyrosine kinase signaling pathways.