Because of strong indications of multiple counting by the multi-channel scaler (MCS) during most of the experiments described in Dendooven et al (2015 Phys. Med. Biol. 60 8923–47), the production of short-lived positron emitters in the stopping of 55 MeV protons in water, carbon, phosphorus and calcium was remeasured. The new results are reported here. With proper single counting of the MCS, the new production rates are 1.1 to 2.9 times smaller than reported in Dendooven et al (2015 Phys. Med. Biol. 60 8923–47). The omission of the conversion from MCS time bin to time unit in the previous data analysis was corrected, leading to an increase of the production rate by a factor of 2.5 or 10 for some nuclides. The most copiously produced short-lived nuclides and their production rates relative to the relevant long-lived nuclides are: 12N (T 1/2 = 11 ms) on carbon (5.3% of 11C), 29P (T 1/2 = 4.1 s) on phosphorus (23% of 30P) and 38mK (T 1/2 = 0.92 s) on calcium (173% of 38gK). The number of decays integrated from the start of an irradiation as a function of time during the irradiation of PMMA and 4 tissue materials has been determined. For (carbon-rich) adipose tissue, 12N dominates up to 70 s. On bone tissue, 38mK dominates the beam-on PET counts from 0.2–0.7 s until about 80–110 s. Considering nuclides created on phosphorus and calcium, the short-lived ones provide 8 times more decays than the long-lived ones during a 70 s irradiation. Bone tissue will thus be much better visible in beam-on PET compared to PET imaging after an irradiation. From the estimated number of 12N PET counts, we conclude that, for any tissue, except carbon-poor ones, 12N PET imaging potentially provides equal quality proton range information as prompt gamma imaging with an optimized knife-edge slit camera.