Experimental data on the frequency of gamma -ray families of total energy above 30 TeV, from measurements at mountain altitudes, have been summarised. Comparison is made with predictions derived for three alternative multiplicity laws: scaling (i.e. mean number of pion secondaries n_s, varies as ln E₀), n_s varies as E^1/4 and n_s varies as E^1/2. It is found that the E^1/2 dependence gives the best fit in keeping with the analysis of other EAS results. It is pointed out that other 'exotic' nuclear interaction models are also possible.