The hypothesis that interstroke intervals and the mean number of strokes per lightning flash to ground vary systematically with latitude is tested using statistical methods. The weighted regression lines relating ⟨log t⟩ and ⟨n⟩ to the latitude, λ in degrees), are ⟨log t⟩=1.75+5.3¿10-4λ and ⟨n⟩=4.4-2.5¿10-2λ, where t is the interstroke interval and n the number of strokes per flash. However, in each case, correlation is not significant at the 5% level of confidence, so that the data generally do not show a systematic latitude dependence. Latitude does appear to affect the distributions of log t in that the tropical distributions of log t are fairly consistent (within the 5% confidence limits using analysis of variance), whereas fluctuations in the nontropical distributions of log t cannot be explained in terms of random variations, but this may be an effect of differing measurement techniques. Year-to-year fluctuations in the observed distributions of the number of strokes per flash are shown to be statistically significant for both a tropical and a nontropical locality and may account for the considerable differences between distributions of n at different localities.<