Applying classical statistical methods to data of 14 low- and mid-latitude locations the reduced standard deviation r.s.d. of monthly foF2 is shown to depend on hour, season and solar cycle phase. With daily percentage departures of foF2 from monthly mean exceeding r.s.d. a comparable analysis applied to individual days shows magnetically disturbed days clearly outside. As experience shows this holds no more when reduction with respect to the true monthly mean is replaced by one to a predicted median e.g. the URSI numerical map.