Nieto (1979) found an excess of optically variable QSOs (OV) near bright galaxies (m < 15.7): 6 observed versus 1.6 expected for r < 5′. The probability involved was p = 5 × 10−3. Because of the small number of OV QSOs in this sample (N = 41, sample 1), this result needed a confirmation. So the same analysis was repeated with a sample of 112 QSOs (sample 3) from Hewitt and Burbidge (1980). Eleven objects were observed at r < 5′ versus 4.4 expected, so p = 4 × 10−3, the sample made up with 71 objects (sample 2) supporting slightly the result found with the first 41 objects. A notable difference between these two samples 1 and 2 is that the objects included in sample 2 are fainter than the objects included in sample 1. Repeating then the same analysis on samples of QSOs at different brightness levels suggests that the excess is related to the apparent brightness of the QSOs.