US—City Directories, 1866–1940

I just pulled two batches. Each had one image with two pages and four columns. On the first page of each image the first column was not readable. Some of the second column could be deciphered. The second page of the image was quite legible and could be easily indexed. I sent the first batch back but still have the second. is there a way to do this without marking a large number of unreadable records? Or if I return them how can I indicate that much of the image is not readable?