For laying hens, the effects of housing system on bacterial eggshell contamination and eggshell quality is almost exclusively studied in experimental hen houses. The aim of this study was to compare eggshell hygiene and quality under commercial conditions. Six flocks of laying hens in furnished cages and 7 flocks in noncage systems were visited when hens were about 60 wk of age. Farms from Belgium, the Netherlands, and Germany were included in the study. The following parameters were determined on eggs sampled at the egg belts: 1) bacterial eggshell contamination, as expressed by total count of aerobic bacteria and number of Enterobacteriaceae; 2) proportion of dirty eggs; and 3) proportion of cracked eggs and eggs with microcracks. Considerable within-flock differences were found in eggshell contamination with total count of aerobic bacteria, both for furnished cages (P <or = 0.001, range 4.24 to 5.22 log cfu/eggshell) and noncage systems (P <or = 0.001, range 4.35 to 5.51 log cfu/eggshell). On average, lower levels of contamination with total count of aerobic bacteria (4.75 vs. 4.98 log cfu/eggshell; P <or = 0.001) were found on eggshells from furnished cages compared with noncage systems. Concerning Enterobacteriaceae, no significant difference in average eggshell contamination between both systems could be shown. The total percentage of cracked eggs was higher (P <or = 0.01) in furnished cages (7.8%) compared with noncage systems (4.1%). This was, however, due to the high percentage of cracked eggs (24%) observed on one of the furnished cage farms. We conclude that bacteriological eggshell contamination and percentage of cracked eggs differed substantially between individual farms using the same housing system. This may also explain some discrepancies between the findings of the present study versus some findings of previous experimental studies or studies on a small number of farms. Although statistically significant, the average differences in bacteriological contamination of nest eggs between both housing systems have limited microbiological relevancy.