In 1997 and 1998 two workshops were held to evaluate how consistent observers were at classifying sea lamprey (Petromyzon marinus) marks on Great Lakes lake trout (Salvelinus namaycush) as described in the King classification system. Two trials were held at each workshop, with group discussion between trials. Variation in counting and classifying marks was considerable, such that reporting rates for A1–A3 marks varied two to three-fold among observers of the same lake trout. Observer variation was greater for classification of healing or healed marks than for fresh marks. The workshops highlighted, as causes for inconsistent mark classification, both departures from the accepted protocol for classifying marks by some agencies, and differences in how sliding and multiple marks were interpreted. Group discussions led to greater agreement in classifying marks. We recommend ways to improve the reliability of marking statistics, including the use of a dichotomous key to classify marks. Laboratory data show that healing times of marks on lake trout were much longer at 4°C and 1°C than at 10°C and varied greatly among individuals. Reported A1–A3 and B1–B3 marks observed in late summer and fall collections likely result from a mixture of attacks by two year classes of sea lamprey. It is likely that a substantial but highly uncertain proportion of attacks that occur in late summer and fall lead to marks that are classified as A1–A3 the next spring. We recommend additional research on mark stage duration.