In the second of two graduate classes with the same title and content,
the volume of on-line messages per person, in the four weeks sampled,
was more than seven times higher than in the first. Between weeks 1-2
and weeks 5-6, the total number of messages per person dropped in the
first class and rose in the second. Each class had a different comput
er-mediated communication (CMC) system and was given different guideli
nes for use. This paper analyses differences in volume and hypothesise
s a relationship between system characteristics, guidelines for use, a
nd the perpetuation of verbal interaction in settings where CMC is man
dated.