Three methods, the enumeration of large non-cleaved cells (LNCC), the enumeration of nucleolar organizer regions (AgNORs), and the estimation of the percentage of Ki67-positive cells, have been compared on a series of 36 follicular lymphomas for their utility in subclassification. All three methods produced similar subdivisions to those obtained using the histological classification of the Working Formulation for clinical usage. Twenty cases have been followed clinically to assess any relationship between these parameters and short-term (less than 4 years) survival. Each technique showed a trend towards identifying patients with a poor short-term prognosis with no one method being superior to any other. Unless larger series with longer follow-up suggest otherwise, there is no indication to use either the AgNOR or Ki67 methods which offer no advantage over the simple and inexpensive Working Formulation or Berard's LNCC counting.