Medicago truncatula (M. truncatula) has been proposed as a model legume for molecular and genetic studies of legumes. While many genetic resources have been developed for this model legume, genetic transformation of the M. truncatula line A17 proved to be a problem. A reproducible transformation method is described for M. truncatula A17. Procedures are detailed that yielded an average regeneration frequency of 35% for recovery of transgenic shoots from cotyledonary node explants. Previously, rooting of transgenic shoots of this line has proven difficult, but media and culture procedures are described that yielded an average frequency of 39% for root induction from 419 phosphinothricin-resistant shoots. Fertile M. truncatula A17 plants transgenic for 35S-GFP, phas-GUS or phas-GFP were obtained. The presence of transgenes was confirmed by expression of transgenes and by genomic DNA blots. Interestingly, although GUS and GFP driven by the phas promoter were very strongly and uniformly expressed in seed cotyledons of most transgenic M. truncatula lines, silencing of the GUS expression from the phas promoter was observed in several lines, indicating the occurrence of novel epigenetic events. The diminution of GFP fluorescence in transgenic M. truncatula occurs despite the presence of GFP transcript and protein. To evaluate the generality and causes of this phenomenon, fluorescence during leaf development from the same 35S-GFP transgene was compared in M. truncatula, rice and Arabidopsis. A substantial decrease in fluorescence early in the development of M. truncatula and rice leaves was found to correlate with chlorophyll accumulation. Several approaches showed that chlorophyll is causally involved in the loss of GFP fluorescence. Removal of chlorophyll from leaves of transgenic M. truncatula, rice or Arabidopsis through etiolation or by extraction with ethanol yielded up to a tenfold increase in fluorescence. Direct evidence that chlorophyll is implicated in the loss of fluorescence from GFP was obtained by mixing solutions of chlorophyll and GFP. At low concentration, fluorescence loss was fourfold greater for chlorophyll b than for chlorophyll a, reflecting their relative interference with GFP excitation and emission. Thus, substantial errors in estimating promoter activity from GFP fluorescence can occur if pigment interference is not considered.