Two cell lines of birch (Betula pendula Roth.), one potentially embryogenic given the right inductive conditions and one which never has shown any embryogenic capacity, were both subjected to conditions inductive and non-inductive for somatic embryogenesis. Cells from these treatments were harvested at intervals over a 3-week period and washed with salt solution to wash off proteins loosely attached to the cell walls. The remaining cells were either freeze-dried whole or the cell walls were isolated. The extracted proteins from these three cell preparations were separated by one-dimentional SDS-polyacrylamide gel electrophoresis and detected by silver staining. Proteins specific for embryogenic cultures under inductive conditions were found in samples from the whole washed cells, whereas in the samples from isolated cell walls and “cell washings”, certain proteins seemed to disappear when the cells entered the embryogenic state. The changes in protein patterns were evident 24 h after the medium has been changed to embryo-production medium.