1. Serial cytological observations were made on one abnormal seedling from the thornless segregants of the F2 progeny from crossing two tetraploid (2n = 28) Rubus species. The number of root-tip chromosomes was found to vary between 9 and 46 per cell, with the mode always at 35. The wide variation in chromosome number decreased with time, until there was almost stability at 2n = 35. A vegetative (stipule) cell was 2n = 39. The numbers lower than 35 were not due to terminal fusion.
2. Evidence from the number of satellites per cell suggests that not always were the same chromosomes being eliminated, although there is a general increase in satellite number with increased numbers of chromosomes per cell.
3. There was a correlation between the chromosome numbers and corresponding cell sizes within the aneuploid series. This may be associated with the incremental change of DNA following the addition or subtraction of individual chromosomes, which are highly similar in Rubus.
4. The cause of the instability was not definitely established. Its initiation is probably attributable to the egg mother cell, as Rubus pollen reacts sensitively chromosome unbalance. No abnormalities were observed in the mitoses, and virus infection was not responsible for the instability.
5. Chromosome number instability of this plant is discussed in relation to other examples from the same and different genera: it is suggested that the changes in chromosome number known in various Rubi may have arisen somatically.