area came from a single plant (sample #46) in which 85% of the late prophase- metaphase I ceils had seven bivalents and one quadrivalent (Fig. 3) indicating that it is heterozygous for a reciprocal translocation. Many cells from both areas had eight bivalents and two univalents (Fig. 4), and a few cells from both areas had nine bivalents and a ring chromosome or a univalent (Fig. 5). Because the aberrant cells noted in sample #46 are so much more frequent than in the rest of the population we have recalculated the values in Table 1 both including and omitting the data from that shrub. When this data is omitted the percentage of aberrant cells drops from 14.2% to 6.6%, compared with 3.8% for non-irradiated shrubs. ‘paliradene meted! Shewattarene! Fig. 4. Fig. 3. Seven bivalents and ring quadrivalent (note arrow) from Artemisia sptnescens sample #46 which is heterozygous for a translocation. 358 The frequent condition of eight bivalents and two univalents (note arrows) in Artemista spinescens. A Mann-Whitney statistical test (Conover, 1971) for differences between the irradiated and non-irradiated populations utilizing all data does not indicate a difference at the usually accepted levels of statistical significance. (Test statistic = 51; w.95 = 57 for a two tailed test. Null hypothesis that the average proportion is same in both populations cannot, therefore be rejected.) If however, the highest value in each population is omitted from the calculations, the difference is significant at the 95% level. (Test statistic = 43; w.95 for a two tailed test is 42. The null hypothesis that there is no difference between the average proportions in both populations can, therefore, be rejected.) 359