EFFECT OF SAMPLING SCALE ON THE ASSESSMENT OF EPIPHYTIC BACTERIAL-POPULATIONS

Bacterial populations on above-ground plant surfaces were estimated at three different biological scales, including leaflet disks, entire le aflets, and whole plants. The influence of sample scale on the estimat ion of mean bacterial population size per unit and per gram and on the variability among sampling units was quantified at each scale, Popula tions were highly variable among sampling units at every scale examine d, suggesting that there is no optimal scale at which sample variance is reduced. The distribution of population sizes among sample units wa s sometimes, but not consistently, described by the lognormal. Regardl ess of the sampling scale, expression of population sizes on a per gra m basis may not reduce variance, because population size was not gener ally a function of sample unit weight within any single sampling scale . In addition, the data show that scaling populations on a per gram ba sis does not provide a useful means of comparing population estimates from samples taken at different scales. The implications of these resu lts for designing sampling strategies to address specific issues in mi crobial ecology are discussed.