SAMPLING ERROR FOR LEAF WATER POTENTIAL MEASUREMENTS IN WHEAT

Leaf cutter psychrometers are a rapid method of obtaining samples for leaf water potential measurements. There is, however, concern that the relatively small cut disc (0.24 cm2) may lead to evaporative losses associated with a high cut surface area to sample volume ratio, and thus result in measurement errors. Experiments in growth chambers with wheat (Triticum aestivum L.) plants at different stress levels showed that leaf water potential values obtained with lamina averaging 7 cm2 (with a very low cut surface to volume ratio) compared quite closely with those from 0.24 cm2 cut discs. The relatively good agreement between techniques may be related to the rapid sealing of the disc in the psychrometer chamber, usually within 3 s after cutting. Using leaf cutter psychrometers, flag leaf water potential was measured, and sampling and experimental statistical errors determined on wheat under control vs. stress conditions. This was done to estimate sample and replication numbers needed to attain a given level of statistical precision in water potential measurements. Stress was induced in greenhouse plants by withholding water from pots, and in the field by severing roots by approximately 5 cm below the soil surface. The experiments were randomized in complete blocks with four replications and 10 water potential samples per replication. With water potential means ranging from -0.82 to -5.23 MPa, sampling error was generally large compared with experimental error. Total error (experimental + sampling) was significantly larger (P < 0.05) for stress than for control treatments. This heterogeneity of error associated with water potential measurements was only partly stabilized by a natural log transformation. Using the mean total error from stress treatment only, a combination of replications and/or samples per replication of about 12 per treatment was required to obtained a 95% confidence interval on a treatment mean of .+-. 0.50 MPa.