Calibration of echo sounders for fish stock assessment are commonly performed using the standard sphere method (Johannesson and Mitson, 1983; Simmonds et al., 1984). To determine the accuracy of the method, direct measurements of target strength (TS) were made of three standard spheres (Copper (Cu) - 23.0 mm and tungsten carbide (WC) - 33.0 mm and 38.1 mm). At the best case range of 5 m, the TS measurements of the spheres differed from the theoretical values (derived by integrated intensities), by -0.1, 0.3, and 0.1 dB with standard deviations of 0.0, 0.3, and 0.2 dB, respectively. The operative measure (derived by peak intensities), differed from the theoretical values by -0.2, O.4, and 0.2 dB with standard deviations of 0.1, 0.3, and 0.2 dB, respectively. To characterize the precision of the method for a fixed pulse length (0.3 ms) and water temperature (18.9°C), a Simrad EK500 echosounder was used to measure sphere TS versus time. Over two 15 hour periods, the measured TS ranged 1.2 dB for a 23.0 mm Cu sphere and 1.4 dB for a 38.1 mm WC sphere. Vector admittance measurements were made of an ES120 transducer versus water temperature (0.06-16.8 °C). Although the measurements were not free-field and consequentially noisy, the trends versus increasing water temperature, indicated decreasing admittance at the operating frequency (119.047 kHz), decreasing resonance frequency, and increasing motional resistance. Judging from these experiments, system calibration at 120 kHz, at a fixed water temperature, using an optimal standard sphere, and a 0.3 ms pulse length, is estimated to be accurate to to.3 dB, and precise to ±0.2 dB for measuring TS; while accurate to ±0.2 dB, and precise to ±0.2 dB for echo integration. Additionally, more pronounced imprecision may be contributed by instabilities in the echosounder electronics. Furthermore, when operating under conditions of varying water temperature, associated changes in transducer performance may cause significant increases in calibration uncertainty. The temperature effects on system gain are consistent with predictions (Blue, 1984), and prior experimental results (Demer and Hewitt, 1993).