This is the first in a series of experiments to study the fatigue properties of equine cannon (third metacarpal) bone specimens from Throughbred racehorses. Monotonic and fatigue tests to failure were performed in four-point bending on diaphyseal specimens in a 37°C saline bath to answer three initial questions. (a) Will a linear variable differential transducer yield the same elastic modulus as strain gauges? (b) Will fatigue results depend on whether the periosteal or endosteal side of the beam is in tension? (c) Are there regional variations in the monotonic and fatigue properties of the cannon bone midshaft? Eighteen left-right pairs of specimens from six horses were used. One beam of each pair was fitted with strain gauges. Fatigue tests were conducted on 24 specimens under load control at 2 Hz; an initial range of 0-10,000 microstrain was used so as to produce failure in a reasonable period of time. There were no left-right differences in the fatigue or monotonic properties, and the presence of a gauge had no effect on modulus measured by a linear variable differential transducer. However, gauge-measured moduli were about 1 GPa less than transducer-measured values. Fatigue life was independent of which side of the beam was in tension, and there were significant variations in mechanical properties around the cortex. The lateral region was stiffer than the dorsal region but the latter had a longer fatigue life. The fixed cylindrical supports used in this experimental eventually produced slight wear grooves, causing artifactual stiffening at the end of the load cycle in some specimens. A second experiment using roller supports confirmed the reason for this stiffening. It also showed that fatigue life was shorter when roller supports were used but regional differences were similar.