Considerable research has been conducted with natural fibres as a source of reinforcement for organic matrices. These fibres have included wood pulp, sisal, hemp and numerous other plants [1–3]. However, little attention has been given to natural animal fibres such as wool and hair. Traditionally, wool fibres have been spun into multilayer fibres in the form of threads, which are then knitted into cloth and utilized for the manufacture of garments. As a composite, wool fibres have been combined with polyester fibres and spun into multistrand yarn as threads, again for use in garments. The major engineering application of these composites has been for thermal insulation for the human body. In general, the mechanical properties were not considered relevant for cloth applications. This study seeks to investigate the mechanical properties of a new form of wool–polyester composite, i.e. the reinforcement of raw wool fibres (before spinning into thread or yarn) with a polyester resin matrix for use in structural applications not associated with clothes. Raw wool is one of the major export commodities of a number of countries, including Australia. It is one of the oldest naturally renewable fibres in use. It is epidermal in origin and grown naturally on the bodies of sheep. The wool is obtained by shearing the fleece from the sheep, and then graded, washed and combed to remove any extraneous matter such as grease, dirt, moisture and vegetable matter. The fibres are generally entangled and crimped. In this study, the wool fibres were supplied by the CSIRO Division of Wool Technology. The fibre was classified as fine, clean carded wool, i.e. washed and combed with all foreign matter removed. The composite matrix was prepared from polyester resin with 1% hardener (methyl ethyl ketone peroxide). Samples of composite sheets were prepared in the laboratory from skeins of wool laid alternatively with layers of resin mixture, and placed in a rectangular mould. This was repeated until the required fractions of wool and resin were achieved. The top of the mould was sealed and hydraulic pressure of 1.2 MPa was applied for a period of 24 h. The pressure was then reduced to 0.6 MPa, the sample sheet was removed and excess solid resin trimmed. The composite so obtained was cured in air. This process was repeated for sample sheets containing wool by mass of 40, 30, 20 and 10 g and a ‘‘control’’ sheet of polyester resin. All test pieces had the wool aligned in the longitudinal direction except for the 40 g (55% mass fraction) transverse samples. Wool composite samples were prepared for a range of fibre fraction by wool mass content corresponding to mass fractions of 21, 40, 52 and 55%. Specimens were cut from the prepared sample sheets for use in three test procedures: flexural (36 test pieces), tensile (36 test pieces) and Izod impact with 48 test pieces. All specimens were conditioned for 24 h at 22 2 8C at a relative humidity of 52 5%. An Instron hydraulic testing machine (model 1114) was used for both the flexural and tensile testing. Automatic data collection was employed to allow further processing of the test data. Flexural test pieces were cut into rectangular shapes measuring approximately 150 mm 3 15 mm (of various thickness, approximately 26 mm) using a rotary saw. Three-point bending tests with a span of 50 mm were conducted. The flexural strength (modulus of rupture) was calculated as 3Pl=2bd2 where P is maximum load, b is test piece width, d is test piece thickness and l is test piece span length. Impact test specimens were cut into rectangular shapes using a rotary saw and notched with a milling machine cutter. Tests and sample dimensions followed the conditions outlined in AS1146 [4]. Impact testing was conducted using a Material Forge model TM 52004 impact tester (an Izod type test). The results from the mechanical tests are given in Table I. These results indicated that there is little influence on the tensile stress or modulus of elasticity with increasing fraction of wool content. However, when the fibres were laid in a transverse position to the tensile load the tensile stress was approximately one half of that for the parallel loading case with the 55% wool fibre test pieces. Moreover, the modulus of elasticity of the transverse samples was similar to that of the parallel fibre samples. The variation of flexural strength with increasing fibre content is shown in Fig. 1. Flexural strength results are also shown for a control sample with 0% fibres, and transverse aligned fibres with 55 wt % wool fibre. The maximum value of flexural strength was at the