The purpose of this study is to provide a new high-quality palaeomagnetic pole for mid- Miocene Australia using results from leucitite lavas in Eastern Australia, given their precise 40Ar/39Ar ages. We review published palaeomagnetic data for Australia during the mid- Miocene and examine palaeosecular variation during the 15-17 Ma interval. The average palaeomagnetic direction from leucitite lavas (D = 190.2°, I = 57.1°, k = 33.9, a95 = 6.7°) and the palaeomagnetic pole (λ=-80.2°, Φ = 95.4°, K = 20.5, A95 = 8.6°, N = 15, Q = 5) are statistically indistinguishable from early mid-Miocene poles from Nandewar Volcano and from the Port Campbell and Glenample Formations, demonstrating that older palaeomagnetic results can be robust despite outdated methodology (e.g. blanket demagnetization, few samples per site). We evaluate palaeosecular variation (PSV) during the mid-Miocene utilizing the combined data set (N = 85/93; cut-off), and calculate a mid-Miocene combined pole for Australia (λ = -80.0°, Φ = 103.6°, K = 26.9, A95 = 3.0°, N = 85, Q = 6). The pole yields a palaeolatitude estimate for the Cosgrove Hotspot of λ15 - 17Ma = 40.3°S ± 3.0°, consistent with the present-day inferred location of the Cosgrove Hotspot at 40.5°S. Scatter of virtual geomagnetic poles (VGPs) is estimated at 15.7°(14.1 17.4 ; 95 per cent limits), and is remarkably similar to the Model G and TK03 values from 0 to 5 Ma and 0 to 10 Ma lavas at comparable latitudes.