Superconducting quantum interference device (SQUID) magnetometers are very high precision instruments: for example, the Quantum Design MPMS-7 instrument capable of measuring an absolute magnetization of ≈10-7 to 10-11 emu (10-10 to 10-14 Am2), corresponding to better than ≈1 ng of magnetite, Fe3O4. However, in biological samples, such precision is rarely achieved. In the presence of ≈100 mg of biological tissue there is a diamagnetic contribution of ≈-9×10-9 emu/Oe so that at 10 kOe the measurement of 1 ng of magnetite in 100 mg of tissue has its precision reduced by a factor of 10, with a loss in accuracy of a factor of 2.5. The extra volume of the biological material also reduces accuracy, typically by ≈25%. We describe here a measurement protocol that increases the obtainable precision and improves accuracy by a factor of 5, and which limits the sample volume effects to ≈2%–3%. This then allows accurate measurement of magnetic biominerals in a biological/diamagnetic matrix. Details on how to prepare, mount, and accurately measure dilute magnetic samples are given. The improvement in data quality comes at the cost of extended measurement periods and slightly increased helium consumption.