The shellfish monitoring system in New Zealand relies on the DSP mouse bioassay (using acetone/dichloromethane extraction) to detect the lipid-soluble toxins NSP, DSP, pectenotoxin and yessotoxin. Where bioactivity is detected, further tests are undertaken to identify the toxin responsible. Areas are closed for harvest on the basis of results from these additional tests. The method is becoming increasingly cumbersome as the number of individual toxin tests that are required increases. One of the bioactive molecules most commonly detected in the screen test, gymnodimine, is not orally toxic. If a screen test could be found which did not detect gymnodimine, but did reliably detect all other lipid-soluble toxins, such a screen test could be used to determine whether areas should be available for harvesting, eliminating the need for subsidiary testing. This project aimed to find a screen test method for the lipid soluble toxins which would give negative results with gymnodimine, but positive results with the toxins mentioned above. Several variants of the present acetone extraction method were tried. The most promising was used for limited trials on samples of known toxicity. The method worked well in detecting the genuine toxins but not gymnodimine. However, the number of nontoxic samples which gave positive results in the mouse bioassay increased using this method. The method appeared to give accurate results at the 10 mouse unit (MU) / 100 g level, but gave too many false positives at 5 MU/100 g shellfish. Subsidiary testing for DSP toxins and pectenotoxin at concentrations between 5 and 10 MU/100 g would still be required if this method were to be introduced. It was therefore decided that the benefits of introducing this method were insufficient to outweigh the cost of method validation.