Abstract Compounds that produce myocardial pathology in vivo can be separated into two main classes—those that are directly toxic to the myocardium and those that are considered to act by way of an indirect vascular or neurologically based mechanism. An in vitro model of myocardium without nervous or systemic influences can be used to differentiate between direct myocardial cytotoxic effects and indirect cardiac pathology arising in vivo from exaggerated vascular or neural pharmacological effects of a number of drugs. In this study direct-acting cardiotoxic compounds are distinguished from those causing cardiac pathology by indirect mechanisms by their different pattern of effects in chick embryonic myocardial myocyte reaggregates (MMRs) cultures. The toxicity of the direct-acting cardiotoxic drugs allylamine (positive control, 50 μm) and doxorubicin were compared with digoxin and isoprenaline, which show both direct and indirect mechanisms in vivo, and the indirectly acting hydralazine and pinacidil. Changes in spontaneous beating activity (SBA), leakage of lactate dehydrogenase (LDH) and cell morphology by light and transmission electron microscopy were used to assess toxicity. The MMRs were cultured for up to 24 hr in a series of concentrations of the five compounds in the range 0.1 to 10,000 μm. Allylamine, doxorubicin, digoxin and, to a lesser extent, isoprenaline were highly toxic to the MMRs, as shown by alterations in SBA, LDH leakage and cellular morphology. In contrast, hydralazine showed a very mild degree of toxicity at the highest concentrations in the absence of LDH leakage; treatment with pinacidil did not show any evidence of morphological degeneration but did cause a dose-related inhibition of SBA. These results are consistent with the view that doxorubicin and digoxin are directly toxic to myocardial cells and also suggests that this is an important mechanism in vivo for isoprenaline. The absence of a significant degree of toxicity with hydralazine and pinacidil is consistent with an indirect toxic mechanism.