Abstract: Three hundred and eleven herb drugs (most of which are used in chinese medicine) and 47 composite recipes were screened for antitumor activity in animals. The drugs or recipes were selected from those commonly used in Chinese medicine in the treatment of neoplasms or tumorous diseases or from drugs that belong to the same genus or the same species as certain known antitumor herbs. The preparations employed included decoctions, tinctures, extracts, electrophoretic fractions and such constituents as alkaloids, essential oils, glycosides etc. Ehrlich carcinoma (both the ascitic and the solid from) and sarcoma 180 were routinely employed for the screening; whereas other mouse tumors including spindle-cell sarcoma, sarcoma AK, reticulo-cell sarcoma L₂, sarcoma B₂₂ and spon- taneous carcinoma of the breast, as well as rat tumors, Yoshida ascites sarcoma and Walker carcinoma 256 were also used in some of the tests. The drugs were given once daily and, with a few exceptions, for 6—14 days. Essential oils and soluble crystalline substances were given intraperitoneally or subcutaneously while the rest were administered through stomach tube. Inhibition of tumor growth by 30% or more in the case of solid tumors or prolongation of survival by 3 days or more in the case of ascites tumors were taken as the criteria for antitumor activity. Of the 358 preparations screened 3 (0.8%) showed definite antitumor effects, namely the essential oil of Atractylodes macrocephala Koid (白术), the tincture of Bombyx batryticatus (殭蚕) and Ye Ke San (a composite recipe meaning "powder for dysphagia"), 350 were ineffective and the results of the remaining 5 were inconclusive. Contrary to our experimental data clinical observations on neoplastic patients have claimed encouraging results for many of the composite recipes tested in the present investigation. As the assessment of antitumor therapy requires large scale and prolonged observations, further clinical and experimental investigations are necessary before definite opinions on the effectiveness of these recipes are to be expressed. The question arises as to the cause of the low incidence of positive screening results. To the authors, two points seem to be worthy of consideration. Firstly, in the crude preparations, like decoctions and extracts, effective antitumor contituents, if present, most likely exist in very low concentrations, and sometimes may even be lost or inactivated during processing. Secondly, the nature, actions and potency of the crude preparations are subjected to the influences of a number of variables. These include differences or irregularities in species, district of cultivation, time of harvesting and conditions of preservation and preparation. The non-uniformity of even one of these variables may introduce significant differences in the quality of the final preparation and this in turn would lead to variability of the screening results and difficulty in assessing the antitumor effects. It would be preferable to extract the crystalline constituents of the herb drugs for antitumor screening experiments or as an alterative to make refined preparations from which certain "inactive" constituents like tannins, sugars, gums etc. are removed so as to render concentration of the effective constituents possible.