Several authors, including Fischer, Wehmeier and their collaborators (1933, 1935) and Waddington, Needham, Nowiński and Lemberg (1935), have obtained the induction of neural tissue in amphibian embryos by the implantation of crude glycogen or of substances extracted from it. Others (Holtfreter 1933, 1934; Woerdemann 1933) have employed it unsuccessfully. It is clear now that the activity is not due to the glycogen itself, but to some accompanying impurity, and it was suggested by Waddington, Needham and Brachet (1936) that the active evocator substance is of a sterol-like nature and forms a loose complex with glycogen and possibly with some protein as well. It therefore seemed interesting to investigate whether any difference could be detected in the evocating powers of glycogen in its two forms—the desmo-form in which it is combined with protein and the lyo-form in which it is free (Willstätter and Rohdewald 1934). Specimens of lyo- and desmo-glycogen were therefore prepared and implanted. Both showed themselves capable of evocation, and their activity was such that they seemed suited for testing the effect of a chemically homogeneous evocating mass on ectoderm isolated from the host. By implanting desmo-glycogen into isolated pieces of ectoderm, we should obtain some idea of what a single chemical stimulus is capable of evoking; does it simply induce neural tissue or is there any tendency for the isolated ectoderm to react by producing a neural organ? This question was raised at the time of the discovery of the activity of the dead organizer (Waddington 1933) and has often been mentioned subsequently (in particular Waddington 1934; Needham 1936
b
). The present results seem to answer it in favour of the first alternative. 2—Methods Preparations of desmo- and lyo-glycogen from rabbit liver were made as follows: 145 g. of liver, which had been cooled in ice as soon as it had been removed from the animal, was minced up with scissors and added in portions to 150 ml. of boiling water. When all the tissue had been added the flask was warmed for a further 20 min. on a boiling water-bath, with occasional shaking. The mixture was then filtered on a Buchner funnel (Filtrate I). The residual tissue was ground up without sand in a mortar, heated on the water-bath for 20 min. with a further 200 ml. of water, and filtered as before (Filtrate II). It was assumed that most of the lyo- glycogen would have been extracted in the first two operations, and most of the remainder was then removed from the tissue by extracting with boiling water five more times. 200 ml. of water were used for each of these extractions, which were allowed to proceed on the boiling water-bath for 1-2 hr. The extract was removed by centrifugation.