By means of heterogenous associations of two parts whose morphogenetic potentialities are different, we tried to find our if the formation of a supernumerary morphogenetic centre comes from one of these parts or from both. 2. Associations of opposite sides of legs of different thoracic metameres give such chimerae that the graft and the receptor side seem to be able to form at least the femoral side which prolong them. 3. The grafts of a ventral side of scape into a dorsal or a posterior coxal side and of a dorsal side of scape into a ventral coxal side determine the formation of a chimera which presents mostly femur tibia and tarsus on the side of the coxal receptor side and tarsus on the side of the grafted scape. 4. The heterogenous associations of a coxa and of grafts coming from more distal segments of the leg (femur, tibia, tarsus) give two types of results. In some cases, the structure of the sides which respectively prolong the graft and the receptor coxa conform to the potentialities shown by each of these parts in a regeneration. In the other cases and particularly with tarsal grafts, the neoformations have, on the side of the graft, articles which, on a normal leg, are localized proximally to the graft. 5. The heterogenous associations in which a level does not normally regenerate determine often the formation of a supernumerary morphogenetic centre which, in about 1/3 of the cases, gives a segmented formation presenting all the articles localized distally to the receptor level. 6. From these results one can deduce that the graft and the receptor may contribute to form a segmented appendage. It seems also possible that some formations come only from the receptor tissues.
