The theory of formation of amboceptors is similar to that for the formation of the other types of antibodies. The cell introduced contains some substance, which acts as a chemical stimulus to some of the body cells provided with proper receptors so that more of these special receptors are produced, and eventually in excess so that they become free in the blood and constitute the free amboceptors. It will be noticed that these free receptors differ from either of the two kinds already described in that they have _two uniting groups_, one for the antigen (cell introduced) named _cytophil-haptophore_, the other for the complement, _complementophil haptophore_. Hence amboceptors are spoken of as _receptors of the third order_. They have no other function than that of this double combining power. The action which results is due to the third body--the complement. It will be readily seen that complement must possess at least two groups, a combining or _haptophore group_ which unites with the amboceptor, and an active group which is usually called the _zymophore_ or _toxophore_ group. Complements thus resemble either toxins, where the specific cell (antigen) is injured or killed, or enzymes, in case the cell is likewise dissolved. This action again shows the close relation between toxins and enzymes. Complement may lose its active group in the same way that toxin does and becomes then _complementoid_. Complement is readily destroyed in blood or serum by heating it to 55° to 56° for half an hour, and is also destroyed spontaneously when serum stands for a day or two, less rapidly at low temperature than at room temperature. Amboceptors appear to be _specific_ in the same sense that agglutinins are. That is, if a given cell is used to immunize an animal, the animal’s blood will contain amboceptors for the cell used and also for others closely related to it. Immunization with spermatozoa or with epithelial or liver cells gives rise to amboceptors for these cells and also for red blood corpuscles and other body cells. A typhoid bactericidal serum has also some dissolving effect on colon bacilli, etc. Hence a given serum may contain a chief amboceptor and a variety of “coamboceptors,” or one amboceptor made up of a number of “partial amboceptors” each specific for its own antigen (“amboceptorogen”). Amboceptors may combine with other substances than antigen and complement, as is shown by their union with lecithin and other “lipoids,” though these substances seem capable of acting as complement in causing solution of blood corpuscles. COMPLEMENTS. As to whether complements are numerous, as Ehrlich claims, or there is only one complement, according to Buchner and others, need not be discussed here. In the practical applications given later as means of diagnosis it is apparent that all the complement or complements are capable of uniting with at least two kinds of amboceptors.