The normal lymphocyte transfer reaction (NLT reaction) is a cutaneous inflammatory episode of delayed onset that is aroused when living lymphocytes from one guinea-pig are injected into the skin of another, and it occurs only in those genetic situations that show it to be an immunological response of the transferred lymphocytes to antigens of the animals into which they are injected (section 4). When the recipient guinea-pig is exposed to 600 r whole-body irradiation before transfer, to delay the onset of an immunological counter-attack on the transferred cells, the NLT reaction evolves in three phases spread over about 6 days (section 3). These are: (a) the first inflammatory episode, of moderate intensity, which reaches its peak at 24 h and remains stationary for 2 further days; (b) the flare-up, starting between the third and fourth days and rising to a peak of violent intensity at about the sixth day; and (c) the fade-out, which is mainly due to an immunological recovery of the host. The same three components may be discerned when the transferred lymphocytes have been presensitized against the tissues of their future recipients, but the pitch of the reaction is much higher throughout and presensitized cells perform as strongly at 24 h as normal cells do when they reach the peak of their activity at 5 to 6 days (figure 2). In general, lymphocyte transfer reactions vary in intensity rather than in tempo: the weak reactions that occur where antigenic disparity is slight, or when relatively few cells are injected, differ from strong reactions-even the very strong reactions caused by presensitized cells-in their general pitch of intensity but not in the relative timing of the various episodes of the response. In terms of the power of a given number of cells to excite an NLT reaction, blood lymphocytes were two to five times more active, and thymocytes ten to twenty times less active, than lymph node cells (section 5). Cells from lymph nodes caused to enlarge greatly by stimulation with human gamma globulin emulsified in Freund's complete adjuvant were not more effective than normal cells in exciting an NLT reaction. Lymphoid cells from foetal or newborn mesenteric nodes gave bold and clear NLT reactions rising to a peak not lower than that achieved by the same number of adult lymphoid cells (section 5). The NLT reaction lent itself very well to a study of inhibitors of the immunological response. Immunosuppressive agents (section 6) were applied, as appropriate, to the cell donor before transfer, to the lymphoid cells in transit, or to the recipient before, during or after transfer. Immunosuppressive agents did not in general affect the first inflammatory episode; with varying degrees of effectiveness they did, however, eliminate the flare-up. The two most effective agents in this respect (Methotrexate and cyclophosphamide) did not, however, oppose the immunological performance of presensitized cells at concentrations more than sufficient to eliminate the flare-up. It is therefore reasoned that conventional immuno-suppressive agents do not affect any distinctively immunological activity of lymphoid cells: they merely prevent the multiplication of the cells activated in the first episode of the NLT reaction, and therefore the transformation of a 'normal' into a sensitized population. Reasons are given for thinking that the antigens which excite the NLT reaction belong to the homograft system and that the reaction as a whole can be construed as a homograft reaction in reverse (section 7). The first inflammatory episode is interpreted as the outward design of a 'recognition' event, i.e. of a distinctively immunological process, unaccompanied by cell division and in no way dependent on it, that occurs when a lymphocyte is first engaged by an antigen of the homograft system and is committed to the evolution revealed outwardly by the flare-up. A quantal theory of the reaction is proposed, according to which the violent response given by normal lymphoid cells at the peak of their flare-up, and by presensitized cells from the outset, is simply an arithmetic multiple of the recognition process, i.e. it simply consists in more cells doing what a relatively small number does in the first inflammatory episode. Immunosuppressive agents act by preventing this multiplication. This interpretation implies that any 'true' immunosuppressive agent which weakens or abolishes the first inflammatory episode must weaken the performance of a presensitized population to an exactly similar degree.