cells are found to be increased in number. There may be some congestion and even ulceration of the intestinal mucosa, but this is not an essential feature.
Diagnosis.— The diagnosis of undulant fever from typhoid is an important practical matter. It is exceedingly difficult in the early stages. Principal reliance has to be placed on the presence or absence of rose spots, of diarrhœa, of joint - complications, of sweats, the locality where and the season in which the disease was contracted, and, if available, the agglutination and precipitation tests.
According to Nicolle, there is a mononuclear increase, 70-80 per cent.; apart from this, microscopical blood examination, unless in a negative sense, is of no value in diagnosis.
Wright has shown that, both as regards sedimentation and agglutination, the germ of undulant fever reacts to the serum test in the same way as, and even more markedly than, Bacillus typhosus. A. weaker dilution (never less than 1 in 30 or 50) than in typhoid must be used. Dead cultures give the reaction, and can be conveniently kept in stock for diagnostic purposes; they are grown on agar, washed off with sterile normal saline, killed by heat at 60°C. or by formalin vapour. These observations have been abundantly confirmed. The agglutination reaction appears early in undulant fever as compared with typhoid, being available for diagnostic purposes by the end of the first week of the fever. It persists long after convalescence, often for years.
The unreliability of the serum test for undulant fever is in part explained by the researches of Nègre and Haynaud, who found that certain strains of M. melitensis agglutinated with normal sera, whereas other strains did not agglutinate with undoubted undulant-fever sera. The latter they name the " paramelitensis group." Bassett - Smith concurs in this. Therefore, before concluding from a negative reaction that any given case is not undulant fever, the serum should be tested with a paramelitensis culture.
After the fever has gone on for several weeks diagnosis is, of course, easier; in the early stages, on
