252 
MALARIA 
attempt to obtain an antigen of low anti¬ 
complementary property but of high 
specificity and sensitivity. Saline emul¬ 
sions of heart blood gave specific results, 
yet sera from 25 known cases of malaria 
fixed complement in only 48 per cent of 
the instances. The chief difficulty encoun¬ 
tered in all of the investigations was the 
lack of sufficient quantities of parasitic 
material to serve as an antigen. This 
obstacle was overcome by the utilization of 
P. knowlesi parasites from rhesus monkeys 
which, unless treated, usually die from an 
overwhelming parasitemia (Coggeshall and 
Eaton 1938a). In these animals it was 
found that specific complement-fixing anti¬ 
bodies appeared shortly after the acute 
attack was terminated either by quinine 
or by immune serum and that the antibody 
titer dropped to a more or less stationary 
level. During the chronic stage of the in¬ 
fection when parasite relapses occur at 
irregular intervals, the titer is low preced¬ 
ing each relapse and is elevated consider¬ 
ably following each relapse. There is no 
evidence to show that specific complement¬ 
fixing antibodies are identical with the pro¬ 
tective antibodies, although frequently ani¬ 
mals with a high protective serum titer 
show a marked concentration of comple¬ 
ment-fixing antibodies. Complement-fixing 
antibodies can be produced with ease in 
rabbits and monkeys by the injection of 
killed P. knowlesi parasites, although this 
procedure does not confer an active im¬ 
munity in the monkey or produce protective 
antibodies in the rabbit. Possible explana¬ 
tions are that separate immunity mecha¬ 
nisms are involved or that perhaps it is 
only a quantitative effect. 
The complement-fixing antigen is present 
in some unidentified constituent of the 
parasite, and it also can be readily detected 
as a soluble antigen. The greatest concen¬ 
tration of the soluble antigen occurs in the 
blood serum at the height of the acute 
attack of P. knowlesi malaria, and, as 
would be expected, the amount is related 
directly to the parasite count. The antigen 
has the ability to fix complement in im¬ 
mune animals and probably is concerned 
with the production of the complement-fix¬ 
ing antibody; at least these antibodies are 
readily produced in normal animals by 
injection of the soluble antigen. When 
normal monkeys are immunized, apparently 
only complement-fixing antibodies are de¬ 
veloped as no agglutinins or protective anti¬ 
bodies appear in the serum, and there is 
no resistance to infection. Clinically, the 
antigen is unstable in acids and alkalies, 
is heat-stable up to 56° C and appears to be 
a constituent of the albumin fraction of 
the serum. 
Complement Fixation Reaction in 
Human Malaria 
In the studies on the P. knowlesi com¬ 
plement-fixing antigen one of the most 
interesting developments was the discovery 
of the broad antigenic power of the antigen. 
It would bind complement in human ma¬ 
laria serum with the same degree of sensi¬ 
tivity as was found with homologous mon¬ 
key immune serum (Eaton and Coggeshall 
1939). This finding immediately suggested 
the investigation of the practicability of the 
reaction as a diagnostic aid in malaria, 
which is a real necessity, especially in those 
low-grade infections or treated cases where 
parasites are too scanty for microscopical 
detection. As pointed out above, all pre¬ 
vious tests depended upon infected human 
material as a source of the antigen, which 
naturally limited their usefulness. In one 
average-size rhesus monkey it is possible 
to obtain approximately 10 cc of packed red 
cells with over 60 per cent containing 
malaria parasites. The antigenic property 
of this material can be shown to exist in 
dilutions up to 1:320, and as the anti¬ 
complementary range is absent or ex¬ 
tremely low, it would seem to be a widely 
adaptable test even for large-scale use. It 
was found that the test was highly specific 
for malaria in that practically no false 
positive reactions were obtained although 
large numbers of sera from normal indi¬ 
viduals, from convalescents from a variety 
of infectious diseases, and from patients 
with strong positive Wassermann reactions 
were used as controls. Patients either with 
