more rapid disappearance of the parasites from the blood, and the greater number of deaths among the white rats.

The agglutinated parasites have a most orderly arrangement in the shape of a rosette, with their posterior ends close around a central point and their flagella at the periphery. Each parasite has its own fine undulating motion.

AGGLUTINATION.

1. By immune serum.-Immunity and agglutinating power go hand in hand. As in late years agglutination has been proven for so many kinds of bacteria by their specific sera, so trypanosomes are found to respond in a somewhat similar way to immune serum.

Laveran and Mesnil probably made the most complete study of agglutination. The trypanosome blood which they used was subjected to defibrination, which of course left the corpuscles and the trypanosomes in the serum. We have found that the substitution of clotting for defibrination will show the agglutination to much better advantage in the hanging drop, since there will be no corpuscles in the field to obscure agglutination. The immune blood and the trypanosome-bearing blood may be drawn up into separate fine capillaries and allowed to coagulate. The clot is then drawn out at one end of the tube, leaving the clear immune serum behind in the one case and the clear serum containing trypanosomes in the other case. Now the dilutions can be made just as in the Widal reaction. We have, however, taken another precaution which Laveran and Mesnil did not observe. We removed all agglutinin from the trypanosome serum before starting the tests. We were led to this by the study of auto-agglutination.

If we should select for our agglutination tests blood in which autoagglutination was already noticeable, we would immediately fall into the error of getting agglutination with ever so great a dilution of the immune serum, because the trypanosomes were agglutinated before we began. Again, if we selected blood in which there was no autoagglutination, but in which there was considerable agglutinin, but still not enough to produce an auto-agglutination, we would still be in error if we attempted to determine the agglutinating power of an immune serum by testing it on trypanosome-bearing serum which was just on the verge of auto-agglutination. We must therefore take into consideration one element which does not enter an agglutination test on bacteria. The bacterial pure culture has a fixed, uniform composition, and until we are able to grow trypanosomes in pure culture on artificial media we will have to consider the element of agglu tinin in the serum which contains the parasites.

Our plan was to draw the trypanosome blood from the rat, allow it to coagulate, draw off the serum containing the parasites, and dilute it with plain distilled water and filter it through a porcelain filter

under the influence of a vacuum.

After four or five washings with large amounts of water, we let it filter until there remained behind a volume of fluid which equaled the original amount of serum. In this residue were the trypanosomes free from agglutinin. Dilutions were then made of the immune serum, and it was tested on the washed parasites. Some immune sera will not agglutinate in a dilution greater than 1:1. An agglutinating power of 1:5 or 1:10 is common. One of our rats showed typical agglutination in a dilution of 1:200. It is interesting to watch in hanging drop an agglutination by a weak serum. At first two or three parasites are seen, joined by their posterior ends. Others come up toward the center of agglutination, recede for some distance, and later join the others. Some disengage themselves from the rosette and then rejoin it, until finally a wellarranged rosette is formed. Two small rosettes will gradually approach each other and then unite to form one mass, which in turn is joined by others of smaller or larger size.

A remarkable fact is that the agglutinated parasites do not lose their motility. There is not the diminution of motility before agglutination that is seen in a typhoid reaction, and while agglutinated each parasite retains a regular vibration. In agglutination with a strong serum the process takes place rapidly. The parasites rush together in great numbers, and the masses may be of macroscopic size. If a strong specific serum is used, the agglutinated masses are very compact and the individual parasites are tightly drawn together, so that there is little motion. In a general way the parasites are all pointed toward a center, but still they overlap and cross each other very much. With a weak serum the parasites are held together in a loose manner, permitting of more individual movement to each organism and more orderly arrangement, and there are fewer parasites to each rosette.

We often found agglutination almost complete, in which case very few parasites were to be seen free in the field. It may be only partial. The parasites may remain agglutinated until their death, or, if the serum is weak, a disagglutination may follow. Specific sera exposed to 55° C. for thirty minutes did not lose the agglutinating power, but a temperature of 65° C. maintained for half an hour destroyed its activity.

Laveran and Mesnil found that trypanosomes killed by chloroform or formalin were agglutinated by the same sera which agglutinate the living, but the parasites have no orderly arrangement in the mass. It is a remarkable fact that Rabinowitsch and Kempner found that "the trypanosome serum shows in no way whatever the property of agglutination."

2. By normal sera.-The action of normal sera in dilution of 1:1 was tested on the trypanosomes. The cat and horse sera were strongly agglutinating. The goat and rabbit sera were feebly agglu20564-No. 11-03-2

tinating. The sera of the white rat, white mouse, and guinea pig were negative.

Laveran and Mesnil state that the normal sera of the chicken and horse caused complete agglutination. The sera of the sheep, dog, and rabbit produced a partial reaction, while the pigeon, frog, guinea pig, white and spotted rats, and sewer rats caused none whatever. They also state that, although the blood of chicken and horse have agglutinating power, they do not protect against infection.

3. In the ice chest.-If trypanosome blood is drawn from the rat under aseptic conditions and sealed in pipettes and placed in the ice box, the parasites will usually join into beautiful rosettes after twenty-four hours. We found them agglutinated in blood which had been eighty-three days in the ice chest.

TRANSMISSION OF THE DISEASE.

1. By intraperitoneal inoculation.-Although white and spotted rats have never yet been found to harbor trypanosomes spontaneously in their blood, we find in them a very susceptible host for experimental inoculations, and it was with them that most of our work was done. Comparative studies show that a heavy blood infection is obtained by intraperitoneal injection sooner than by any other form of inoculation. The blood to be injected is mixed with 1 c. c. of saline solution or bouillon and injected with a hypodermic syringe. The period which elapses between the intraperitoneal inoculation and the first appearance of the parasites in the blood is variable, depending upon the amount of trypanosomes injected and the stage of develpment of the injected parasites.

Rabinowitsch and Kempner place the time at three to seven days, although they observed parasites in a few instances within the first day.

Laveran and Mesnil give three to seven days as the average time before the blood infection. They found a few in the circulation after five or six hours.

Jourgens found that the first presence of parasites in the tail blood occurred three or four days after inoculation, seldom later. After inoculation with 1 c. c. of blood he found them in tail blood in several instances after twenty minutes.

The parasites appeared in the tail blood in our cases usually on the second day. This was chiefly because we injected as a rule larger doses and used heavily infected blood in which were many developmental forms.

There has been considerable discussion as to where the principal seat of multiplication takes place in intraperitoneal inoculation. Rabinowitsch and Kempner regard the peritoneal fluid as a better nutritive medium for the development of the parasites than the blood and think that the chief seat of development and multiplication is in the peritoneum. In one to five days after intraperitoneal injection

of trypanosome blood they found in the peritoneum several engaged in development. They think that as soon as development is perfected in the peritoneum the parasite disappears from the peritoneal fluid into the blood. They give some weight to the fact that in one case of intravenous injection of parasites multiplication did not occur until the fifth day. Laveran and Mesnil state that there is less multiplication in the blood than in the peritoneum.

Our observations speak for the blood as the principal seat of development. On examination of the peritoneal fluid at varying times after injection we did not find division forms, nor did we find the first existence of parasites in the tail blood accompanied by division forms. The most reasonable explanation seems to be that if a large amount (0.75 c. c.) of heavily infected blood be injected into the peritoneal cavity, the small young forms and the long slender adults pass at once by the lymphatics into the general blood stream in sufficient numbers to be detected in cover-slip preparations within twentyfour hours or even within twenty-minutes, but the rosettes and other large division forms which are injected into the peritoneal cavity are prevented by their size from passage through the lymph channels and remain behind in the peritoneal cavity until their division is complete, when the young then pass through the lymph passages into the blood, leaving the peritoneum permanently free from parasites.

Daily examination of the blood for one to two days after the first appearance of parasites in the tail blood shows at most only a very gradual increase in the number of parasites present, but suddenly there comes an enormous swarming of the blood with parasites, and the presence of rosettes and other division forms indicate that multiplication is going on in the blood. The slide from which the microphotographs (Pl. III, figs. 11, 12) were made shows at least two dozen rosettes and was taken from the rat on the second day after the first appearance of parasites in the tail blood. It is not unusual to find parasites in the proportion of 1:2 red blood cells. Exceptional instances are met with in which the blood corpuscles are outnumbered by the trypanosomes. The duration of the period of multiplication is often no more than twelve to twenty-four hours. By the fourth day after the first entrance of parasites into the blood the height of infection has been reached. Rabinowitsch and Kempner found no division forms in the blood on the fourth day after the first appearance of parasites in the tail blood.

2. By subcutaneous inoculation.-Blood infection occurs by this form of injection a little later than by intraperitoneal inoculation. The shortest time in our cases between inoculation and the appearance of parasites in the blood was three days. Multiplication proceeded in the blood until it swarmed with myriads of parasites. This would seem to be additional evidence to the superiority of the blood over the peritoneal fluid as a nutritive medium for the development of the parasites, for in cases where the parasites had advantage of

the peritoneal fluid in addition to the blood their number did not reach a point beyond the number obtained by subcutateous injection.

3. By intrastomachal injection.-We have read some discussion bearing on the natural mode of infection with trypanosomes in which infection followed the eating of a trypanosome rat by a healthy one. It was held that this could not be considered a case of intrastomachal infection because the possibility of the entrance of the parasites through wounds about the mouth, lips, or teeth had not been excluded.

We therefore arranged a series of experiments in which we thought all likelihood of infection through wounds was removed and that infection occurred through the stomach. Twelve white rats were chloroformed sufficiently to prevent any struggling. Then a smallsized catheter well oiled was passed into the stomach without encountering any resistance; injection of trypanosomes was made through the catheter, and the rats were then placed in separate cages and examined daily for parasites in the tail blood. Eleven out of the 12 developed blood infections fully as heavy as was obtained by any other form of inoculation. The time which elapsed before their appearance in the blood was as follows: Two in four days, one in five days, three in six days, three in seven days, and two in eight days. We see that infection is considerably delayed by this method, the earliest being in four days and the latest in eight days.

4. Transmission by feeding.-The success of the intrastomachal injections naturally lead to a series of experiments to determine whether infection would not occur by feeding when all precautions were taken to prevent any wounding about the teeth or mouth.

Seven white rats, apparently free from mouth wounds, were put into separate cages to prevent fighting, and they were fed with soft food, so that no wounds would result from the gnawing of bones. They were each given a single feeding with the entire blood of a trypanosome rat. No other part of the infected rats was given to them. Trypanosomes appeared in the tail blood in five of the seven at periods of time which averaged six days.

Wild rats were then subjected to similar feeding experiments. In the blood of five wild rats we found parasites after three, seven, eight, nine, and ten days. Some of the rats had enormous numbers in their blood, while others had comparatively few.

We conclude from these experiments that infection may take place through the digestive tract and that the spread of the disease among wild rats may be due to feeding upon one another, especially since the instinct of fighting and pluck is so well implanted in them and is brought into action on slight provocation. We found it necessary to separate the wild rats in our stock cage to prevent losses from injuries inflicted on the weaker ones by the stronger.

After beginning the feeding experiments we were surprised to read the results obtained by Rabinowitsch and Kempner. They made intrastomachal injections through a stomach tube in four rats, but were