Occasionally, there is no relapse whatever, but convalescence follows the initial paroxysm. This occurred in 10 out of 181 of our cases, and Murchison found that of 2425 cases reported by various authors no relapses occurred in about 30 per cent. Carter describes these under the name of the abortive form, and found them to constitute 23.8 per cent. of all his cases. It is probable, however, that in many cases so regarded either a relapse of very transient duration has been overlooked, or else that an attack of ephemeral fever has been regarded as of specific nature. In ordinary cases the duration of the intermission averages six or seven days, but here, again, considerable variation occurs. In 139 of our cases where its duration could be accurately determined it was as follows:

3 days in 4 cases.7 days in 64 cases.11 days in 1 case.
4 days in 3 cases.8 days in 22 cases.12 days in 1 case.
5 days in 12 cases.9 days in 9 cases.13 days in 1 case.
6 days in 12 cases.10 days in 9 cases.20 days in 1 case.

Despite these variations in the duration of the initial paroxysm and of the first intermission, the average date of the occurrence of the relapse in any large series of cases is about the twelfth day from the primary chill.

The relapse is ushered in with the same striking abruptness as the initial attack. The temperature again rises rapidly to 104° or 106°, and then pursues a continuous course resembling ordinarily that of the primary paroxysm. The difference between the maximum of the two paroxysms is rarely more than 1.5° or 2°, though either may be much milder than the other; as a rule, the highest temperature is attained on the last or penultimate day of the first attack. The duration of the relapse averages three or four days, though it may last but a few hours or a single day, and yet exhibit a rise of 5°, 6°, or 7°; or, on the other hand, it may be prolonged to six, seven, or even more days. Lyons, observing the disease in the Crimea, reports some relapses as having lasted twenty-one days, though it is improbable that a greater duration than seven days occurs without the presence of some complication. The relapse usually terminates by crisis, with an abrupt fall to an abnormally low temperature; though we observed at this time, much more frequently than at the close of the first paroxysm, a gradual subsidence of fever, or lysis. Again the patient regains strength and appetite, but in a considerable proportion of cases subsequent relapses ensue. As a rule, the second, third, and later relapses are attended with a febrile movement of shorter duration and of less severity than the first two paroxysms, and are also separated by intermissions of increasing length. Meschede13 found from a study of 360 cases that the average duration was for the first paroxysm six or seven days; second paroxysm, four or five days; third paroxysm, three or four days; fourth paroxysm, one or two days; fifth paroxysm, one day.

13 Loc. cit.

In a remarkable case given in full [below], the duration of the paroxysms and intermissions were as follows:

First paroxysm, 8 days; first intermission, 9 days.
Second paroxysm, 5 days; second intermission, 1 day.
Third paroxysm, 1 day; third intermission, 6 days.
Fourth paroxysm, 6 days; fourth intermission, 8 days.
Fifth paroxysm, 5 days; fifth intermission, 9 days.
Sixth paroxysm, 4 days; sixth intermission, 10 days.
Seventh paroxysm, 3 days; seventh intermission, 11 days.
Eighth paroxysm, 3 days; followed by convalescence.

The proportion of cases in which more than a single relapse occurs appears to vary in different epidemics. Murchison found that in 1500 cases reported by various authors a second relapse occurred 109 times (1 out of 14); a third relapse, 9 times (1 out of 166); and a fourth relapse, once. Of 182 cases noted carefully by ourselves, a second relapse occurred 24 times (1 out of 7½); a third relapse, 5 times (1 out of 36); a fourth relapse, once; and in the above-mentioned case six or seven relapses.

It follows that the total duration of the morbid process varies from the average of about eighteen or twenty days, in cases with a single relapse, to forty, sixty, or even ninety days. Of course the occurrence of complications may lead to very great modifications of the febrile movement and of the total duration of the disease.

There are several additional points about the febrile process requiring mention. In all the paroxysms there is a greater tendency to local or general perspirations than is met with in other continued fevers, and occasionally there are rigors or slight chills about the same hour on several days after the invasion or on the day preceding the crisis. It has been noted also that, even when the temperature is very high, the quality of the heat, as judged by the feeling of the skin, is different from that in typhus fever, and that the peculiar pungent irritating sensation known as calor mordax is rarely marked. But a more important peculiarity is the fact that the extreme temperatures (106°, 107°, or 108°) that are frequently observed in relapsing fever for several days in succession do not appear to involve any great increase of danger, and in particular are not attended with the production of the grave nervous symptoms so often met with in connection with hyperpyrexia in typhus and typhoid, and often regarded as the direct result of the exalted temperature itself. This striking fact is of much interest in its bearing on the theory of hyperpyrexia, and may possibly be explained by some marked difference in the conditions of heat-dispersion in these different diseases.