Doring reported the findings to be almost always positive at the beginning of the attack.

The blood examinations show otherwise a striking and rapid decrease of erythrocytes, which is explained by their solution. From numerous cases in which blood counts were made I take two from Boisson. In one patient, before the paroxysm, he found 1,700,000; after the paroxysm, 670,000; in the second, before the paroxysm, 2,400,000; after, 1,600,000. These figures do not at all correspond with Ponfick's, yet the correctness of the latter's is by no means questioned, though we must assume an extraordinary rapid regeneration to explain them. The frequent occurrence of normoblasts, megalocytes, and microcytes in the blood would indicate such a regeneration. F. Plehn suggested that a thickening of the blood as a result of profuse diarrhea and vomiting must also be taken into consideration.

Bastianelli found in every case of blackwater fever a polynuclear leukocytosis. [The polynuclear leukocytes often constitute 90 per cent, of all leukocytes.-Ed.] Increase of the blood platelets has. likewise been asserted.

Examinations of the blood serum for dissolved hemoglobin are in too small numbers to deduce conclusions of any significance. Boisson, who had the opportunity of conducting examinations accurately in the hospital at Lyons, found in one case the blood serum reddish in color. Spectroscopically oxyhemoglobin, methemoglobin, and urobilin were seen. [On the other hand, there may be no hemo globinemia, even while hemoglobinuria persists.-Ed.]

It is, moreover, doubtful whether in every case the solution of the red blood corpuscles takes place in the circulating blood-that is, whether in every case a hemoglobinemia can be found in the peripheral vessels, or whether there are cases, as has been asserted, in which hemoglobinemia does not occur at any stage of the paroxysm. For these latter possible cases the hypothesis has been suggested that the destruction of the erythrocytes takes place in the kidneys.

Berthier found, in two cases examined spectroscopically, only the bands of oxyhemoglobin, while the color was the same as that of normal blood serum, and assumes that malaria hemoglobinuria is not the result of hemoglobinemia, but of hemorrhages into the kidneys, from which the urine extracts the blood coloring matter. We do not wish to question the renal hemorrhages (apoplexie renal) found by Pellarin, but we do believe that Berthier took the blood from his patient at a time when it would be impossible to say with certainty that no hemoglobinemia had existed. He took the blood always several hours after the beginning of the paroxysm; in other words, at a time when the liver and kidney would already have removed the greater part of the hemoglobin. Attempts made by ourselves in several cases of paroxysmal hemoglobinuria taught us caution in this respect.

We may refer here to the similarity of the symptoms of black water fever with those of paroxysmal hemoglobinuria. In both we see a chill, rise of temperature, anxiety, cyanosis, and pain in the region of the kidneys as constant symptoms. It seems, therefore, justifiable to assume that in blackwater fever only a part of the symptoms are to be attributed to the toxic effect of the malarial parasites, while the others should be attributed to the destruction of red blood corpuscles.

The duration of the disease is very variable. In mild cases we seldom see more than two or three paroxysms and often only one. In these the duration of the disease is not more than two to five days. Severer cases with subcontinued fever may last ten days, though the very severe ones, with fulminating symptoms, may produce death in two or three clays. According to Corre, the duration of the favorable cases is three to fifteen days (in one case, twenty two days); that of the fatal cases, two to twelve days.

In the majority of cases blackwater fever terminates in recovery. This is naturally preceded by a pretty long convalescence. It is needless to say that malarial cachexia follows in many cases, and these cannot, therefore, be regarded as cured.

The mortality statistics differ widely. Factors depending on the time, place, and individual seem to influence the severity of the case. How far the therapy is to be taken into consideration will be discussed in the section on Therapy.

Corre in Nossi Be had a mortality of over 50 per cent, in cases treated outside the hospital; of 28 per cent, in hospital patients. Barthelemy-Benoit had in Senegal, 25 per cent.; Guiol in Madagascar, 31.6 per cent.; Cassan in Goree (West Africa), 32.1 per cent.; Berenger-Feraud, who took his statistics from 286 cases treated in West Africa, 66 deaths-about 23.1 per cent.; Steudel from 18 cases-17 per cent.; F. Plehn, 39 cases-over 10 per cent. ;

A. Plehn, 53 cases-9.8 per cent.; O'Neill (Madagascar), 50 cases- 4 per cent. Pampoukis estimates for Greece a mortality of 6.6 per cent. Segard found the cases in Madagascar not so pernicious as generally considered.

The exitus letalis occurs in different ways. Sometimes the subjective symptoms improve immediately before and the patient dies unexpectedly in syncope. More frequently a typhoid condition with algid symptoms develops: Small, compressible, rapid pulse; cool skin; clammy sweat; crusts on lips and tongue; stupor and occasionally coma or convulsions (Guillaud). In other cases the skin is dry, the pulse soft and small, hemorrhages occur from the nose, mouth, or intestine, the alas of the nose and lips become crusted, and an obstinate singultus (an especially common symptom) develops. The end is ushered in by complete or almost complete anuria, involuntary fecal evacuations, and delirium. These symptoms may continue several days. This form lasts the longest, death occurring frequently, in the second week, when the icterus may have almost disappeared (Raimond).

The diagnosis of blackwater fever is based on the cardinal symptoms-fever, icterus, hemoglobinuria-and on the presence of the malarial parasite.