THE RELATION OF ATMOSPHERIC CONDITIONS TO INTERMITTENT FEVER.

BY HENRY BIXBY HEMENWAY, A.M., M.D.,

OF EVANSTON, ILL.

Among the various State Boards of Health, perhaps none have done more to raise the standard of the medical profession, and to defend the public against impostors, than that of Illinois. None has probably done more for the science of medicine than that of Michigan. What is said. of the Boards, applies especially to their executive officers, who have given the distinctive character to the bodies of which they are members. Dr. Henry B. Baker's laborious care in collecting statistics, and zeal in spreading information are well known, and his work is highly valued. Like other mortals, he is liable to errors of judgment, and we fear that in such an error originated his articles on "Malaria and the Causation of Inter

NOTE.-For tables and diagrams 1 to 12, inclusive, see Dr. Henry B. Baker's paper on “Malaria and the Causation of Intermittent Fever," published in this THE JOURNAL OF THE AMERICAN MEDICAL ASSOCIATION, Vol. xi, p. 651.

mittent Fever," published in THE JOURNAL OF THE AMERICAN MEDICAL ASSOCIATION (Vol. xi, p. 651, etc., and Vol. xv, p. 561, etc.), having been read before the appropriate Sections in 1888. and 1890.

The first article, coming from so high a source, commanded attention; its repetition demands the most careful investigation. If the conclusions are fallacious, the fallacies should be clearly exposed, If his reasoning is correct,, the discovery is important, for it shows that the individual can easily protect himself from the disease by wearing suitable clothing, and keeping his surroundings as even as possible-keeping the house warmed by fires at night, for example. On the other hand, coming from so high an authority, by many the ideas will be regarded as facts until proved otherwise.

The articles in question do not show the judicial spirit of impartial investigation usually found in Dr. Baker's studies, but rather show a labored, and, as Dr. Lee characterized it when the article was first read (THE JOURNAL, Vol. xi, p. 653), "ingenious" argument to demonstrate a pet theory. They show how unconsciously a careful student may sometimes be misled and blinded by impressions. That the results reached are consistent, neither with each other, nor with facts, we shall try to point out.

His first conclusion is that "intermittent fever is proportional, either directly or indirectly, to the average daily range of atmospheric tempera

ture." This sounds well, but reminds one of a test for flour once given by a good lady: "Take a pinch of the flour and throw it against the wall. If it is good it will stick-or else it won't." If the proportion is sometimes direct and sometimes inverse, does it not seem that any argument based upon it lacks stability? If we grant that this double relationship is a possible cause of the fever, has the author presented sufficient proof of the existence of such a fact? The proof is in the form of tables and diagrams, which we will examine.

Figure 1 of Dr. Baker's series shows that during the late war, the number of cases of intermittent fever in the United States Army, and the average daily range of temperature, both increased up to May. In June we find a much greater rise in the curve of cases, but a fall in the daily range of temperature. In July there is a steep decline in the curve of cases, but a rise in the other line. The daily range of temperature continues to rise until, in August, it reaches its highest mark. From July the number of cases increases until September. From their highest points, both lines decline until, in December, they touch bottom. It will be seen that part of the time the lines go in the same direction, but that during much of the time the case line variations are a month behind the other. This Dr. Baker explains by saying that it is due to the accumulation of old cases.

There are two elements of fallacy in this table.

Reference to figure 2 shows that the line representing the daily range of temperature in Lansing is very different from that for the average of the State of Michigan. The case line in Fig. 1 represents a wide range of country-wherever the National troops were stationed. Some were in Texas, some suffered in Florida, Many shook at Charleston, and some were taken sick in the Northern cities. If Lansing atmosphere, in the interior of the State of Michigan, is not to be taken as a basis of comparison for the State of Michigan, how much less should the sickness in the whole army be compared with the atmospheric conditions around a single building, in one corner of the battlefield.

Secondly. Fig. 1 is a partisan table. Such a study as we are now engaged in is not a control experiment. If it could be shown that intermittent fever increased and diminished, either exactly with or inversely with the daily range of temperature, keeping other conditions stable, one might perhaps be justified in making deductions therefrom. Since other conditions cannot be kept stable, they must be carefully examined, and eliminated as possible elements before drawing conclusions, especially if the variation in case and daily range of temperature lines do not bear a constantly definite relationship to each other. It must be remembered that post hoc is by no means always propter hoc. We would therefore have expected at least a report on the mean monthly temperature, and should have liked to have seen