has recently been so much improved that this complication no longer appears to be of any material significance. As a matter of fact, we have tested out this question by means of tests which indicate the specific effect of insulin and have found that the toxicity of the recent batches is altogether due to insulin. The details of the rat test have not been completely worked out at the present time, and the purpose of the first of this series of papers dealing with this subject is to call attention to one of the factors which appears to play an important part in the accuracy of the test, namely, the influence of the atmospheric temperature to which the animals are exposed after the injection of the drug.

The work was carried out as follows:

Young, healthy, albino rats from a standard strain and weighing about 50 to 60 grams were put on a standard diet of the following composition:

This diet, which has been extensively used in the Hygienic Laboratory, has given entire satisfaction in the bio-assay of arsphenamine. Rats kept on this diet show a normal growth curve and appear to be in excellent physical condition. As soon as the animals had reached 100 to 110 grams, they were used for the test. The food was withdrawn 18 hours before the test, at which time the animals were weighed and injected subcutaneously with graded doses of a recent lot of insulin, kindly supplied to us for experimental purpose by Doctor Clowes, director of research of the Eli Lilly Co. The drug was kept in the refrigerator (10° C.) until used, and was diluted on the day of the test so as to yield a 10 per cent solution in sterile physiological saline. The original product, as received, contained 10 units per cubic centimeter, and the diluted solution therefore contained 1 unit per cubic centimeter. The doses given in the table refer to the undiluted preparation and are expressed in cubic centimeters per kilo bodyweight.

That the atmospheric temperature may play a rôle in the toxicity of insulin was suggested to us by the differences in the results obtained on relatively cool days as compared with those obtained on hot days. The problem of keeping the temperature constant was solved by working on relatively warm days and exposing one set of rats to the ordinary atmospheric temperature and another set to air in a cabinet (ventilated) which was cooled to the desired temperature by means of a crushed ice and salt mixture. An automatic tempera

ture register was employed. As a rule it was quite easy to keep the temperature fluctuations within 1° C.

Three ranges of temperature were used, 15° to 17°, 18° to 22°, and 28° to 30° C. These temperatures about cover the range of atmospheric indoor temperature in this country during the various seasons of the year, with exception of unusually warm days.

In all, 270 rats were used, divided into lots of 30 animals. The results are summarized in the table. It will be noted that the difference in the mortality rate is not appreciable between the animals exposed to 15° to 17° C. and 18° to 22° C., respectively. There is, however, a great difference between the percentage mortality in the latter group of animals and that of the group exposed to 28° to 30° C.

It is furthermore evident that the time of survival of those animals which ultimately died on a given dose of insulin is progressively shortened with a rise in atmospheric temperature. We also observed that the characteristic symptoms of insulin poisoning in the rat made their appearance much more rapidly at a high temperature. These symptoms consist in salivation and a gradually increasing weakness. Finally, the animal passes into coma, the body feels cold, and some animals develop convulsions, rolling over sidewise. Respiration is greatly depressed and often so shallow that the animal appears to have died, although it lives for a considerable time longer. Death is always preceded by respiratory failure.

It is difficult to give a satisfactory explanation for the temperature effect described. It is possible that it may be due to differences in the rate of absorption of the drug from the subcutaneous tissues. A low atmospheric temperature may delay absorption, owing to a more or less active peripheral vasoconstriction; and, vice versa, a higher temperature may promote absorption by peripheral vasodilatation.

At any rate, the fact remains that the toxicity of insulin in albino rats is greatly influenced by the room temperature, and this factor must be controlled in work of this kind.

The effect of atmospheric temperature on the toxicity of insulin.

DENGUE FEVER.

By C. ARMSTRONG, Passed Assistant Surgeon, United States Public Health Service.

Introduction.

Definition.--Dengue fever is an acute, insect-borne fever of unknown etiology which is endemic in the Tropics and which at times becomes epidemic. It may spread to temperate regions in the hotter portions of the year. In typical cases the disease is characterized by a sudden onset; an initial erythema; pains in the head, trunk, and limbs; fever of short duration, which shows a saddle-back curve; a slow pulse; marked leucopenia; a terminal rash; slow convalescence; and practically no mortality. Its pathology in uncomplicated cases is unknown.

Importance. The disease is important because it attacks large numbers of people, causes much suffering, and incapacitates its victims for varying lengths of time. It is especially of military importance, as whole Army units may be disqualified for duty by an epidemic of the disease.

Geographical distribution.-Dengue is mainly confined between parallels 32°, 47′ N. and 23°, 23′ S., but has been known to extend beyond these limits in hotter portions of the year, as, for instance, to Philadelphia, Constantinople, and Athens.

Previous epidemics.-What was probably dengue fever was first described in 1779 at Cairo by Gaberti; the following year the disease was described by Rush in Philadelphia and by Bylon' in Batavia. Many epidemics have been recorded since, some of which are as follows: Spain, 1793; Peru, 1818; India and Suez, 1825; United States, Mexico, West Indies, and South America, 1826-1828; India, 1824-1828; Arabia, 1835; India, 1836 and 1844: Bermuda, 1837; Egypt, 1845; India, 1847; Senegambia, 1845-1848; Brazil, 18451849; United States and West Indies, 1850-1854; India, 1853-54; Tropics of Eastern Hemisphere, 1870-1875; Louisiana (U. S. A.), 1872; Tripoli, 1878; Caribbean, North America, and Egypt, 1880; Caledonia, 1884-85; Fiji Islands, 1885; Texas (U. S. A.), 1885; Tripoli, 1887; Asia Minor, 1889-90; Texas, 1894 and 1897; Hawaii, 1903; Texas, 1907 and 1918; Bermuda, 1915; Egypt, 1916; and southern United States, 1922.

Etiology.

As dengue is a mosquito-borne disease it is natural to infer that the causative organism is present in the blood stream, and various authors have demonstrated that this is true by the injection of

1 Stitt suggests that Gaberti may have described relapsing fever and thinks that Rush should have the honor for the earliest recognizable description of dengue.

2 Cited from Hirsch.

volunteers. It has also been demonstrated that the organism is filterable. Cleland, Bradley, and MacDonald, 1916, produced dengue in volunteers by the injection of washed corpuscles as well as of serum and plasma, thus indicating that the virus is present in all elements of the blood. With the exception of these few facts the nature of the virus is unknown.

From time to time numerous workers have reported various "causative organisms," but the reports lack confirmation.

In 1873 Charles described an organism in the blood; in 1886 McLaughlin described a coccus; in 1903 Graham described a hematozoan; in 1910 Nagib Ardate described bodies in the corpuscles which he thought were the same as Graham had described. Eberle, in 1904, described his plasmeba, and in 1906 Reiche described some very actively mobile translucent bodies in the blood of dengue cases. In 1904 J. C. D. Allen described spirochetes from the sputa of several cases, and in 1919 McMullin suggested that the disease was possibly an anaphylactic reaction brought about by repeated injections of protein by the mosquito. Craig has suggested that the causative organism is probably a spirochete, from certain analogies which dengue bears to yellow fever.

Couvy, in Beirut, 1914, described short, slender spirochetes having two or three turns and fine extremities, which he found in the blood of patients drawn two or three hours before the rise of fever, but not at other times. In 1921 he again found spirochetes not only before the onset of fever, but from 3 to 48 hours thereafter. They were not numerous. Blood inoculated into rabbits caused fever, and Couvy found spirochetes in their blood at the time of onset and relapse. Transfers were made through three rabbits without attenuation. In two animals, crushed infected sand flies gave febrile attacks, and spirochetes were found. While the author considered this outbreak to be one of dengue fever, his charts would well illustrate phlebotomus fever, and there seems to be some doubt as to the identity of the fever which he was studying.

Holt, 1922, described polymorphous organisms seen in blood of patients and inoculated animals.

These observations, however, all lack confirmation, although many attempts to find the organism have been made.

Epidemiology.

Climate.-Hirsch noted that seashore cities and towns upon large rivers were especially liable to the visitation of dengue, but the disease may also travel inland as has been the case in India and the United States.

Cited from Hirsch.

With the exception of temperature, there is very little dependence of the disease upon climatic conditions. It flourishes in wet weather but it also occurs in times of drouth. The epidemics at Philadelphia. 1780; Goojeret, 1824; St. Thomas, 1827; Senegambia, 1860; Southern States, 1922, and many others occurred in very dry weather. Frosts soon bring the disease under control.

Age and sex distribution of cases.-When the disease spreads among a nonimmune population it has been repeatedly observed to attack both sexes and all ages indiscriminately, although Argramonte, in Habana, 1905, states that the disease did not occur or at least was not recognized in children under 5 years of age.

Diffusion and numbers attacked.-The spread of dengue is similar to that observed in yellow fever, but more rapid. In regard to the numbers attacked and the rapidity with which they are stricken, epidemic dengue is second only to influenza. At Austin, Tex., 1885, it is estimated that 16,000 out of 22,000 population were attacked; at Cairo, Egypt, in 1880, four-fifths of the people are said to have suffered with the disease; at Lima, Peru, 1818, only a few persons are said to have escaped; in Galveston, Tex., in 1897, it is estimated that one-half of the population suffered, and in 1922, 60 per cent; in Monroe, La., 1922, perhaps one-fourth had the disease. In many epidemics, however, smaller proportions of the population are often attacked, owing to the advent of cool weather or perhaps to the presence of an immunity to the disease. At Monroe, La., 1922, the writer observed a sharp, severe outbreak among the poorer negro classes on the eastern boundary of the city, where piped water was not supplied and screens were seldom found. This localized epidemic had attacked perhaps 1,000 persons before its existence was known to the local health authorities. The disease spread slowly in the better districts of the town for several weeks until the epidemic was terminated by frosts. The same sort of spread was observed by Manson at Amoy, China.

Influence of economic status, crowding, etc.-Before Graham demonstrated that dengue could be transmitted by the mosquito, older writers had attributed considerable importance to filth, poverty, and overcrowding. These conditions are important in so far as they are related to the life habits of the insect carriers. The disease is practically confined to cities, showing little tendency to spread to rural areas and villages.

Case chronology.-There are but few figures bearing on the case chronology of dengue (except in military groups); and owing to the great rarity of deaths from the disease, mortality statistics are also lacking. Kennedy gives the accompanying curve for an epidemic of dengue in India, 1912, among a group of soldiers (Fig. 1). The rise and fall in this epidemic curve are seen to be very sudden,