THE DISTRIBUTION OF YELLOW FEVER IMMUNITY IN NORTH AMERICA, CENTRAL AMERICA, THE WEST INDIES, EUROPE, ASIA, AND AUSTRALIA, WITH SPECIAL REFERENCE TO THE SPECIFICITY OF THE PROTECTION TEST WILBUR A. SAWYER, JOHANNES II. BAUER, AND LORING WHITMAN hwrz the bdm-atories of the international Hedth ~iuision, Rockefeller Foundation, New York Received for publication December 8, 1936 A survey of the world for yellow fever immunity in man has been carried on during the past five years by the International Health Division of The Rockefeller Foundation with the co- operation of the many governments concerned. The surveys of Africa and South America produced evidence of a much wider distribution of yellow fever than had been suspected, and required separate reports. The final report for Africa has been published ' by Sawyer and Whitman (1) with a review of the results of the earlier participation in the survey by Beeuwkes and his associates (2, 3) and Stefanopoulo (as reported by Boy6 (4)). The survey of South America by Soper and his colleagues is notyet complete, but preliminary reports may be found in publications of Kerr and Pat50 Camargo (5), Sawyer (6), and Soper (7). In the other parts of the world investigations were made in as many countries as seemed necessary to determine the general boundaries betwe, i the recently infected and the non-infected regions and to find out where intensive local investigations were required. The communication now being presented is the final report on these surveys outside of Africa and South America and includes obser- vations for North America, Central America, the West Indies, Europe, Asia, and Australia. Reports have been made for some of the countries by Hughes and Sawyer (8), and Sawyer (6), but the essential findings of the earlier investigations have been com- bined with those now set forth. 137 138 W. A. SAWYER, J. H. BAUER AND L. WHITMAN DISTRIBUTION OF YELLOW FEVER IMMUNITY 139 The immunity survey was possible because an attack of yellow fever leaves a lasting immunity, the presence of which can easily be demonstrated. To determine the presence of immunity, a specimen of the blood serum of the person to be tested is injected together with a relatively large amount of living virulent yellow fever virus into a susceptible animal. The survival of the test animal is taken as an indication that specific immune substances, the result of a previous attack of yellow fever, were present in the serum and protected the animal against a fatal infection. Shortly after the discovery of an animal susceptible to yellow fever by Stokes, Bauer and Hudson (9), rhesus monkeys were used in such protection tests by Beeuwkes, Bauer and Mahaffy (10) to determine whether yellow fever had been present in cer- tain communities. Because of the excessive cost of monkeys, the extent to which such tests could be applied for epidemiological investigation was necessarily limited. The important discovery by Theiler (11) that white mice are susceptible to yellow fever when inoculated intracerebrally made possible a wider applica- tion of the protection test to epidemiological investigations. The technique used throughout the immunity survey in de- termining the presence of yellow fever immune substances in serum was that of the intraperitoneal protection test in mice, as worked out by Sawyer and Lloyd (12). The blood specimens were random samples from persons who had lived all their lives in the locality under investigation, As in the African surveys an attempt was made to obtain 25 specimens or more from adults and the same number from children in each locality, but in cer- tain regions in which it was highly improbable that yellow fever had been present in recent years adults alone were tested. If less than 10 sera were tested in one of these age groups in a lo- cality, the results for that age group were omitted from the maps presented with this report, as though no specimens had been collected. In the previous surveys of Africa persons aged 16 years or less were classed in the tables and on the maps as children and those aged 17 years or over as adults, but on the maps in this report persons who were under 15 years of age were considered as children and those aged 15 years or over as adults. The specimens were collected for us through the kind coopera- tion of local medical authorities. SPECIFICITY OF THE YELLOW FEVER PROTECTION TEST AS SHOWN BY THE RESULTS OBTAINED WITH SERA FROM NON-YELLOW FEVER COUNTRIES From time to time questions have been raised as to the degree of specificity of the yellow fever protection test and the reliability of the survival of mice inoculated with living virus and an un- known serum as an indication of the presence in the serum of yellow fever immune substances acquired as the result of experi- ence with yellow fever virus. Inasmuch as the protection test was being employed extensively in epidemiological investigations, it was considered of prime importance that information be se- cured regarding its specificity. There was already much evidence that protective antibodies appeared regularly after an attack of yellow fever, but not enough evidence to determine to what extent protective substances might be encountered in the sera of persons who have never been infected with yellow fever virus. With a view to obtaining the needed information and establish- ing a standard with which the results of the epidemiological `surveys could be compared, a large number of serum specimens were tested from widely separated regions in which exposure to yellow fever virus seemed impossible as the disease had never been known to be present. The countries from which these specimens were collected were Australia, Ceylon, China, Java, India, Federated Malay States, Philippine Islands, and Syria. The results are given in table 1 and map I. As seen from this table, only two specimens, or 0.23 per cent, of a total of 876 speci- mens gave protection against yellow fever virus. Both of these came from India, and as far as we have been able to ascertain neither of the donors had ever been exposed to yellow fever infection. Although the results indicate that the protection test is highly specific, and the proportion of normal sera showing yellow fever virucidal properties is exceedingly small, the fact that such sera are encountered at all seems to justify a more detailed discussion. 140 W. A. SAWYER, J. H. BAUER AND L. WHITMAN Clear cut and highly specific results are obtained in the intra- peritoneal yellow fever protection tests in mice only when massive doses of virus are used. In the early stages of the yellow fever protection test work, a 10 per cent emulsion of infected mouse brain was used as a source of virus. The results were sometimes TABLE 1 Results of protection test survey in non-yellow fever countries Australia, hlelbourne.. Ceylon, Colombo.. . China, Peiping.. Java : Batavia.. Bentjoeloek. . . India: Madras.. Chingleput.. . . . . . Calicut. Vizagapatam.. Lahore.. . Murree.. . Shillong.. . Calcutta. . . Bombay.. Madura. . Malaya, Kuala Lumpu Philippines, Manila.. Syria, Beirut.. Total.. . . . . 5 r. -- - 2 2 1 2 5 8 20 ; 2 -- 29 44 18 ---- 34 44 118 3 1 20 23 25 24 17 10 1 4 19 9 17 5 16 2 1 17 2 4 6 6 15 6 7 -- 84 3 -- 92 64 61 23 58 17 28 9 22 26 23 46 27 42 180 88 3.3 87 2 2.3 25 61 34 30 19 52 41 55 52 27 60 ---- I.29876 2 0.23 -~ 2 ( - confusing and not clear cut in that a relatively large proportion of specimens gave inconclusive results in which three or four mice of a group of six used for testing a specimen survived. After the amount of the virus used was doubled in September, 1932, and a 20 per cent infected mouse-brain suspension was used, the pro- DISTRIBUTION OF YELLOW FEVER IMMUNITY 141 YELLOWFEVERIMMUNITY INEUROPEASlA,AND AUSTRALIA MAP I. DISTRIBUTION OF IMMUNITY TO YELLOW FEVER IN EUROPE, ASIA, AND AUSTRALIA Outer circles represent adults. Inner circles represent children. Black sectors show the proportion with protective serum. Circles crossed by diagonal lines signify that no tests, or less than 10, were made in the age group. 142 W. A. SAWYER, J. H. BAUER AND L. WHITMAN DISTRIBUTION OF YELLOW FEVER IMMUNITY 143 portion of sera giving such partial protection became very small and the test became more clear cut and highly specific. The quality of the test was further improved by using only strains of mice which were known to be very susceptible to yellow fever. Among the large number of specimens received from India, there was one from a Tamil, male, 20 years of age, who lived in Chingleput and had been all his life in the Madras Presidency. This specimen gave an inconclusive result in one test with 10 per cent virus and protection in another. Unfortunately, there was not enough serum left for further studies and, as we have not been able to secure another specimen from this donor, it is listed among the positives in table 1. A specimen giving similar re- sults was obtained from a 13-year-old child of Calcutta. It was tested three times with a 10 per cent virus emulsion, and gave inconclusive results twice and full protection once. A second specimen from the same child secured eight months later was tested against 20 per cent virus and the result was a clear cut negative. The result is therefore classified as negative. Among the group tested from Chingleput, Madras Presidency, there were two specimens which gave full protection on repeated tests with 10 per cent emulsion, both when the serum was undiluted and when diluted 1: 10. Both donors were male Tamils who had lived in the Chingleput District all their lives. One of them, "P. B." was 17 years old, and the other, "N," was 40. Second specimens were obtained from both fifteen months after the first bleeding. The serum of "P. B." was tested twice against 20 per cent virus-brain suspension and gave an inconclusive result in one test and a clear cut negative in the other. This specimen is classed as negative in table 1. As there was a time interval of fifteen months between the withdrawal of the first and second specimens and a difference in technique in testing them, it is im- possible to determine in retrospect whether the protective sub- stances so prominantly present in the first specimen had dimin- ished in concentration in the course of fifteen months, or whether the negative result with the second specimen was entirely due to the larger amount of virus used in the tests. The serum of the other positive donor, "N," is of great interest. As far as we have been able to ascertain, this donor had never been out of India and therefore had never been exposed to yellow fever infection. The first specimen was received in May, 1932, and was tested against 10 per cent virus-brain suspension. It showed full protection undiluted, as well as in dilutions of 1:2 and 1: 10. A second specimen which was received in June, 1933, was tested against 20 per cent virus suspension and it protected fully when tested undiluted, as well as in a 1: 10 dilution. A dilu- tion of 1: 16 gave inconclusive results, and dilutions of 1: 32 and 1:64 showed no protection. A third specimen was received in April, 1935, and was again tested against 20 per cent virus-brain suspension. It protected fully when tested undiluted and in 1: 2 dilution, while dilutions of from 1: 4 to 1:32 gave entirely negative results. Here then is an instance in which the serum of a person who had presumably never been exposed to yellow fever virus exhibited yellow fever virucidal properties to a degree fully com- parable with that of serum taken from individuals a few years after attacks of the disease. To see whether this supposedly non-specific power would protect against other virus diseases, Dr. Thomas M. Rivers and Dr. Leslie T. Webster, of The Rocke- feller Institute for Medical Research, tested this serum for us by . protection test in mice against the virus of lymphocytic chorio- meningitis and that of St. Louis encephalitis, but found no pro- tective properties against either. Bacterial agglutinins and bactericidal substances have Iong been demonstrated in normal human and animal sera. More recently, trypanocidal substances in normal human serum have been reported by Culbertson (13) in cases in which the possi- bility of trypanosomal infection was definitely ruled out. Very little is known regarding the non-specific virucidal substances in normal human serum; that is, in the serum of persons never infected with the causative organism of the disease under con- sideration. It is true that protective substances against the viruses of herpes and poliomyelitis in normal human serum have been reported by numerous investigators, and a theory has been advanced by Jungeblut and Engle (14) and others that the oc- currence of virucidai and other antibodies in normal human 144 W. A. SAWYER, J. H. BAUER AND L. WHITMAN DISTRIBUTION OF YELLOW FEVER IMMUNITY 145 serum is the result of normal physiological maturation of the human body. But as both these viruses are practically ubiqui- tous in their distribution, the possibility of a missed subclinical infection as the cause of immunity cannot be definitely ruled out. On the other hand, the possibility of contracting yellow fever in India can be safely ruled out. Also we know of no other infec- tion which gives cross immunity to yellow fever. The possibil- ity of cross immunity between yellow fever and dengue has been excluded. Specimens were received from Philippine Army scouts who had recently recovered from dengue, and some of the sera from Miami, Florida, were from persons who were convales- cent from dengue in 1934. As shown in tables I and 2, neither group showed evidence of protection against yellow fever virus. Furthermore, the work of Snijders, Postmus, and Schiiffner (15), and of Stefanopoulo and Callinicos (16) offers additional evidence that no cross immunity exists between yellow fever and dengue. From the evidence accumulated in the course of this study of serum specimens from parts of the world where yellow fever presumably never occurred, we have reason to believe that yel- low fever virucidal substances are present in the serum of some human beings who have not been exposed to yellow fever virus. If the protection test were made sufficiently sensitive, it would probably be possible frequently to demonstrate the presence of such virucidal substances in normal human serum in varying degrees of concentration, ranging from total absence to a level equal in rare instances to that of the active specific immunity following infection, as exemplified by the serum of "N" of India. Although of considerable academic interest, so sensitive a test would be of little value in epidemiological surveys. The test as performed now, using massive doses of virus, is highly specific. Had the present technique been developed when most of the sera shown in table 1 were tested, it seems probable that only one of the 876 specimens would have shown full protection. On the other hand, because the test is so se$ere, there is a probability that some specimens having small amounts of specific yellow fever immune substance due to past attacks of yellow fever may be missed. Absolute specificity could hardly be expected of any biological test involving the use of living animals. COUNTRIES IN WHICH YELLOW FEVER IS ABSENT BUT OCCURRED IN THE MORE DISTANT PAST From the historical records it is apparent that yellow fever was at one time present in epidemic form in parts of southern Europe and in the eastern and southern regions of the United States of America. It has also appeared several times as intro- duced cases in southeastern Canada. In none of these regions have there been outbreaks in recent years. A few cities were selected for investigation to see if the protection test would confirm the evidence of history that yellow fever is now com- pletely absent. The results of the immunity survey are shown in table 2 and map II. The historical data presented below with regard to yellow fever in Europe and North America are from Augustin (17). In Italy specimens were taken in Rome, Naples, and Leghorn. Yellow fever has never been observed in Rome or Naples, but there was a small outbreak at Torre Annunziata near Naples in 1883, the last recorded appearance of yellow fever on shore in Italy. In 1804 Leghorn suffered from a yellow fever epidemic in which 655 persons died, and the disease was brought to the port in 1821 and 1828 but did not spread. As would be expected, none of the specimens from Italy gave protection. The oldest blood donors in Leghorn and Naples were aged 74 and 64 years, respectively. Spain and Portugal were repeatedly infected by ships coming from the Americas or West Africa, and many towns and cities were involved. The great epidemic in Barcelona in 1821 caused about 20,000 deaths. In Malaga in 1803 and 1804, yellow fever caused the death of 6884 and 11,486 persons, respectively. The last epidemic in that city, with 242 deaths, occurred in 1821. In 1890 there were 5 cases in persons who had visited a ship in the harbor, the last appearance of yellow fever in Spain. The age of the oldest of the Malaga blood donors was 63 years. In Valencia the only recorded cases were in sailors from Barcelona in 1870. The oldest donor was 66 years of age. Lisbon had many visitations, and its epidemic of 1857 caused 5652 deaths. The last introduction into Lisbon was in 1879, with the exception of one suspected case in 1880. The age of the oldest blood donor TABLE 2 Results of protection test survel~ in countries in which occasional epidemics or introduced cases mere present in the more distant past t 8 9 6 - 23 - - $2 31 g' - l! l! / - 3 - j 5 7 1 -- 3 - 24 15 25 63 23 49 13 75 -- 481 1 22 13 23 69 -- 380 1 1828 1888 1899 1867 1897 Italy: Rome ............................ Naples. .......................... Leghorn. ......................... Spain : ~~alaga .......................... Valencia ......................... Portugal, Lisbon. .................. Canada, Toronto ................... United States: Annapolis, Md ................... Jacksonville, Fla ................. Miami, Fla ...................... Tallahassee, Fla .................. Galveston, Texas ................. Total.. . . . . . . . . 148 W. A. SAWYER, J. H. BAUER AND L. WHITMAN DISTRIBUTION OF YELLOW FEVER IMMUNITY 149 from various parts of Canada. Some had traveled in countries known to be free from yellow fever. The disease is said to have been introduced into Quebec in 1805, 1812, and 1864; into Hali- fax in 1842, 1861, and 1878; and into Prince Edward Island in 1880. In Quebec there were said to have been 55 cases and 6 deaths in an army regiment in 1805. The age of the oldest blood donor was 79 years. The one protective serum recorded in table 2 came from Canada and, like the two positive sera from India, it came from a person who in all probability had never been exposed. The first sera from Canada were tested early in 1932 when 10 per cent mouse- brain suspension was used regularly as the source of virus. The infected mouse brains used in testing some of these sera had been kept overnight in the frozen state in an icebox, contrary to our usual practice. Two specimens showed full protection, and seven gave inconclusive results in which only half of the test animals survived. These nine sera were retested with a fresh 10 per cent mouse-brain suspension, and all with the exception of one gave negative results, according to the present criteria which require that the result be classed as definitely negative if not more than two mice survived out of six or five. This one serum from "M. H." was tested twice more and continued to give par- tial protection each time. Five months later a second specimen was secured from the same person and was tested four times. The results were once negative, twice positive, and once incon- clusive. In the final analysis this serum is classified as positive for yellow fever protection, although we feel that if a 20 per cent virus suspension had been used, the results might have been negative or inconclusive. The serum evidently had slight pro- tective properties, probably non-specific. In the United States of America only a few places were studied, and none of the moderate number of random samples of serum had protective power against yellow fever virus. The results are given in table 2. It had been shown previously by Sawyer (18) that protective sera could be obtained in the United States by seeking out aged persons who remembered having had yellow fever in the epidemics which formerly occurred in the United States, the last in 1905. In the present study the ages of the oldest blood donors in Annapolis, Miami, and Galveston were 58, 68, and 75 years, respectively. The donors from Jacksonville and Tallahassee were children. The limited evidence is consist- ent with the belief that yellow fever has not recently been pres- ent in the United States. The specimens from Galveston, Texas, were supplied by Sharp and Hollar (19), who have discussed the significance of the pro- tection test results. They reached the conclusion that the sup- posed cases of yellow fever in that city in 1897 were not of that disease but were more probably part of the prevailing epidemic of dengue. Serum from five persons who had had dengue and from one who supposedly had had yellow fever in that year were submitted in addition to the random samples and were found to be devoid of protective power against yellow fever virus. YELLOW FEVER IMMUNITY SURVEY IN THE WEST INDIES The early American colonization records indicate that the islands of the Greater and the Lesser Antilles were hotbeds of yellow fever for at least three centuries. In some of them the disease seems to have been present continuously, while in others it was introduced from time to time. But after the mechanism of the transmission of yellow fever from man to man had become understood through the work of Walter Reed and his associates in Havana in 1900, and appropriate anti-mosquito measures had been gradually introduced, the disease disappeared from most of the islands within the first decade of the present century. In some of them, however, the disease persisted much longer, and cases were reported as late as 1921 in Martinique and Guadeloupe, but in recent years no new cases have been observed. As this history indicates, these islands are capable of playing a very im- portant r61e in the spread of yellow fever to the American con- tinent. They lie on important maritime trade routes and are regularly visited by ships from all over the world which are en- gaged in the Central and South American trade. They lie on the main air routes between North and South America. More- over, some of the islands are in fairly close proximity to the South TABLE 3 Results of protection test survey in countries in which occasional epidemics have occurred G 0 1934 E il id 1931, 1932 5 r i 2 1934 1934 E 224 49 2 41 48 50 2 40 50 / 83 52 9 7.3 48 3 6.3 56 3 5.4 64 9 4.1 39 12 27 55 63 74 46 2 4.3 41 23 -- - ,177 30 2.5! 21 22 2 19 20 2 22 21 9.0 10.0 25 9 36.0 19 3 15.8 27 3 11.1 59 ! 39 11 20 19 5 2 2 -- E6 30 15.3 00.0 8.4: - - : ; i : L 2 _- .- - -- 134 15 8 19 13 43 20 27 26 1 9 17 44 13 19 17 -- $46 47 9 16 5 11 2 1 1 4 29 22 19 18 6 -- 211 22 3 5 6 4 17 6 2 2 5 1 6 23 12 8 12 4 j ' ij 4 -- : - 1916 1908 1905 1900 1914 Barbados,f,Bridgetoun.. Cuba: Havana. Santiago.. Holguin. Nuevitas. Cienfuegos.. Pinar de1 Rio.. Jamaica : Kingston. . Port Antonio.. Port Maria.. Puerto Rico: San Juan. Caguas. Arecibo................... Humacao. L' owa . . St. Lucia, Castries.. . Trinidad: Port of Spain.. . San Fernando.. La Brea.. _. . Tunapuna.. . Total.. MAP III YELLOW FEVER IMMUNIlY IN NORTH AMERICA CENTRAL AMERICA AND THE WEST INDIES MAP III. DISTRIBUTION OF IMMUNITY TO YELLOW FEVER IN PARTS OF MEXICO, CENTRAL AMERICA, AND CGBA Symbols as in map I 152 W. A. SAWYER, J. H. BAUER AND L. WHITMAN American mainland, in the interior of which yellow fever is en- demic at the present time. For these reasons, it was considered important that an immunity survey be undertaken to determine whether or not yellow fever is still endemic in certain representa- tive islands. Accordingly, serum specimens were secured for yellow fever protection tests from persons of various age groups living in Barbados, Cuba, Jamaica, Puerto Rico, St. Lucia, and Trinidad. The results of this survey are summarized in table 3 and maps IT and III. As seen from the tabulated summary, a total of 1177 specimens were examined. All of 821 sera from persons under 20 years of age gave negative results. Of 356 sera from adults of 20 years of age or over, there were 30 specimens, or 8.42 per cent, which showed definite protection against yellow fever virus. The majority of the protecting sera were from persons over 40 years of age. The youngest adult showing immunity to yellow fever was a Cuban, 28 years old. According to his age, this donor was born in 1908, which is the year when the last case of yellow fever was reported from Cuba, and he may have had an immunizing attack in his early infancy. Recio (20) reported testing a total of 41 sera from donors ranging from 7 to 70 years of age in Cuba, and found 12 positive for yellow fever immunity, the youngest immune donor being 36 years old. The protection test results and the disease records together seem to indicate that the disease has not been recently present in the West Indies. YELLOW FEVER IMMUNITY SURVEY IN MEXICO AND CENTRAL AMERICA It is generally believed that yellow fever was long endemic in Mexico and in Central America, and Boyce (21) is inclined to think that it was present more or less continuously from the sixteenth century to the time of the publication of his treatise (1911). Although explosive epidemic out'breaks were relatively infrequent, there were sporadic cases occurring practrcally all the time, which suggested endemicity. When there was a large influx or accumulation of non-immune human subjects, the dis- ease tended to assume epidemic form. The tragedy which over- DISTRIBUTION OF YELLOW FEVER IMMUNITY 153 took the early builders of the Panama Canal is familiar to every- one. The last widespread epidemics occurred in 1919, 1920, and 1921, when numerous cases of yellow fever appeared over a wide area in Mexico and Central America. After the principal epi- demics had died down, an outbreak occurred in Salvador in 1924 and there were a few cases in Guatemala and British Honduras in that year. Since then occasional cases of fever with jaundice have been encountered, but none has been definitely diagnosed as yellow fever. In fact, we had opportunity to test sera from some of these cases and found no immunity to yellow fever. In the course of the survey of Mexico and Central America, a total of 2271 serum specimens were examined. They were col- lected in 44 different towns or communities, and ages from 5 to 70 years were represented among the donors. The results of the protection tests with these sera as well as the number of speci- mens tested from each locality are summarized in tables 4 and 5. The survey of Mexico was begun early in 1932, and most of the results have already been analyzed by Mejia (22). Since Dr. Mejia's report was written, however, additional specimens have been tested, and the results of the entire survey are sum- marized in table 4. This tabulated summary and map 3 require little explanation, but there are a few points which need further elucidation. It will be noticed that three sera among the 321 obtained from children under ten years of age showed protection against yellow fever virus. These specimens came from Merida, Tampico, and Manzonillo in 1932, and the ages of the donors at that time were given as 8, 8, and 7 years respectively: If the ages of these donors were given correctly, two of the children must have been born in 1924 and one in 1925, i.e., two and three years after the last recognized case of yellow fever in Mexico. It is difficult to explain the yellow fever immunity in these chil- dren supposedly born several years after the disease had appar- ently disappeared from their country. Several possibilities sug- gest themselves, however. These sera were tested early in 1932 when 10 per cent brain-virus suspension was routinely used. As pointed out above, the test as performed then was slightly oversensitive and occasionally produced false positives. It would DISTRIBUTION OF YELLOW FEVER IMMUNITY 155 seem remotely possible that the protective substances in all three of these sera were of the non-specific variety, and not acquired as a result of infection with yellow fever virus. Unfortunately, this possibility could not be further explored as we have not been able to secure second specimens from any of these children. Dr. Mejia has suggested that the ages may have been incorrectly given and that these three children were probably older than stated. There remains another possibility, that the yellow fever infection in a mild unrecognized form lingered in the communi- ties for several years after the last epidemics had died down. There were no protective specimens, however, among those ob- tained from children born after the year 1925, and as there have been no definitely diagnosed cases of yellow fever reported since 1922, it would seem reasonably safe to conclude that yellow fever is not endemic in Mexico at the present time and has probably been completely absent from there during the last ten years. As seen from the tabulated summary, the proportion of im- munes in Mexico increases progressively with the age of the do- nors. In the age group of 10 to 14 years, 8.55 per cent gave full protection among 386 specimens tested; of 59 sera from donors 15 to 19 years of age 28.81 per cent were positive, and for the adults of 20 years and over the percentage became 42.72. As a result of repeated exposure to epidemics, the percentage of im- munes in some localities was very high, as for example in Merida, where 26 of 40 adults, or 65 per cent, and in Payo Obispo where 10 of 15 adults, or 66.6 per cent, proved to be immune. The results of the survey in Central America, shown in table 5, correspond in general with those obtained in Mexico. The dis- tribution of immunes among the various age groups is similar, although the percentage is somewhat lower than in Mexico. It will be noticed in table 5 that of a total of 135 sera from children under 10 years of age there was one protective specimen from Salvador, bringing the percentage of immunes for this age group to about the level found in Mexico. This positive serum came from a child in Usulatan, Salvador, in 1934, when 20 per cent virus was being used in all tests. At that time the child was said to be 9 years old, which would mean that he was born in 1925. TABLE 5 Results of protection test survey in Central America 1921 1934 1919 1934 1920 1934 1924 1934 1918 1935 1910 1935 1924 1935 1 1935 .- - Y - 6 1 16.7 12 1 8.3 16 3 18.8 6 1 16.7 10 4 1 25.0 13 2 15.4 3 15 426.7 43 7 16.3 20 945.0 43 1023.2 17 1158.8 56 1425.0 17 423.6 53 5 9.4 26 5 19.2 66 5 7.6 8 112.5 31 2 6.5 37 1232.4 67 1420.9 16 20 20 1 5.0 5 1921 1924 1936 11 12 6 1 16.7 14 1 7.1 49 2 4.1 21 838.1 45 920.0 21 1 4.8 47 1 2.1 13 1 7.7 42 1 2.4 23 1565.2 52 1630.8 1905 1919 1921 1936' 1936 14 5 35.7 21 1152.4 76 2026.3 1 1934 11 1 9.1 `20 945.0 50 1029.0 1919 1934 4 20 1155.0 56 11,19.6 1919 1934 - - - 5toByem 5.3 7.1 .3.8 - - 1 ; )r -_ 1 1 - - 29 4 15 26 3 1 3 6 13 11 6 9 1 5 5 6 10 12 4 6 1. El Salvador: Usulatan. ................. Sonsonate ................. San Miguel ................ San Salvador .............. 2. Guatemala: Retalhuleu ................ Puerto Barrios ............ Zscapa .................... Guatemala City ........... 3. British Honduras: Stann Creek ............... Belize .................... 4. Honduras : Puerto Cortea ............. Cholutecs ................. Amapala .................. 5. Nicaragua: Bluefields ................ Rivas. ................... Chinandega .............. 6. Costa Rica: I San Jose.. . . . . . . . . . . . 2 3 10 2 17 Alajuela. . . . . . . . . . . . . . . . . 1910 4 6 13 23 Liberia.. . . . . . . . . . . . 521.7 46 510.9 1899 5 15 5 20 735.0 45 Puntarenas.. . . . . . . . . 7 15.5 lQO1 2 11 10 15 213.3 38 Port Limon.. . 2 5.3 1902 . . . . . . 4 17 8 15 426.7 44 7. Panama: 4 9.1 1910 David.. . . . . . . . . . . 3 15 15 12 45 Bocas de1 Toro.. 1 . . . . . . . 9 a 13 26 830.8 56 Canal City and Panama 8 14.3 1995 City.. . . . . . . . . . . 6 8 17 64 1320.3 95 1313.7 1905 ---_-- ---------_ Total _...............,,,. 135 1 0.74 317 6 1.90 244 17 7.00 486 14229221,182 16614.94 1932 1934 1934 1934 1934 1934 1934 - 158 W. A. SAWYER, J. H. BAUER AND L. WHITMAN According to the records available, yellow fever was reported from San Salvador, the capital of the republic, as late as 1924, but no cases are known to have occurred in Tisulatan, the town in which the child lived, since 1921. Here again, as with the three similar specimens from Mexico, we can offer no definite explanation for t,his finding. We believe, however, that the fac- tors, whatever their nature, which enabled us to demonstrate yellow fever virucidal substances in the blood of the three chil- dren of the same age group in Mexico, are also responsible for the positive finding in this instance, and we are inclined to the con- clusion that yellow fever persisted for a time after the last re- corded cases. As with the Mexican cases, we were unable to secure another specimen of serum from this child and the test therefore could not be repeated. In studying the results in table 5, it will be noticed that of the seven countries included in this survey, Costa Rica and Panama stand out in sharp contrast to the others. In Costa Rica yellow fever has not been reported since 1910 in any of the five towns where sera were obtained for the immunity survey. For some unknown reason these towns apparently escaped infection during the epidemics of 1919, 1920, and 1921, which were widespread over Central America. This is fully corroborated by the pro- tection test results. It will be noticed that there were no im- munes among the 115 persons under 20 years of age examined from Costa Rica. In Panama the last case of yellow fever oc- curred in 1905, and none of the 21 donors of protective serum was born after that date. The Panamanian territory south of the Canal Zone is being studied at the present time. On the basis of the data presented above, we are inclined to the conclusion that yellow fever is not endemic at present in the American continent north of the Panama Canal. In fact, our results indicate that the disease was widespread in Mexico and Central America in the past but probably disappeared entirely within recent years. It is true that the sera from very young children tested in the course of this survey have been relatively few considering t,he wide territory covered, but nevertheless the findings seem significant because in them are represented the DISTRIBUTION OF YELLOW FEVER IMMUNITY 159 cities and towns which are of greatest interest from an epidemio- logical' standpoint. Most of these cities are either important centers of communication or places which were known as notori- ous hotbeds of yellow fever until relatively recent years. Whether the disease may be present in some limited rural areas in the form of the so-called "jungle yellow fever" recentIy re- ported from Brazil and Colombia (Soper (7, 23, 24)) is impos- sible to determine definitely from the results of our survey. But the absence of immunity in children born within the last ten years as well as the absence of recognized cases from both urban and rural communities during this period would seem to render the existence of this type of yellow fever in Mexico and Central America highly improbable. SUMMARY 1. Evidence is presented to show that the yellow fever protec- tion test in mice, as used in epidemiological surveys, is highly specific. 2. Of a total of 876 human sera from Asia and Australia, where yellow fever has presumably never been present, only 2 speci- mens, or 0.23 per cent, showed protection against yellow fever . . vu-us. Among 481 sera from Italy, Spain, Portugal, Canada, and the United States, where yellow fever was formerly present but is now absent, only one was protective. 3. A total of 1177 sera were tested from the following islands of the West Indies: Barbados, Cuba, Jamaica, Puerto Rico, St. Lucia, and Trinidad. Among them were 821 from persons under 20 years of age, none of whom gave blood-serum which was pro- tective against yellow fever virus. Of 356 sera from adults over 20 years old, 30, or 8.42 per cent, showed immunity to yellow fever. 4. A total of 1089 sera were tested from Mexico. There was a large percentage of protective sera from donors of all age-groups except very young children. All sera from children born after the year 1925 gave negative results, suggesting that yellow fever has probably disappeared from Mexico within recent years. 5. A total of 1182 specimens were tested from the seven Cen- DISTRIBUTION OF YELLOW FEVER IMMUNITY 161 160 W. A. SAWYER, J. H. BAUER AND L. WHITMAN tral American countries. The results for El Salvador, Guate- mala, British Honduras, Honduras, and Nicaragua were similar to those for Mexico. The sera from persons under 20 years of age in Panama and Costa Rica gave no protection against yellow fever virus, while the percentage of immunes in the older persons considered as one group was about the same as in the other Cen- tral American countries. We wish to express our appreciation for the whole-hearted co- operation of the medical anthorities and other persons of the vari- ous countries included in this survey who were instrumental in collectini and sending the sera for protection test. Needless to say, without their cooperation these studies would not have been possible. We received specimens from the places mentioned through the courtesy of the following: Dr. Jean Macnamara and Dr. Margaret Ashton from Melbourne, Australia; Dr. Nichols and Dr. R. Briercliffe from Colombo, Ceylon; Dr. J. B. Grant, Dr. II. L. Amoss, and Dr. I. C. Fang from Peiping, China; Dr. C. Bonne, Dr. W. M. Bonne, and Dr. Pet from Java; Dr. K. S. Shah, Lieut. Col. R. Knowles, Lieut. Col. S. S. Sokhey, Dr. C. G. Pandit, Licut. Col. 1-I. H. King, Major W. J. Webster, and Dr. Vankatraman from India; Dr. Peter H. Martin and Dr. A. W. Kingsbury from the Federated Malay States; Major G. C. Dunham of the U. S. Army Medical Corps, from the Philippines; Dr. H. A. Yenikomshian from Syria; Dr. Lewis W. Hackett from Italy; Dr. Rolla B. Hill from Spain and Portugal; Dr. J. G. FitzGerald and Dr. D. T. Fraser from Toronto, Canada; Dr. Henry Hanson, Dr. G. N. MacDonell, and Dr. Mark F. Boyd from Florida; Dr. J. H. Janney from Maryland; Prof. W. B. Sharp and Dr. E. 1). Hollar from Texas; Dr. A. M. Walcott from the islands of Barbados, St. Lucia, and Trinidad; Dr. H. P. Carr from Cuba and Mexico; Dr. B. E. Washburn and Dr. T. B. Turner from Jamaica; Dr. G. C. Payne, Dr. 0. Costa Mandry, and Dr. Silva from Puerto Rico; Dr. C. A. Bailey from Mexico; Dr. D. M. Molloy from Central America; Col. J. F. Siler and Dr. L. 13. Bates from Panama. REFERENCES (1) SAWYER, W. A., AND WHITMAN, L. i Trans. Roy. Sot. Trop. Med. and Hyg., 1936, 29, 397. (2) BEEUWKEB, H., AND MAHAFFY, A. F.: Trans. Roy. Sot. Trop. Med. and Hyg., 1934, aa,39. (3) BEEWXES, H., MAHAFFY, A. F., BURKE, A. W., AND PAUL, J. H.: Trans. Roy. Sot. Trop. Med. and Hyg., 1934,28,233. (4) BOY$ L.: Bull. de 1'0ffice int. d'Hyg. publ., 1933, 26, 1015. (5) KERR, J. A., AND PATINO CAMARGO, L.: Revists de Higiene (Bogota), 1933, 2,63. (6) SAWYER, W. A.: The Harvey Lectures, 1934-35, Series xxx, The Williams PI: Wilkins Co., Baltimore, 1936, 6692. (7) SOPER, F. L.: Quarterly Bull. Health Org. League of Nations, 1936, 6, 19. (8) HUQHES, T. P., AND SAWYER, W. A.: Jour. Amer. Med. Assoc., 1932,99,978. (9) STOKES, A., BAUER, J. H., AND HUDSON, N. P.: Amer. Jour. Trop. Med., 1928, 8,103. (10) BEEUWKER, H., BAUER, J. H., AND MAHAFFY, A. F.: Amer. Jour. Trop. Med., 1930, 10, 305. (11) TAEILER, M.: Ann. Trop. Med. and Parasitnl., 1930, 24, 249. (12) SAWYER, W. A., AND LLOYD, W. L.: Jour. Exper. Med., 1931,64,533. (13) CULBERSTON, J. T.: Arch. Path., 1935,20,767. (14) JUNQEBLUT, C. W., AND ENGLE, E. T.: Jour. Amer. Med. Assoc., 1932, 99, 2091. (15) SNIJDERB, E. P., POSTMUEI, S., AND SCEI~FFNER, W.: Ned. Tijdschr. v. Gen- eeak., 1934, 3, 3253. (16) STEFANOPOULO, G. J., AND CALLINICOS, G.: Compt. rend. Sot. biol., 1932, 110,123O. . (17) AUGUSTIN, G.: History of Yellow Fever, 1909. New Orleans, Searcy and Pfaff. (18) SAWYER, W. A.: Jour. Prev. Med., 1931, 6,413. (19) SHARP, W. B., AND HOLLAR, E.: Amer. Jour. Trop. Med., 1935,16,247. (20) RECIO, M. A., Bull. Acad. Med., Paris, 1934,112, No. 35. (21) BOYCE, SIR RUPERT: Yellow Fever and Its Prevention. E. P. Dutton h Co., New York, 1911. (22) MEJIA, H.: Bull. de 1'Office int. d'Hyg. publ., 1936, 28, 1298. (23) SOPER, F. L.: Revista, de Higiene (Bogota), 1935,4,49. (24) SOPER, F. L.: Revista de Hygiene e Saude Publica (Rio de Janeiro), 1936, 10, 107.