66 tients after radiation treatment for ankylosing spondylitis,149 etc.), and its etiological role was established without question by the high incidence of leukemia in the Japanese exposed to the atom bomb.150.151 Somefindings in the Japanese tend to support radiation etiology in this Marshallese boy. In his case the latent period was quite long, but the incidence of leukemia in Japan is still (after >25 years) higher among the exposed group than amongthegeneral population. Also, the incidence has been higher among those exposed in childhood, particularly males, in whom the granulocytic form of leukemia has been prevalent; and the same form has been more prevalent in cases with longer latent periods and lower exposures.!52 On the basis of 2 cases per year per million people per rad amongthe exposed Japanese, the expectancy amongthe exposed Marshallese would bea total of about 0.33 cases. However, the normalincidence of leukemia in the Marshallese, from the sketchy statistics available, appears to be low, about 20 cases per year per million compared with 60 to 70 in the U.S.; thus the numberof spontaneous cases expected amongthe exposed Marshallese would be a total of 0.02 to 0.067. Theratio of radiationinduced to spontaneous expectancy (0.33:0.02 or 0.33:0.067) indicates that chances are about 5 to 15 times as great that this Marshallese case was radiation-inducedas that it was spontaneous. Review of the hematological data on subject No. 54 over the 18-year period before the developmentof leukemiais of interest. The greater degree of hematologic depression in him than in the other exposed boysindicateseither a greater radiation dose to the hematopoietic tissues or greater sensitivity or proclivity to marrow depression. (None of the peripheral granulocyte levels in the other boys, exposed or unexposed, plotted individyally, wereas low as his.) Perhaps his radiation dose was actually greater, since he was the youngest Rongelap child exposed and maytherefore have been in closer contact with the fallout source on the ground. Being younger mightalso imply greater radiosensitivity. In view of the later developmentof the granulocytic form of leukemia, it is of interest that the early bone marrow injury was characterized by havingthe greatest effect on the granulocytes, with the thrombocytic and lymphocytic cells less affected and the erythrocytic cells least. The erythro- cytic system did not show evidenceoffailure until near death. In spite of the long-term findings, morbid evidenceof the disease was not apparent until the last few months. Oneyearbefore the diagnosis of leukemia the peripheral blood cells appeared normal except for a possible increase in atypical monocytes. A preleukemic syndromehas been described for some cases of nonlymphocytic leukemias. 153.154 Linmanand Saarni!54 state that the frequency of the syndrome may beashigh as twocasesin five and that preleukemic findings “reflect the early stage of myeloproliferative disorders which will eventually fulfill the criteria to be classified as myelomonocytic leukemias.”’ The syndromeis char- acterized by various findings, notall necessarily present, such as anemia with erythrocyte abnormalities, thrombocytopenia,atypicalplatelets in the peripheral blood, and neutropenia.453.154 Chromosome abnormalities are reported in about half the cases.153 Isolated neutropeniais an occasional!preleukemic finding; in the case reported here, the relative neutropenia is the only finding that might be considered as preleukemic. If the preleukeinic phase is considered to date back to near the time of radiation exposurein this case, then it had a longer course than is usually noted.!54 The possibility that the granulocytic reserve was reduced in subject No. 54 as a result of radiation exposure was indicated byhis granulocyte response being lower than thatof his peers in response to challenge by childhood infections. One mightspeculate that the frequent infections to which the Marshallese are exposed may have plaved a role in accelerating the development of a radiationinduced mutant leukemic clone. This would be similar in some ways to the role thought to he played by the thyroid-stimulating hormone in enhancing the development of tumorsin radiation- injured thyroid glands, as has been seen in the Marshallese years after exposure of their thyroid glands to radioiodinefrom the fallout. The above findings do not rule out the possibility of a viral etiology for this leukemia. Because of the developmentofthis case of leukemia, complete hematological examinations on the exposed Marshallese population are now being carried out semiannually instead of annually. B. OTHER MALIGNANCIES Malignancies other than thyroid cancer and leukemia (see Table 35) appeared to have a higher