| |242-257|105-124|68-79|41-44| | | | | | | | | Kermadec Islands | 12 | 251 | 115 | 75 | 43 | | | |244-255|110-121|71-81|40-46| | +----+-------+-------+-----+-----+-------+--------- Totals | 28 | 250 | 114 | 74 | 42 | | | |242-257|105-124|68-81|40-46| | ----------------------+----+-------+-------+-----+-----+-------+--------- The measurements indicate that there is a gradient in size from small in the north to large in the south; however, there is a definite separation in average measurements--ten millimeters in length of wing and four millimeters in length of exposed culmen--between the two populations which are designated as _G. a. candida_ and _G. a. pacifica_. In studying material from Micronesia and the Hawaiian Islands, I (1948:57) pointed out the similarities between birds of the Marianas and the Hawaiians and separated these from terns found in the Caroline Islands. The systematic position of the White Tern in the Gilbert and Ellice islands will remain in doubt until specimens are available for examination. _G. a. royana_ is provisionally retained as the name for the Fairy Tern of the Kermadecs and Norfolk Island; there is considerable overlap in measurements between _G. a. royana_ and _G. a. pacifica_. Measurements have given evidence of the degrees of structural resemblance of the White Terns of the different islands, but it is not certain that the groupings made on this basis are natural; more data is needed on ecology and life history. Of particular importance is to learn whether these birds fly regularly from island to island. On the basis of eleven months of rather continuous observation in Micronesia, I suspect that the White Tern has little tendency to make inter-island migrations. This might account for the differences in size in the populations at Guam in the Marianas (_G. a. candida_) and at Ulithi in the Carolines (_G. a. pacifica_) where only approximately 400 miles of open water separate the two islands. The occurrence of the distinct _G. a. microrhyncha_ in the Marquesas may be accounted for by such nonmigratory behavior. Mayr (1945a:27), however, is of the opinion that White Terns found in the Bismarck Archipelago, the Solomons, Santa Cruz and New Hebrides islands may not breed there, which is another way of saying that they are migrants. Swarth (1934:221) and Murphy (1936:1268) record the wandering of the White Tern to the Galapagos Islands, probably from breeding grounds at Cocos Island. Swarth suggests that the tern is not established at the Galapagos because of the presence of colder water in the area. Murphy (1936:1166) is of the opinion that the South Atlantic White Terns are sedentary, but reports evidence of pelagic migration in the Pacific at the Kermadecs. The fact that _G.