300 U. S. NATIONAL MUSEUM BULLETIN 217 



must have survived the winter. But an important change in the 

 environment of arctic aquatic life has been remarked by John Krog 

 (1954) to occur some time before the waters are ice free. Ice is 

 impermeable to air and in the underlying pools of water which remain 

 unfrozen during the northern winter the oxygen supply becomes 

 much depleted because of the thickness of the covering ice. In late 

 winter the added pressure of water trickling under the ice from 

 melting snow begins to crack the ice on streams and cause the "over- 

 flows" which are so dangerous to travelers. As the season progresses 

 more of this circulating water is returned below the ice by the pressure 

 changes resulting from heat and barometric variations. As this 

 circulation increases, aerated water is mixed with the oxygen-deficient 

 sluggish streams and pools beneath the ice. The restoration of the 

 aquatic oxygen supply begins some time before the breakup of ice 

 occurs and before the temperature of the water changes appreciably. 

 It appears that the metabolism of arctic aquatic plants and animals 

 starts a spring revival as soon as the first localized melting starts the 

 circulation and reoxygenation of the waters. Thus, although the 

 formal breakup of ice is often later than the principal arrival time 

 of the birds, the beginning aeration of the water is earlier. I suspect 

 that this process initiates an efflorescence of aquatic growth some time 

 before the ice breaks and starts production of a new crop of aquatic 

 food in time for the arriving birds, for while watching beaver trap- 

 pers cut holes through several feet of ice in April, I have found the 

 buds of lilies pressed against the bottom of the ice. A few days after 

 the ice melts from Alaskan lakes the lily buds rise above the water, 

 having grown four or five feet from the bottom before an appreciable 

 change in water temperature takes place. 



That the curves for mean temperatures during the time of migra- 

 tion vary greatly from year to year can be seen in figure 15. At 

 both Umiat and at Bettles the mean temperatures of May differed 

 by as much as 6° to 8° C. between 1948 and 1953. The cold mean 

 of —9.1° C at Umiat in May 1952 was but little warmer than the 

 mean of —12.5° C. there in April 1948. In a longer series of arctic 

 records the mean temperature of May can vary ±5° C. If birds 

 guided their arctic migration upon a fixed mean temperature, in some 

 years they would have to vary the date of their arrival by several 

 weeks, whereas the annual variation in arrival of a species at the 

 arctic terminus of their flight is not more than a few days. 



While daily temperatures in May are not so variable as in April, 

 between May 15 and 20 temperatures as low as —16° and as high as 

 18° C. were recorded at Anaktuvuk in 1951 (fig. 5). Wide spring- 

 time fluctuations in temperature are so frequent in the Arctic that 

 birds must commonly encounter days of more severe cold there than 



