NATURE OF THE STREAM 125 



here it may lead to errors. Among those who used it to trace 

 the Gulf Stream in its varying courses during the next twenty- 

 five years were Franklin's nephew^ Colonel Jonathan Wil- 

 liams, and Colonel William Strickland, whose temperature 

 observations gave the first certain evidence of the northeastern 

 branch that brings beneficial climate to Norway. They also 

 discovered mysterious and alien water within the Stream itself. 

 During a cruise of the packet Eliza from Halifax to England 

 in 1810, pools of cool water were found in the middle of the 

 warm current, as much as 15° colder than the surrounding sea 

 and as much as 200 miles wide. It was suggested at the time 

 that icebergs had drifted into the main current and had melted 

 while trapped within the warmer water. As we shall see later, 

 the true significance of these cold patches has been explained 

 only during the past few years. 



The general eighteenth-century concept of the Gulf Stream 

 was that of a fixed and permanent current, but this was seri- 

 ously disturbed by the observations of Captain John Hamil- 

 ton. From his air and water temperatures and current records, 

 taken during a series of twenty-six voyages between Europe 

 and the United States, he concluded in 1825 that the River, 

 continually shifting its position, is so unsteady that definite 

 limits cannot be assigned to it. At the same time more evi- 

 dence was obtained of its northwesterly and southwesterly 

 branches. Somewhat similar conclusions were reached by the 

 celebrated German scientist von Humboldt, and also by 

 Colonel E. Sabine, who were convinced — though without 

 proof — that the Gulf Stream is affected by changes in the 

 strength of the trade winds. 



So far, the increasingly clear conception of a great Atlantic 

 system of currents had grown out of numerous independent 

 and scattered observations rather than from a systematic in- 

 vestigation. The close of this period of relatively random obser- 



