1880.] 
63 
[Nichols. 
after the temperature has reached 4°, “ the cooling on the surface 
still continues, but the cooled layers being lighter remain on the 
surface and ultimately freeze. The ice formed thus protects the 
water below which remains at a temperature of 4°, even in the 
most severe winters, a temperature at which fishes and other 
inhabitants of the waters are not destroyed*.” 
I jwesume that these and similar statements may be traced back 
to the time of Rumford’s experiments on the propagation of heat 
in liquids. He seems to have believed that the temperature of 
4° C. prevailed generally, except in a comparatively thin stratum 
at the top which was cooled to the freezing point. He even 
supposed that ground-ice could not be formed in rivers unless they 
overflowed their banks so as to cover ground already frozen. We 
know, however, now, that this is not the case, and Rumford him- 
self 1 admits as possible that the temperature of large bodies of 
water may fall below 4° without freezing at the surface, provided 
the water is agitated by a strong wind. 
I may mention, in this place, that I found on consulting the 
unpublished observations of the U. S. Signal Service, that the 
water at the various lake stations is frequently of the temperature 
of 32° F. (0° C.) from top to bottom for several days before the 
surface is frozen over. At these stations it is true, as before 
stated, the water is seldom much over fifteen feet deep, but I am 
informed by Mr. Chesbrough, formerly City Engineer, that in 
Chicago, on Lake Michigan, the same temperature of 32° F. has 
been observed from top to bottom at the inlet to the tunnel of 
the water-works, which is two miles from the shore, where the 
water is thirty-two feet deep. 
In this connection it may be w T ell to correct another wide-spread 
misconception. It is frequently stated that the water at the bot- 
tom of deep lakes, such as those of the Alps, remains at a constant 
temperature — that of its greatest density — the year around. 
That the temperature at considerable depths is constant, is approx- 
imately true, for the variations, though real, are often only of a 
few tenths of a degree, but that the temperature is that of the 
1 See Vol. i, p. 320, of the edition of Rumford’s Works published by the American 
Academy, 1870-75. 
