21» PKOCEEDINQS OF TIJE AMERICAN ACADEMY 



nislies but few examples of this extraordinary occurrence, — a complete 

 circumzenithal circle. On the 24th of January, 1838. Lambert* saw at 

 Wetzler a circle, nearly complete, centred about the zenith, with vivid 

 prismatic colors. On the 11th of July, 1749, Anderonf witnessed at 

 Norwich, about five o'clock, p. M., when the sun was nearly 25° high, a 

 white circle around the zenith. Bravais resorts to two expedients for 

 explaining the enlargement of the circumzenithal arc into a complete 

 circle, in a few rare cases. In the first place, the light may strike the 

 vertical side of the prism too obliquely to be transmitted, so that, after 

 being once or twice reflected upon other vertical sides, it may emerge 

 from sides opposite to the usual ones. In the second place, each point 

 of the arc, originally produced, causes a parhelion circle, all of which are 

 superimposed upon the arc itself, as far as it extends. This last opera- 

 tion, however, would produce light without any discoloration. In the 

 halo seen at Cambridge, the centre of the circle was decidedly south of 

 the zenith. This fact requires us to suppose that the parallel axes of 

 the prisms were not exactly vertical. A current in the atmosphere 

 mi^ht change the direction of the descending particles of ice, but could 

 the lateral motion, with the air, and not in it, develop any new resist- 

 ance which would direct their axes away from the zenith ? 



I will now exhibit an experiment with an equilateral triangular prism 

 of glass, and also a hollow one filled with water. The axis is vertical, 

 about which it is made to revolve rapidly by clock-work. With a single 

 prism and sunlight, or any bright and circular artificial light, all those 

 features of the halo may be artificially produced which have been re- 

 ferred to the action of many prisms of ice, with vertical axes : the 

 single prism, in its motion, assuming, in rapid succession, all the possible 

 positions of these many prisms in the atmosphere. The halos them- 

 selves can be produced artificially, either by a conical prism, or by arti- 

 ficial crystals formed .upon a plate of glass, as shown by Brewster I and 

 others. § 



The sun and moon are sometimes encircled by what are called coro- 

 na?. A corona may be distinguished from a halo in many ways. 1. It 

 is much smaller even than the smallest of the two halos. 2. It is not 

 rigidly bound to almost invariable dimensions, as the halo is. 3. When 



* Pogg. Ann. Physik und Chemie, xlvi. p. 660. 



t Phil. Trans, xlvi. p. 203. 



} A Treatise on Optics. Amer. edit. 1835, pp. 232, 233. 



<> Amer. Journ. xvi. 398. 



