VOL. XXXII.] PHILOSOPHICAL TRANSACTIONS. 645 



The next figure is c, in which the pellet has shot out 6 of those small 

 bodies of equal length, and set at equal angles; of this kind I saw a consider- 

 able number. 



The next step in the crystallization is d, in which those bodies are lengthened, 

 and have shot out a great many more from their sides, at equal angles, but 

 unequal lengths, growing continually shorter and shorter, till they terminate in 

 a point. I measured some of these, and found them to be about one quarter 

 of an inch in breadth. I saw but very few of them in perfection, for the col- 

 lateral shoots were so exquisitely fine, as to be liable to be broken in their fall, 

 or confounded together by the least degree of heat. 



Of the next kind, e, I saw a very great number, which being examined by 

 the microscope, plainly appeared to be nothing but the former in disorder. 

 The edges of these were in general very irregular, but some of them happened 

 to be so indented, as to look like the jagged leaves of plants. 



The next kind, f, had twelve points regularly disposed, and probably might 

 consist of two of the former so joined together, as to cut their angles 

 equally. 



Perhaps also those Mr. Morton describes, as consisting of radii, which, 

 instead of terminating in a point, grow larger, as they advance from the centre, 

 might be formed from two of the kind, c, so joined at the centre, as to cut 

 each others angles unequally; for in the progsess of the crystallization, these 

 radii would quickly unite. 



Lastly, that sort which Descartes compares to roses, and of which he has 

 given a figure in his Treatise of Meteors, may be nothing but the kind e, 

 when the points are rounded off by being gently thawed. 



I saw but very few figures of 1 2 points, and those mostly imperfect in one 

 respect or other. 



Observations for four Years on the Aurora Borealis, made at Lyn-regis in 

 Norfolk. N° 376, p. 300. Abridged from the Latin, 



These observations were made in the years 1718» 19> 20, 21, in the winter 

 and spring months of these years, and were mostly of the usual kind. The 

 more extraordinary ones, were Jst, that of Sept. 5, 17 18, about 10 in the 

 evening, when the appearance in the north was as shown in fig. 12, pi. 16. 

 The next night, Sept. 6, between 8 and 10 o'clock, several columns of light 

 were observed like those represented at aa in the same figure, but not so bright 

 as the pyramids observed the night before, which were carried towards the east, 

 but these toward the west. Dec. 1 9, between 8 and 9, several rays of light. 



