application of other particles on the windward side, and by the rebound to the lines already laid of those particles which are driven violently against the surfaces between the lines. On smooth, narrow bodies, as this process is continued, the deposits along the sides or edges soon become so thick and long as to meet in the middle. On rough surfaces new lines and centers of groups are begun on all projections, however slight, and the particles rebound to them from the surrounding surfaces.
Fig. 3, a section of rough board, illustrates this. The deviation from the perpendicular in the frost-forms on the edges is due to the fact that the board was not accurately facing the wind.
There is, of course, a great variety of forms produced in different storms, all wonderful for delicacy of design and perfection of finish such as could not be imitated in any
 | |
| Fig. 5. | Fig. 6. |
material. Among them may be shown a branch of balsam fir (Abies Fraseri) (Fig. 4) which bears the heavy fringe of the storm of December 28th, when the wind blew at the rate of fifteen to twenty-five miles an hour, and the temperature was fifteen degrees above zero.
Fig. 5, a pillar and standpipe, shows the perfect fir-tip pattern of January 3d. Wind, fifteen to thirty miles an hour; temperature, ten degrees below zero. The lower temperature and swifter wind account mainly for the difference between this form and the preceding one. The leeward sides of pillar and pipe are thinly coated by the rebound of particles from the house wall.
