408 ON SENSORIAL VISION. 



ously varied by that cause. They are very frequent. In 

 the great majority of instances the pattern presented is 

 that of a lattice work; the larger axes of the rhombs 

 being vertical. Sometimes, however, the larger axes are 

 horizontal. The lines are sometimes dark on a lighter 

 ground, and sometimes the reverse. Occasionally at 

 their intersections appears a small, close, and apparently 

 complex piece of pattern work, but always t-^o indis- 

 tinctly seen to be well made out. The lattice pattern if 

 constant, and if always upright, might be explained by 

 the habit of looking fixedly at a lattice window, with a 

 view to notiuGf the order of succession of colours in the 

 Ocular Spectra, which this mode of viewing them shows 

 finely. Occasionally, however, the latticed pattern is 

 replaced by a rectangular one, and within the rectangles 

 occurs in some cases a filling in of a smaller lattice 

 pattern, or of a sort of lozenge of fillagree work, of which 

 it is impossible to seize the precise form, but which is 

 evidently the same in all the rectangles. Occasionally 

 too, but much more rarely, complex and coloured 

 patterns like those of a carpet appear, but not of any 

 carpet remembered or lately seen, and in the two or 

 three instances when" this has been the case, the pattern 

 has not remained constant, but has kept changing from 

 instant to instant, hardly giving time to apprehend its 

 symmetry and regularity before being replaced by an- 

 other ; that other, however, not being a sudden transi- 

 tion to something totally different, but rather a variation 

 on the former. 



(i I.) Hitherto I have mentioned only rectilinear forms. 



