656 Ancient Rock Inscriptions on the Lake of the Woods. [July, 
ceeded to carve his initials in the rock with my hammer the 
moment we landed. The weathered and rough character of the 
carving afforded no clue as to the tool used. In size the charac- 
ters varied from about three to twelve inches. There was no in- 
dication of ochre having been rubbed into the carving. The 
characters figured in Plate x1x were scattered over the rock sur- 
face in all directions and in greater numbers than are represented; 
and although the typical ones are gathered together on one sheet, 
that arrangement by no means shows their relative positions. The 
chief advantage to be derived by archzologists from an acquaint- 
ance with such inscriptions is the tracing out the similarity or 
identity of the individual characters with those of inscriptions 
found in other parts of the continent. There is little hope of any 
coherent meaning or narrative ever being derived from such iso- 
lated groups of characters. 
The similarity of some of the characters now figured to those 
described by Mr. Branner from the boulders of Alagéas is a 
striking and suggestive one. For example, No. 2, Plate x1x, is 
identical with the left-hand figure of Mr. Branner’s %, even as to 
the number of whorls and their direction. No. 25 is almost 
identical with æ and 4 of Mr. Branner’s plate. The horse-shoe 
or part-circle shape is distinctly common to both, as may be seen 
by comparing X and d of the Brazilian inscriptions with 17, 18, 
21, 23, 24 and perhaps 29 of those from the Lake of the Woods. 
Nos. 7, 8, 15, 20 may be compared with Mr. Branner’s asterisk 
as simpler forms on the same principle. The circle is also com- 
mon to both sets. Nos. 1, 4, and 14 are similar in character to 
the 3d form from the top of Mr. Banner’s X. No. 10 has nearly 
the same shape as the 3d on the top row of the same group. 
No. 12 is not unlike Mr. Branner’s c and No. 19 is on the same 
principle as the chain_of small circles of his J. But there is no 
need of straining the comparison. The coincidence appears to 
be too strong to be purely accidental, although considering the 
remoteness of the two regions in question, much more abundant 
material for comparison would be required before inferences, even 
of the most general sort, could be drawn. 
= The island on which were found the other inscriptions to which 
nie ok have alluded, is one of the many steep rocky islands known 
_ among the Indians as Ka-ka-ki-wa-bic win-nis, or Crow-rock 
island. The rock is a hard greenstone, not easily cut, and the 
