ROTH] FIRE, STONE, TIMBER, TOOLS 73 
applied (sec. 8). Certain celts were very probably intended for 
wedges. 
7. As to the types of celt and ax found in the Guianas, there are 
certain more or less distinct groups to one or the other of which I 
believe all specimens may be assigned. These groups can be recog- 
nized as follows, but it is only proper to admit at the outset that the 
distinctive peculiarities of some may become imperceptibly merged 
into those of others (except perhaps in the first series), there being 
no hard and fast line of demarcation. In all probability the type 
largely depended upon the shape in which the water-worn pebble 
was originally found (sec. 9). 
i, The first group is distinguished from all the others by the 
presence of a cutting edge at each extremity (pl. 3 A, B), of which one 
may be larger than the other. The celt is elongate and curved, but 
there is no notch or groove for the attachment of a handle as men- 
tioned by Gumilla (sec. 6). Personally, I am inclined to think that 
it may also have been held by the middle in the hand (sec. 8). 
ii. The second group (pl. 3 C, D, E) includes those comparatively 
small specimens with the butt more or less trimmed in the smaller ones 
for insertion into the socket of a haft, as with certain of the clubs 
(sec. 153). The cutting edge is approximately straight. 
iil. Larger specimens with truncated butts (pl. 3 F, G, H) and 
rounded cutting edges comprise a third group, the members of which 
are further characterized by being comparatively squat and heavy. I 
see no reason why these celts may not have been used without handles. 
iv. Narrow flattened celts with markedly pointed butts (pl. 3 
I, J, K) are distinctive of this class. They may be described as nearly 
petaloid in form and as showing but very rarely (I believe in only two 
specimens in the Georgetown Museum) any indication of a groove. 
The cutting edge is straight or rounded. These celts may have 
been hafted into oval-slit boards (sec. 8). 
v. The so-called grooved axes, each provided with a notch or 
groove above and below, which may be either very sharp and narrow 
or correspondingly blunt and open, all intermediate stages being 
noticeable (pl. 4). In the former extreme the butt is more or less 
squat and squared; in the latter it is markedly convex, occasionally 
almost attenuated (IT, pl. x, fig. 4), but its edge, straight or convex, 
remains blunt. The blade, with no sign of a groove across it, is gen- 
erally rounded, but in some specimens takes the shape of a European 
hatchet. The method of fixation of celts belonging to this fifth type 
is a Matter that is still in doubt (sec. 8). 
8. Inthevery rare specimens still extant of celts fixed in rectangular- 
section wooden handles (sec. 153), the stone is secured by fitting 
its base into an opening cut to receive it with exactness, and by the 
60160°—24——6 
