341 
The last specimen of vitrified fort examined was from Knock- 
farrel ; the vitrified portion was very distinct, but owing to the 
small quantity at my disposal, the preliminary experiments had to 
be omitted. The rocks collected in the immediate vicinity were for 
the most part red granite and gneiss. 
The specific gravity was 
2*85, and the chemical composition 
Silica, 
54*42 
Peroxide iron* 
15*19 
Oxygen ratio as 1:2. 
Alumina, . 
19-72 
Magnesia, . 
2-49 
Lime, 
6-42 
Oxygen quotient, 0’57. 
Soda, 
0-07 
Potash, 
0-61 
98-92 
It will be seen by comparing the whole analyses, that the two 
surfaces of the vitrified fort Duno’deer, differ from each other in 
very slight degree, and, from the presence of alumina and soda, it 
appears most probable that the felspar with which it is associated 
has played an important part in the formation of the glass, its 
fusion being aided by the alkali furnished in the ashes of the wood 
used, the lime might also be derived from the same source. 
The vitrified fort, Knockfarrel, resembles that of Duno’deer, al- 
though not closely, and from the fact of micaceous rocks being found 
in the vicinity, it is most likely that these rocks entered into the 
composition of the vitrified material. 
Of the small quantity of metallic iron found in Duno’deer it per- 
haps is scarcely necessary to say much ; but it is quite possible that 
it may either have existed as such in the basaltic rocks, or have been 
reduced from the state of oxide, a constituent of augite and olivine. 
The following general conclusions may be drawn from the infor- 
mation afforded by the analyses : — 
ls£, That these vitrified surfaces must have been produced by the 
bringing together of certain rocks, and exposing them to a high 
temperature. 
2ci?, That these rocks were not collected without regard to their 
relative powers of melting. 
3 d, That such a collection of rocks, basaltic, gneiss, and felspathic, 
would at high temperature produce a glass with wood ashes, simi- 
lar in composition to these vitrifications. 
I subjoin a tabulated view of the analytical results : — 
VOL. iv. 2 z 
