DR. ANGUS SMITH ON r PEAT. 327 



early; nor do I know if any bodies give off marsh-gas if 

 merely kept moist and exposed to the air, as the upper 

 layers of peat often are with us, without being immersed 

 in water. 



However this may be, the effect of air and moisture in 

 reducing plants to peat is rapid ; and it is seen best in 

 plants of delicate form, small plants, such as mosses. 

 Hypnum purum, triquetrum, and tamariscmum, with a 

 Galium (as named by Dr. Sclmnck, F.R.S.), were those 

 which grew above the peat which was found most rapid in 

 growth at Cairn Monearn, and of which, in all probability, 

 the whole was composed. This will explain why the struc- 

 ture was so fine, and why it shrunk into such a small space 

 and became so heavy ; it did so because the feeble stems 

 crumbled on a slight attack by the atmosphere. Had they 

 been strong pieces of wood, they would of course have re- 

 sisted ; and thus we find peat of strong fibre very light and 

 bulky. It is often supposed that wood makes peat ; and we 

 hear of forests falling and becoming peat. The forests may 

 interrupt drainage and aid mechanically ; but when large 

 pieces of wood become entangled in mosses they endure very 

 long. It may be that the outside or surface, after being- 

 attacked, prevents the action on the rest. We find forests 

 preserved for ages, and real peat formed in a few years ; 

 this is enough to lead us to give up the opinion that trees 

 form peat bogs. If any peat is formed of the wood of 

 trees, it can only be after a very long time, when the thick 

 stems have decomposed. 



When I say that the formation of marsh-gas is not an 

 essential to peat, so far as I know, I do not mean to deny 

 that it does always occur, especially in warmer regions 

 than I have studied ; and indeed we know, from several 

 sources which I have quoted, that the formation takes 

 place; but then this is the action which allies itself to 

 ordinary vegetable decomposition, and not to peat-making. 



