REGENERATION MEASURES AND THE FORMATION OF SALTPETRE CI 
When the Fpilobium plants have grown older, the nitrate reaction ceases: 
this phenomenon is due not merely to causes connected with plavt physiology 
(older plants are not so addicted as younger plants to gather saltpetre in 
their tissues), but also to the fact that the nitrification of the soil decreases. 
Hjokar high degree, too, the reaction depends, ont howmwitnerenomnd 
iSvartlected by thevfires -Where it has been very. badly burmt, the 
saltpetre reaction of the plants seems to abate more rapidly than 
where it has been less severely attacked. The raspberry, too, seems to 
prefer the less severely burnt to the more severely burnt places. The altera- 
tion which is brought about by fire in the bacterial flora of the ground is 
very thorough-going. Saltpetre-forming and saltpetre-destroying bac- 
förta Wiwiiie hi arealtosethett lacking in. the unburntsöromo duns 
migrate. Samples of soil from the burn nitrify a solution of sulphate of 
ammonia, though only very slowly (see table 6). The denitrification of 
GILTAY'S solution proceeds fairly quickly and develops rather large gas-bub- 
bles, while samples of soil from the adjacent unburnt woods produce no 
change, or only an extremely slow change in the nitrate content of the solu- 
tion (cf. table 10). As in the forest mould soils, so too on burns, the nitrifica- 
tion bacteria are accompanied by denitrifying bacteria (see HeESSELMAN, 
FOT 
The power to produce ammonia, on the other hand, has been only very 
slightly inereased on the burn; the results are somewhat various—it being 
sometimes greater, sometimes less, than in samples of soil from the adjacent 
woods—but the differences are relatively small (see, too, table 3). When, 
however, the ground has burnt less badly, so that a real covering of humus 
is left on the ground, the power to produce ammonia may prove to be very 
considerable, for instance on the spots where one meets with raspberries 
and singing nettles' (see, too,ftable 3, NOS. 9, 10, 19): 
Samples of soil from the burn form very considerable quantities of salt- 
petre on being, stored. Even where the layer of hummus conmsists of 
only a quite thin charred crust, the capacity to form saltpetre may 
prove quite rfespectableltilthas in a series of experiments a sam- 
BUS TOIASORAE firo mu the rommen, I taken, 10, ASU St oroa tormed 
more saltpetre nitrogen per kilogram of soil than similarly treated 
samples from a manured potato patch or a sample of mould soil 
from a grey alder stand. The ground where this sample was taken (see 
fig. 19) consisted of a clayey moraine, overlaid by a quite thin covering of 
small fragments of charcoal and particles of humus. As far as one can 
judge, however, the formation of saltpetre gradually abates, probably more 
rapidly on very severely burnt than on slightly burnt ground; but even 
where the development has gone so far that the ground is overspread with a 
low dense covering of Polytrichum juniperinum, an active formation of salt- 
petre can be observed in stored samples (see table 14). 
In denshiring for culture, the ground is not burnt so severely, as a rule, 
as was the case with the burn at Selsjön. The remaining covering of humus, 
sprinkled with larger or smaller fragments of charcoal has a considerable 
depth,' as a rule. Beneath the charred layer the humus-covering, even when 
it formerly consisted of a very tough and tenacious raw humus has a rather 
mould-like structure. I have not examined in detail how soon after the fire 
