468 
Copeland and Kahleriberg. 
jury, and found alum entirely harmless. Aluminum is nearly 
always present in Lycopodium , and is an occasional constituent 
of many flowering plants: according to G-aze (6), Al 2 O s makes 
up 0.3259 per cent, of the dry weight of Hydrastis roots. 
The poisonous character of salts of copper is universally rec¬ 
ognized, and makes itself apparent whenever any salt which 
dissociates to form copper ions, enters the plant through the 
leaves (when it is used as a fungicide) or from the ground. 
Molisch (16) finds the sulphates of nickel, cobalt, zinc, and cop¬ 
per to be poisonous. Tschirch (29) holds that copper is not a 
poison, but that its salts are corrosive and therefore injurious; 
if his view is correct, it must be the salts that are effective 
in our experiments, but we would not like to rely upon this 
proof. According to Gunther (7. Ref,) copper acts upon fungi as 
a stimulant when very dilute; when not so dilute, as a poison. 
The concentration of copper salts (dissociated) necessary to kill 
Lupinus in twenty-four hours according to Kahlenberg and 
N 
True (9. p. 96), is gg-^. 
are fatal at a concentration of 
On the same authority nickel salts 
N 
and cobalt salts at the 
25,600’ 
same point. Phillips (23) quotes Frey tag to the effect that 
nickel has about three-fifths of the toxic power of cobalt. Other 
authority makes the difference greater; thus Heald gives 
N 
Moo 
Co S0 4 and 
N 
51,200 
Ni S0 4 as solutions just permitting the 
growth of Zea Mais , and Richards (25. p. 686) finds the most 
favorable concentration for the vegetable development of Asper¬ 
gillus niger to be 0.002 per cent. Co S0 4 or 0.033 per cent. 
Ni S0 4 ; their toxicity is probably in the same proportion, and 
the per cent, data are comparable because nickel and cobalt 
have about the same atomic weight. Our results agree with 
those of Richards, that cobalt is decidedly more poisonous than 
nickel. 
The influence of iron upon plants depends upon whether it 
occurs in solution as a crystalloid or as a colloid substance. In 
our experiments enough iron dissolved to discolor the solution, 
but it was probably in large part colloidal. That it didnotdif- 
