“ Predisposition ” and “ Immunity ” in Plants. 309 
right in England, Klebahn in Germany, and others, have led to 
the recognition of a state of affairs even more complex than this, 
however, for they have disclosed the facts that not only are there 
several species of rust on our cereals capable of distinction by 
visible characters, and, therefore, morphological species, but that 
each of these actual species may behave differently towards 
different species of cereals or other grasses, thus giving rise to 
physiological races or forms, each with its own distinctive biolo- 
gical characters, but indistinguishable by means of the microscope. 
Eriksson especially has done so much, and such able work 
in connection with these points, that the reader may be referred 
to his papers for further details. Here it will suffice to point out 
that in his latest summary of his work on this species question 1 , 
he accepts no less than thirteen “ species ” of Puccinia on the 
cereals, in place of the three hitherto mentioned, breaking up 
P. graminis into two, P. rubigo-vera into eight, and P. coronata 
into three distinct species, in addition to more than twenty 
specialised varieties or races. With these matters we are not 
now concerned: I mention them here merely to show that P. 
graminis of the text-books is not a single or simple species. 
Of these “ species ” I selected some time ago one which forms 
the characteristic Brown Rust of the Bromes, for an exhaustive 
investigation. I may remark that at the time this work was 
begun, eighteen months ago, the species was known as Puccinia 
dispersa (Erikss.), a form or variety of the old P. rubigo-vera of 
De Candolle 2 , but during the progress of the investigation two 
discoveries have been made, both of importance in other con- 
nections, but not bearing on my work, which is concerned only 
with the behaviour of the Uredo stage. It has been shown (1) by 
Eriksson 3 that this Brome-rust does not form its secidium on 
Anchusa, as does the typical P. dispersa (the old P. rubigo- 
vera ), and (2) by Muller 4 that the secidium-form is developed 
not on Anchusa but on Pul monaria and Symphytum. 
The plan of mjr work was to test (1) how far and in what 
manner this particular IJredine is specialised to the Bromes, and 
the nature of this specialisation ; (2) what are the conditions of 
infection and the relations between host and parasite during in- 
fection and incubation. 
In September last 5 1 gave a brief account of the principal results 
of a long series of infections, the details of which have only been 
withheld because I was anxious to solve some questions depending 
1 Ann. des Sc. Nat. 1901, vol. xiv. p. 101. 
- Eriksson, Die Getreideroste, 18. 
3 Eriksson, Ann. des Sc. Nat. 1901, vol. xiv. p. 101. 
4 Bot. Centr. Beihefte, B. x. 1901, p. 182. 
3 See British Association Report , Glasgow, 1901, and Ann. of Bot. vol. xv. 1901, 
p. 560. 
23—2 
