Sorosphaera Graminis. 793 
attacked by the Sorosphaera were not so well developed as other healthy 
grasses in the same area, and did not so commonly flower. It was not 
always possible to determine with certainty the species of grass, but fre¬ 
quently the parasite was found in the roots of Poa Annua , a common grass 
weed on gravelled paths and a short-lived annual. Roots and stems of 
Veronica Chamaedrys , growing close to some of the diseased grasses, were 
microscopically examined, but found to be free from disease. 
Structure and appearance of diseased roots. Almost without exception, 
the roots of diseased plants were in parts hypertrophied, although, as a 
general rule, the swollen portion was not invaded by the Sorosphaera , but 
was found to be the seat of ‘ eel-worm ’ infection. Probably the eel-worms 
form one means by which infection with the Sorosphaera is effected or con¬ 
veyed, but as nearly all the grasses in the vicinity of my diseased area had 
their roots attacked by eel-worm, this might account for the fact of all the 
grasses infected by the Sorosphaera showing also the presence of eel-worm. 
However, I found the Sorosphaera much more commonly on the actual 
roots bearing the tubercles with eel-worm than on the non-tuberculate 
roots of the same or close neighbouring plants ; this points to the eel-worm 
as being concerned in producing infection, at any rate the eel-worms would 
enlarge the area of disease by helping in the dispersal of spores. Proto - 
myces Rhizobius , Trail, has been stated to be the cause of the formation 
of small tubercles on the roots of Poa Annua. I have but very rarely 
observed this fungus in the roots, and never actually in the tubercles them¬ 
selves ; these latter, in my opinion, are caused by the presence of the worm, 
and both the Protomyces and the Sorosphaera are secondary infections, and 
mostly confined to parts of the roots which show no signs of hypertrophy. 
The swollen portion of the root may be terminal; not so commonly, 
however, the root may continue its growth, leaving the swelling behind ; in 
this latter case the root usually becomes twisted. The swellings, which are 
often of crescent shape and of small size, are shown in PI. LXI, Figs. 1 and 2; 
in transverse section they show a central stele surrounded by an enlarged 
parenchymatous cortex, the individual cells of which are of normal size. 
They contain cavities caused by the eel-worms which may be seen in them. 
Rarely some of the cells show the presence of the amoebae of the Soro¬ 
sphaera ; the latter are, however, usually to be found in the root bearing 
the tubercle, though frequently the grass plant infected by the nematode is 
quite free from the Sorosphaera. Both the tubercles and the roots them¬ 
selves are free from reserve starch grains. In roots in which the disease is 
in an early stage we find the amoebae of the Sorosphaera in the outer 
cortical cells, some of which may be completely filled by the parasite, 
whilst others may contain but a single mononucleate amoeba or a few 
small-sized independent amoebae. The root-cells themselves are not hyper¬ 
trophied to any great extent, although many of them are considerably elon- 
