226 Ward . — Thames Bacteria , III. 
the description of No. 23, a form studied very carefully and 
for long periods — I have kept it under continuous observation 
for some months — with interesting results. 
No. 23: B. radiatus (Zimm.). Figs. 36-42. 
This form is very common in the river at all seasons 
apparently. 
It occurs on the separation-plates as slightly motile rods 
2-6 fx x about o*6 w. After three months in culture the 
plate-colonies gave similar measurements. No success was 
obtained with Gram’s method, though it stains readily by 
ordinary methods, and impression preparations are remarkably 
beautiful. 
Plates at 12-15 0 gave ochre-yellow colonies, more or less 
regularly circular, and often surrounded by a sort of yellow 
halo (Fig. 36). The central part opens out like a star, and 
the clouds around are found to consist of floating or c wander- 
ing islets 5 of much coiled and curved filaments or series of 
rodlets in a zone of liquefaction. 
Plates at 20° : in two days there were circular bluish-grey 
clouds 10-15 mm - diameter, but so tenuous as to be nearly 
invisible. Under | these give the appearance of central 
tresses breaking up into flocculent granular masses. Liquefies. 
(Fig- 37 -) 
Plates at 18-21° in forty hours show the characteristic bluish- 
grey clouds, circular and very indistinct, which under the 
exhibit coils and tresses radiating like mycelial strands from 
a dumb-bell-like or spider-like centre, with serpentine tresses 
running about : the whole hardly yellowish. On the third 
day the circles are about 5 mm. diameter, and pale ochre 
with bluish margin ; granular, opaque and zoned, and begin- 
ning to run and sink. The centre opens out into reticulated 
or sieve-like clouds of granules ; the margin has radial tresses 
like striae. All is composed of tresses of filaments falling 
into cloudy aggregates of rodlets (Fig. 37 d and e). On the 
fourth day liquefaction is commencing, and most of the 
