Ward ’ — Thames Bacteria , III . 247 
i?. subflavus and i?. ochraceus are also connected by this form 
to the true Proteus type. 
Having regard now to all the facts, there can be little doubt 
that my Nos. 7, 23, and 31 are allied to the Proteus type ; and 
if we admit this, and bear in mind that they all become yellow 
as the cultures progress, the same is true of Nos. 17, K, and 
B. arbor escens , in spite of their assumption of the yellow 
pigments at so early a date in their development. My own 
conviction is that we have in Hauser’s three £ species ’ of 
Proteus , and in my six forms 7, 17, 23, 31, K, and B. arbor es- 
cens a set of varieties of the same species, the definition of 
which rests on their agreement in size and form, mode of 
growth, formation of zoogloeas and tresses, method of lique- 
faction, requirements as to temperature and so on, but that 
they vary in detail as to rapidity of growth and liquefaction, 
and consequent differences in the extent and appearance of 
the colonies, and as to the intensity of pigmentation, which 
may well be assumed to vary with the above. 
But if this is so, we shall have to go further, and to regard 
Frankland’s B. aquatilis as belonging to the same series — it 
may well be a merely weaker variety of his B. arborescens — 
and Zimmermann’s B. radiatus also, for this is almost certainly 
identical with my No. 23. 
The most serious difficulty here is as regards No. 17, for 
it shows unquestionable resemblances to Zimmermann’s 
B. ochraceus. If these are also identical, then we shall have 
to consider the question of a possible (indeed probable) alliance 
between this Proteus group and the whole of the golden- 
yellow liquefying forms which I have brought together as 
Group IX. 
The following varietal forms of this type may be sufficiently 
characterized by diagnosis as below. 
Varieties K and ( 5 ). Fig. 63. 
Habitat. In the Thames in 1893-4. 
Morph. Slender inactive rods 3 to 4 ^ x 0-75 \i. 
