. Prichard, in his elaborate treatise on the _Eastern Origin
of the Celtic Nations_, and the Rev. W. D. Conybeare, (in his _Theological
Lectures delivered in Bristol College in 1831-33_) has shown that it is by
thus analysing the grammatical structure, which forms the very skeleton of
languages, rather than by confining our attention to mere vocabularies,
that we may best detect their true affinities, and has illustrated this
doctrine by a few Welsh examples. In the _West of England Archaeological
Journal_ is exhibited (I believe by the same author) the identity of verbal
forms in the Welsh and Latin languages.

Nevertheless, Archdeacon Williams maintains that two languages may have a
common vocabulary, but different grammars[3]:

"The Latin language, whether from Pelasgic or Achaean influence, adopted
at an early period the Hellenic grammar; and, under the skilful hands
of the bilingual Ennius, became that polished interpreter of thought,
which yields in regularity and majesty to the Greek alone. The Cumri
either retained, which is more probable, a still more ancient, or
invented a grammar, now peculiar to themselves. This, although it be
simple and scientific in the highest degree, is so completely at
variance with all the other grammars of the civilised world, that
scholars who have to acquire it late in life feel the strongest
repugnance to its forms and principles, and are tempted to regard a
language more fixed and unchangeable in its principles than any other
existing, as more slippery and grasp-escaping than the Proteus of the
Grecian mythology."

Since I wrote these extracts, I have been much gratified by the perusal of
Archdeacon Williams's _Gomer_, which I recommend to all interested in this
inquiry.

BIBLIOTHECAR. CHETHAM.

[Footnote 2: In _Transactions of the Royal Society of Edinburgh_, vol.
xiii.]

[Footnote 3: In his _Gomer_ he shows that the Latin and Cymraeg display
great similarity in the tenses of the substantive verb.]

* * * * *

PHOTOGRAPHIC CORRESPONDENCE.

_Box Sawdust for Collodion._--The following will be of some use to your
photographic readers:

I find that, by treating box sawdust with nitric and sulphuric acid (in the
same manner as cotton), and then dissolving it in ether, it gives a far
more sensitive collodion than either cotton or paper, and the pictures
produced by it are of unequalled bri