sing of books in an
ordinary private dwelling. Thus the arrangement of the bays made by the
projections could not be carried out without extensive structural
alterations in one house out of twenty in the country, and not one house
out of a thousand in London. His ideas, however, are wholly practicable
and admirably thorough when applied to the annexe library. It is
interesting to see Mr. Gladstone's calculations as to shelf
accommodation. They were disputed at the time by some cavilling critics,
but have since been shown to be accurate. Mr. Gladstone is speaking[52]
of the bookcases round the walls and the projecting arms, and he
says:--'I will now exhibit to my readers the practical effect of such
arrangement in bringing great numbers of books within easy reach. Let
each projection be three feet long, twelve inches deep (ample for two
faces of octavos), and nine feet high, so that the upper shelf can be
reached by the aid of a wooden stool of two steps, not more than twenty
inches high, and portable without the least effort of a single hand. I
will suppose the wall-space available to be eight feet, and the
projections, three in number, with end pieces, need only put out three
feet five, while narrow strips of bookcase will run up the wall between
the projections. Under these conditions, the bookcases thus described
will carry about 2000 volumes.

'And a library forty feet long and twenty feet broad, amply lighted,
having some portion of the centre fitted with very low bookcases, suited
to serve for some of the uses of tables, will receive on the floor from
18,000 to 20,000 volumes of all sizes without losing the appearance of a
room . . . . while leaving portions of space available near the windows
for purposes of study. If a gallery be added, there will be accommodation
for a further number of 5000, and the room need be no more than sixteen
feet high.'

This estimate of shelf accommodation may be compared with one which was
made by Mr. Justin Winsor, the well-known librarian of the Harvard
library. He says:--'The book room of the Roxbury branch of the public
library of Boston is fifty-three feet long by twenty-seven feet wide,
and having three storeys of eight feet each in height will hold 100,000
volumes. . . . . I doubt if any other construction can produce this
result.'

The building at Hawarden cost, I believe, 1000_l._, but whether this is
with fittings or not I do not know. It is certain that for men whose
b