bifurcation of the two rivers-a bifurcation which differs from other larger (and more important) examples only so far as to depend upon the state of water in the Clyde. I give the words of the attentive angler who describes it It is

a singular circumstance that salmon and their fry have occasionally been taken in the upper parts of the Clyde, above its loftiest fall, which, being 80 feet in height, it is utterly impossible for fish of any kind to surmount. The fact is accounted for in this way. After passing Tinto Hill, the bed of the Clyde approaches to a level with that of the Biggar Water, which is close at hand, and discharges itself into the Tweed. On the occasion of a large flood the two streams become connected, and the Clyde actually pours a portion of its waters into one of the tributaries of the Tweed, which is accessible to and frequented by salmon."

V. The Dee.-We have already noticed at some length the fall of this river. The results for the Dee I have based upon levels ascertained by repeated barometrical measurements by Dr Skene Keith and Dr Dickie of Aberdeen, which have been kindly communicated to me by the latter. This gentleman also confirms my statements by his personal knowledge of the Dee-that it does not exhibit any cataract from its mouth up to the Linn.

The source of the Dee, rising between Ben MacDhui on the east and Braeriach on the west, is 4060 feet high, and most probably it is the highest source in the United Kingdom. The highest spring on Ben Nevis is only 3602 feet, according to my barometrical measurements—that is, 766 feet below the top of the hill; another spring, on one of the highest hills of the Grampians, Ben Aulder, reaches a height of 3650 feet.

An Analysis of Plate-Glass. By Messrs J. E. MAYER and J. S. BRAZIER.

In going over the analyses of the different varieties of glass which have been recorded, we find that but little attention has been paid to the composition of plate-glass, a material which is almost becoming a necessary of life. It is, moreover, remarkable that no analysis of the plate-glass manufactured in Great Britain has ever been published.

The following pages contain the results obtained from the analyses of three different specimens of plate-glass, which we undertook at the request of Dr Hofmann.* These specimens were procured at the three most extensive plate-glass manufactories of England, which are,

1. The British Plate-Glass Company, St Helens, Liverpool.

2. The London Thames Plate-Glass Company, Bow Creek, Blackwall.

3. The London and Manchester Plate-Glass Company, Sutton, St Helens, Liverpool.

For the purpose of analysis, these specimens of glass were reduced to the most minute state of division, which was effected by levigating in the usual manner. None of the

* I am indebted for these specimens to the kindness of Mr Fincham of the British Plate-Glass Works.-Dr A. W. Hofmann.

specimens, whilst digesting in water, gave any reaction with the most delicate test-papers.

To determine the extent of their solubility in water, from four to five grammes were digested in that menstruum for about forty-eight hours, the clear solution in each case yielded on evaporation but a slight residue, too small for determination.

The specific gravity of these specimens of glass is as follows:

British Plate-Glass,

London Thames Plate-Glass,

London and Manchester Plate-Glass,

2.319

2.242

2.408

A qualitative examination shewed the presence of silicic acid, potash, soda, sesquioxide of iron, alumina, lime, and, in one case, traces of manganese.

The silicic acid was determined in the usual manner by fusion with pure carbonate of potash. The sesquioxide of iron, the alumina, and the lime, were afterwards precipitated from the hydrochloric filtrate.

To determine the alkalies the glasses were decomposed by means of hydrofluoric acid, in an apparatus recommended by Brunner,* which consists of a leaden capsula with a flat bottom, about 6 inches in diameter, and 4 inches high, in the centre of which is placed a small leaden ring about an inch and a half high, which serves as a support for a platinum dish. The leaden capsula has a cover fitting perfectly tight.

To set the apparatus in action it is necessary to cover the bottom of the capsula with a layer of pulverised fluor-spar about half an inch in thickness, and to pour upon it some sulphuric acid, sufficient to form a thick paste. A weighed portion of the finely-powdered glass, after being put in the platinum dish, is covered with water, and placed on the leaden ring. The whole is then kept at a gentle heat either on a sand-bath, or by means of a spirit-lamp.

By a few preliminary experiments we found the action on

* Poggendorff's Annalen, xliv., p. 134.

VOL. XLVII. NO. XCIV.-OCTOBER 1849.

Y

the glass to be exceedingly slow when covered merely with water; it was then suggested to us by Dr Hofmann to try, instead of water, a strong solution of ammonia; we found that the hydrofluoric acid being much more rapidly absorbed by this latter agent, the decomposition was facilitated in a remarkable manner.

The first of the two following Tables shews the amount of substance employed; the results obtained are exhibited in Table II.

*These numbers were obtained in an indirect determination of the alkalies.

A Table is subjoined, containing analyses of several varieties of plate-glass, in order that the composition of the plateglass in this country may be compared with that manufactured abroad. The Venetian glass was analysed by M. Berthier, the Bohemian mirror-glass by Peligot, and the French glasses by Dumas.*

* Comp. Knapp's Technology, vol. ii., p. 16.