557 
of Edinburgh , Session 1881-82. 
of escape being possible. The upward force must necessarily have 
been slight, as it was insufficient to raise the bar lying on surface 
and weighing about 5 lbs. This view the author finds borne out 
by the manager of the foundry where this experiment was made, as 
he states that he has noticed the same action at times, but only 
ivhere air was present , such as when draughts of air played over the 
surface at casting. The presence of air or gas also accounts for 
movements which have at times been observed to take place in the 
hemispherical moulds used for casting balls, as the author has been 
informed by moulders accustomed to this work, that sufficient 
opening of the two halves of the mould is noticeable to show the 
hot solid metal within, but that this is due to air or gases entangled 
in the casting. 
(10) From some experiments made by the author with other metals, 
it appears that when the solid pieces float that their bulk is also 
increased as in the case of cast iron. Thus, when a piece of gun 
metal was placed in a crucible of molten gun metal it sank, but 
shortly reappeared ; on being taken out it was found that it had 
increased in length by -^th part, whilst the shrinkage allowed for 
such an alloy would be even rather less than that. 
Phosphor bronze also floated in molten phosphor bronze after 
sinking, behaving like the cast iron and gun metal. Type metal 
gave doubtful results with light pieces, heavy pieces going to the 
bottom of ladle and remaining there till melted. Lead, when tried 
also with large pieces, sank to the bottom, where the piece could be 
felt remaining there till melted ; and incidentally it may be men- 
tioned that when very large pieces of cast iron are used these 
remain so long at bottom of ladle as to have their ends melted off 
before rising to surface. 
5. On an Oxycyanide, and a New Oxide, of Bismuth. By 
M. M. Pattison Muir, M.A., Fellow, and Prselector in 
Chemistry, of Gonville and Caius College, Cambridge. 
1. Boedeker ( Annalen , 123, 61), by the action of potassium cyanide 
on bismuth nitrate, obtained a reddish-brown salt, to which he 
assigned the formula Bi 2 0 5 .2H 2 0. In an examination of the oxides 
